This is an archive of past discussions about Wikipedia:AutoWikiBrowser. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
Another common pattern is word [http://en.--whateversite--.org/wiki/word] which should be replaced by [[whatever:word]].
I would be wary of implementing the [http://en.--whateversite--.org/wiki/word] versions on their own. I have seen quite a few cases where that is used as footnotes. That may not be the correct usage, but converting it to an interwiki link would be worse as it would result in an unintelligible sentence.
Example: Alfred Tennyson's works[1] are should not become Alfred Tennyson's workss:Author:Alfred_Tennyson are.
I needed code for my tool since people didn't know which form to enter in. It has since become convenient to just paste the URL in and watch the magic happen. I hope the AWB devs implement this for the list maker parts of the interface.
Javascript fixTitle code
functionfixTitle(e){// Convert from the escaped UTF-8 byte code into Unicodes=unescape(decodeURI(e.value))// Convert secure URLs into non-secure equivalents (note the secure system is considered a 'hack')s=s.replace(/\w+:\/\/secure\.wikimedia\.org\/(\w+)\/(\w+)\//,'http://$2.$1.org/')// Convert http://lang.domain.org/wiki/ into interwiki formats=s.replace(/http:\/\/(\w+)\.(\w+)\.org\/wiki\/([^#{|}\[\]]*).*/i,'$2:$1:$3')// Scripts paths (/w/index.php?...) into interwiki formats=s.replace(/http:\/\/(\w+)\.(\w+)\.org\/.*?title=([^#&{|}\[\]]*).*/i,'$2:$1:$3')// Remove [[brackets]] from links=s.replace(/[^\n]*?\[\[([^[\]{|}]+)[^\n]*/g,'$1')// '_' -> ' ' and hard coded home wikis=s.replace(/_/g,' ').replace(/^ *(w:|wikipedia:|)(en:|([a-z\-]+:)) */i,'$3')// Use short prefix form (wiktionary:en:Wiktionary:Main Page -> wikt:en:Wiktionary:Main Page)s=s.replace(/^ *(?:wikimedia:(m)eta|wikimedia:(commons)|(wikt)ionary|wiki(?:(n)ews|(b)ooks|(q)uote|(s)ource|(v)ersity))(:[a-z\-]+:)/i,'$1$2$3$4$5$6$7$8$9')// Put back ine.value=s}
A general implementation (suitable for general fixes) for foundation links from the code above:
Find \[http://(\w+)\.(\w+)\.org/wiki/([^{|}\[\]<>"\n]+) +([^]]+)\] replace with [[$2:$1:$3|$4]]
Find \[\[(?:wikimedia:(m)eta|wikimedia:(commons)|(wikt)ionary|wiki(?:(n)ews|(b)ooks|(q)uote|(s)ource|(v)ersity))(:[a-z\-]+:[^{}\[\]]+)\]\] replace with [[$1$2$3$4$5$6$7$8$9]]
I've been using my own regexes for this (though, not as good as the combo above) and would love to see this implemented. Rocket000 (talk) 00:17, 20 June 2008 (UTC)
Would be good to implement this.. Not sure why the first one is needed in the ListMaker...? If you can elaborate/be a bit more specific Dispenser, i shall get this implemented. —Reedy17:44, 23 June 2008 (UTC)
Partially implemented. rev 3036 (code exists, but not in use. As per the discussion page, it doesnt seem to actually work as a general fix or in list maker.......) —Reedy22:20, 3 July 2008 (UTC)
While the code I provided above is good what I had coded it for (a user page input routine) it wasn't good enough for a general fix (potential language issues). Thus I've coded the following which should be nearly problem free:
Python white list code
familiesIWlist={'wikipedia':'w','wiktionary':'wikt','wikinews':'n','wikibooks':'b','wikiquote':'q','wikisource':'s','wikiversity':'v',}forminre.finditer(ur'\[http://([a-z0-9\-]+)\.(\w+)\.org/wiki/([^{|}\[\]<>"\s?]+) +([^]\n]+)\]',text):ifm.group(1)=='commons':iwPrefix='commons'elifm.group(1)=='meta':iwPrefix='m'elifm.group(1)infamiliesIWlist:# don't allow http://sources.wikipedia.orgcontinueelifm.group(2)infamiliesIWlist:iwPrefix='%s:%s'%(familiesIWlist[m.group(2)],m.group(1))else:# TODO: Prevent iw linking on [[Wikipedia]] where it's used as referencescontinuetext=text.replace(m.group(0),'[[%s:%s|%s]]'%(iwPrefix,m.group(3),m.group(4)))
I've implemented the above code in my commonfixes.py library. Over time I've noticed a few problems:
Due to the haphazard way naming was done the above code will try to make download.wikipedia.org and sources.wikipedia.org into invalid interwikis. This can be solved using interwiki table, which would include more links than we can cover in regexes.
Edit on the "Wikipedia" article raises question wheather those links can be linked. And I think I read somewhere that they aren't suppose to be directly, but it still doesn't answer the question about interwiki.
[ { | } ] and more are not valid characters, as well as their escaped counter parts like %7B. However, this highly unlikely since people typically don't link to invalid pages anyway.
For those of you still interested in linking I've come up with a set of regex. The whatever.wikipedia.org glitch is avoid by only allowing languages with two to three letter character codes.
However you should likely add (?<![*#:;]{2}) to the beginning to avoid changing lists and (?![^<>]*</ref>) to the end to avoid changing links in references.—Dispenser01:56, 23 November 2008 (UTC)
According to WP:WAWI, AWB would have to detect the difference between a convenient link to material like wikisource page and a reference link like a wikisource policy page. Luckily that's why we created namespaces.—Dispenser08:07, 29 December 2008 (UTC)
Have care with this suggestion. There are some cases where the ol's deliberate (say, when there is also a colstart or style attribute) rather than as holdovers from some person marking his content up in HTML. --Izno (talk) 17:56, 16 February 2010 (UTC)
There are a number of edge cases where line breaks, comments, templates, and attributes are involved that need to be kept in mind. Some of them may even occur in articles. E.g. (see source):
Foo barLorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.bat
A simple find and replace that finds <br> at the end of a value in a template, and removes it as it is un-nescesary. See Error 59 in WikiProject: Check Wikipedia.
If Not: honorific prefix
Find (with regex multiline, case sensitive) :[\t ]*<[\s\/\.]*br[\s\/\.]*>[\t ]*([\t\n ]*?)(\]\]|}}|\|)
Replace (with regex and multiline, case sensitive: $1$2
We just need to avoid: honorific prefix, honorific-prefix as the request says. I left a note to Infobox officeholder and they 'll probably fix this too. -- Magioladitis (talk) 08:55, 12 October 2010 (UTC)
The point with hon-pref and hon-suf is that in some circs they want the pref and/or suff on the same line as the name, in others they don't. It would be bnest for the template to have the smarts to distinguish, but that would need work and consensus. RichFarmbrough, 16:35, 22 October 2010 (UTC).
Any templates under section headers for "Notes", "References", or "External links" should not be changed. I would be happy to make a more comprehensive list if the developers express interest in implementing this. --GoingBatty (talk) 00:06, 23 February 2012 (UTC)
Yes, this should apply to templates within {{multiple issues}} that have |section=yes. Only templates following a header so that it doesn't change templates in the zeroth section that would apply to the entire article. Not to notes/references sections since these templates would apply to the whole article (especially No footnotes, Refimprove, and Unreferenced). GoingBatty (talk) 01:22, 15 August 2012 (UTC)
If the title and url parameters in a citation template are defined, then the title is linked. If the title contains certain characters, then the link is broken and the display is malformed. The standard fix is to encode the offending characters in the title field, however some editors have been wrapping the value in <nowiki>. While the visual rendering is fine, the metadata in the COinS output is mangled, with the strip markers exposed; see Help:Strip markers for details.
We need to remove <nowiki> tags and encode characters in the title and chapter fields.
remove <nowiki> and </nowiki>
change newline to space
change [ to [
change ] to ]
change | to |
This needs clarification. Is it requested to remove "encode[d] characters" (as well as "nowiki" constructs)? Or is that certain characters (e.g., square brackets) should be represented by their codes ("[")? ~ J. Johnson (JJ) (talk) 21:57, 19 January 2012 (UTC)
Added in revision
Do not include unimportant prefixes in {{DEFAULTSORT}}
I only see "In some categories, sort keys are used to exclude prefixes that are common to all or many of the entries, or are considered unimportant (such as "List of" or "The")." It doesn't say in all. -- Magioladitis (talk) 00:24, 26 November 2010 (UTC)
Adding "The" at the end is unlikely to ever make a difference, but AWB already excludes "The" from DEFAULTSORT tagging. If you look at the edit you linked, AWB did not include "The". McLerristarr | Mclay106:21, 26 November 2010 (UTC)
Most people assume that you've got to keep the case that the source is using, but the MOS advises changing this to standard title case. So, may I suggest pushing a citation's title parameter through:
to fix the most in your face, block-caps titles. I don't think that would leave you with any false positives. Cheers, - Jarry1250(t, c, rfa)16:17, 20 June 2009 (UTC)
When do you propose to convert the case of citation titles? Just when all in uppercase? Do you have some example articles? Rjwilmsi10:27, 28 June 2009 (UTC)
Well, in a perfect world, a citation title of "EXAMPLE: Lorem ipsum" would be converted as well, but the false positive/pointless edit rate would be too high I fear. So yes, just when all in uppercase for maximum efficiency. I would like to see this as a general fix if possible, though I haven't tested the FP rate myself yet. I shall set about finding you an example now. - Jarry1250[ humourous – discuss ]10:31, 28 June 2009 (UTC)
I've had some experience programming reflinks with this, you can get most of the cases right. Here some edges cases
Newspaper Archive: MINOR STORY OF THE DAY; MAN BITES DOG
65_PDF.pdf
SPACE PROBE 56T LAUNCHES
A.I.D.S. EPIDEMIC STILL SPREADING
FOREIGN AIDS STILL MISSING
ATLAS USER EQUIPMENT INTRODUCTION
FIRST ROBOTICS GIVES HOPE
NAVSTAR GPS
J P PENNY
Those are some example I can think off the top of my head. It also a good idea to apply it to the author/first/last/publisher fields as well.—Dispenser12:13, 28 June 2009 (UTC)
Hey, thanks Dispenser. As written, the code doesn't touch .pdf (lowercase), capitalises "Of", and turns GPS to "Gps". The rest it gets right; hopefully, a few tweaks and it should be read to roll. - Jarry1250[ humourous – discuss ]18:58, 30 June 2009 (UTC)
Here's a much improved function for converting to useful title case, which is more palatable than block caps (I personally prefer sentence case, but that would be more controversial / less widely deployable. It works on all the examples above (and some more I invented), with the exception of acronyms that could be words UNICEF, etc. GPS has no vowels, and is therefore easy to capitalise.
publicstaticstringProperCase(stringTextToFormat){List<String>smalls=newList<String>{"and","of","the","but","in","to","a","an"};if(TextToFormat.ToUpper()==TextToFormat){TextToFormat=newCultureInfo("en").TextInfo.ToTitleCase(TextToFormat.ToLower());//Ignore first wordsStringFirstBit="";if(TextToFormat.Contains(" ")){intIndex=TextToFormat.IndexOf(" ");FirstBit=TextToFormat.Substring(0,Index);TextToFormat=TextToFormat.Substring(Index);}foreach(Stringsmallinsmalls){TextToFormat=Regex.Replace(TextToFormat,"([^a-zA-Z0-9])"+small+"([^a-zA-Z0-9])","$1"+small+"$2",RegexOptions.IgnoreCase);}TextToFormat=FirstBit+TextToFormat;String[]Bits=TextToFormat.Split(" ".ToCharArray());for(inti=0;i<Bits.Length;i++){//Capitalise consonant only words, plus a few obvious onesif(Regex.IsMatch(Bits[i],"^([BCDFGHJKLMNPQRSTVWXZ]{2,}|UK|USA)$",RegexOptions.IgnoreCase)){Bits[i]=Bits[i].ToUpper();}}returnString.Join(" ",Bits);}else{returnTextToFormat;}}
Maybe you should use a dictionary from a spellchecker to ensure words like GNU, LIDAR, and CBDTPA stay uppercased? You might also be able to capitalize names Ted Stevens.—Dispenser18:37, 2 July 2009 (UTC)
Yeah... it's a question of how much in the way of resources one chooses to give over to such a minor (albeit intensely annoying to me) thing as capitalisation... hopefully the major acronyms can be hardcoded, and the rest left to the individual editors to catch. As the default Is This Sort Of Capitalisation, We Needn't Worry About Names Of People. - Jarry1250[ humourous – discuss ]18:41, 2 July 2009 (UTC)
And no matter how much effort we throw at the problem, there will always be "yet another exception", e.g. CAT scan
This could be implemented as a general fix that users would have to explicitly turn on via the options menu (off by default) and could be disabled for bots. Question then is just what fields is this required on beyond the 'title=' field of a citation template? Rjwilmsi15:38, 8 July 2009 (UTC)
I have for some time thought such downcasing would be good for the 'pedia but lack the know-how to implement such a fix. I have already built in some case conversions into my formatting script to downcase certain combinations – mainly prepositions of fewer than 5 letters, per the MOS – but not restricted to titles. Most of my fixes are highly specific cases, such as "(\w )O(f|n|r) A(n? \w)" and "(\w )A(nd|t) T(he \w)", or "([Rr]unner)[\- ][Uu](?:ps)\b". I also downcase commonly capitalised "B(oard (?:of |))D(irectors?\W)" and "N(on[\s\-])[Ee](xecutive )D(irectors?\W)". Yet I constantly make a note of exceptions, so the rate of FP is tending to zero. My sources script also aligns cases for sources to the relevant WP article, even to camelcase where necessary, such as "AllMusic".
Sites frequently treat metadata as some sort of dustbin, and some often have FULLCAPS metadata. I suspect part of the fix could lie in WP:Reflinks not importing overcapitalised metadata 'as is'. The problem isn't going away and few people bother to fix these when they do occur; there may be the belief that we should cite sources verbatim including the errant formatting. As suggested by Dispenser above, I assume that we can easily ask for specific cases such as acronyms that are not ordinary words to be exempted ("CAT scan" and not "CAT"). -- Ohc ¡digame!¿que pasa?02:15, 17 September 2013 (UTC)
Multiple hyphens should be removed, especially "--", which is used a lot. In "text--text" it should be replaced with an em dash and in "text -- text" it should be replaced with an en dash. But it should work for more than two hyphens as well. McLerristarr (Mclay1) (talk) 10:19, 2 August 2010 (UTC)
"text -- text" is often used when the intent is to add an em dash, not an en dash (i.e., the intent is to have an em dash with spaces around it). In such cases it should of course convert to an em dash with the spaces removed. I do not see how a program could tell which correction to make. DGG ( talk ) 17:36, 15 April 2011 (UTC)
I disagree. I feel that the intent of "text -- text" is to have some kind of dash longer than a hyphen which is set off by the spaces. This is a stylistic choice, and in my AWB settings, I attempt to respect the original writer's intentions by replacing unspaced instances with em dashes and spaced ones with en dashes. See WP:EMDASH. MANdARAX•XAЯAbИAM18:35, 15 April 2011 (UTC)
Usually is not good enough. I would like to understand why editors ever put such data in the template. Rjwilmsi17:50, 2 August 2010 (UTC)
They put data in order to add extra information before the parentheses that contains the local name closes. I am not sure how we do it in other cases. Do we have two pairs of parentheses one next to the other. Btw, "usually" in this case is equal to "always with probability 1" i.e. in all the cases I observed. -- Magioladitis (talk) 22:01, 2 August 2010 (UTC)
This is a language-rendering template, and use of |extra= for birth/death dates ought to be discouraged because it's not its vocation to collect this data. Biographies will usually contain {{persondata}}, which is where this metadata should be centralised. I fail to see why it should be duplicated in this, or indeed any other, template. -- Ohc ¡digame!¿que pasa?02:28, 17 September 2013 (UTC)
It's not that easy. Think of a math formula. Replacing "2 km² - 400 m²" with "2 km² – 400 m²" doesn't make it more correct. I'm currently using the regex "([\\w'\\)>\\]\\u00C0-\\u024F“]) +-(,?) +(?=[\\w'\\(\\[\\u00C0-\\u024F„])" → "$1 –$2" in my tool to avoid that. --TMg12:53, 11 May 2014 (UTC)
Duplicate report filed:
Currently we fix dashes in limited circumstances. I think there are several other cases that we could cover:
Use of an em dash rather than an en dash in dates and page ranges (there are a surprising number of these)
Spaced em dashes, though whether that should preferably be corrected to an unspaced em dash or a spaced en dash I don't know
Spaced hyphen, in the circumstances where we currently correct a spaced double hyphen. Since these frequently occur in lists, we should fix to a spaced en dash, though many occur in date and page ranges and would need to be ordered before corrections for spacing.
I don't think any of this would be controversial. The spaced hyphen in particular is aesthetically rather unpleasant to a lot of editors.—kwami (talk) 00:51, 21 June 2011 (UTC)
The automatic title bolding, that bolds the title of the article at the start of the article, needs to work for titles that are italicised, resulting in 5 apostrophes around the title. McLerristarr (Mclay1) (talk) 05:17, 11 August 2010 (UTC)
I only mean for titles that are already italicised before AWB works on the article. It would be impossible to program it to know which articles needed italicised titles. McLerristarr (Mclay1) (talk) 03:24, 12 August 2010 (UTC)
No changes inside single straight quotes. This request arises from discussion at the previous feature request: "no_changes_inside_quotation_marks". Most (if not all) quotes have a non-word character on one side of the single straight quote. Most (there are exceptions) apostrophes have a word character on both sides. I'd rather have an occasional miss than be beaten up for a false positive. Could we use a non-word character outside the single straight quote marks as a distinction for super-cautious editing? Lightmouse (talk)
I thought a general fix was an edit. This is not an edit, it's almost the opposite of an edit. I suspect it would need to be an option. Not everybody would need or want to use it. Lightmouse (talk) 13:56, 31 August 2010 (UTC)
You may be referring to single curly quotes. I should have made clear that I'm talking about single straight quotes. Thanks. I've amended the text above. Lightmouse (talk) 14:06, 31 August 2010 (UTC)
I note that HideMore is, by definition, an option for cautious editing. Perhaps the misses are rare enough that it could be added to HideMore for all users until we can selectively choose options. Lightmouse (talk) 14:08, 31 August 2010 (UTC)
When an unbalanced bracket is shown, continue checking by ignoring that bracket type. E.g. If someone uses 1) 2) numbering style you can still find an unterminated <tag! Mark Hurd (talk) 16:42, 30 July 2013 (UTC)
It would be useful to have an explicit AND function else we have to do "..REGEX1 .. REGEX2.. OR .. REGEX2 .. REGEX1.." with 6 combinations for three things, etc. But this is very low priority. RichFarmbrough, 20:02, 31 August 2009 (UTC).
Added in revision
It would be great if it was possible in the skip section to be able to select more words so that if any of those are present then the page is skipped. —Preceding unsigned comment added by Mephiston999 (talk • contribs)
A better phrasing would be... "it would be great if AWB could do more to help people who are entirely n00bish about regex" =) Separating it such that you could have a list of the skip log of which word it matched would be nice as well. –xenotalk19:21, 26 August 2009 (UTC)
Required. Trying to skip if contains 'Category:Living people' or 'Birth date and age|1900|1|1'. I know no way of doing that with regex because | is the separator and thus (Category:Living people|Birth date and age|1900|1|1) don't work. Regards, SunCreator(talk)02:50, 29 April 2010 (UTC)
I found another way around this if anyone is interested. If you use the pre-parse option, you can specify a skip option, then you pre-parse the list again with another skip option. Basically you can recurse your list as many times as you want and doing this is the same result as having multiple skip options, although it is somewhat slower because you reload the article multiple times. Regards, SunCreator(talk)23:29, 29 April 2010 (UTC)
I noticed Regex (Word1|Word2|Word3) is logical OR. How does one use it to find out: word1 AND word2 AND word3? Regards, SunCreator(talk)19:33, 2 May 2010 (UTC)
Exactly the point. You have to do ( Word1.*word2.*word3| Word1.*word3.*word2| Word2.*word1.*word3| Word2.*word3.*word1| Word3.*word1.*word2| Word3.*word2.*word1) (Or (1(23|32)|2(13|31)|3(12|21)) )
This is bad enough but if you have 4 its even more complex . (In fact you can skip all with word 1, delete the list and copy back the skip list, skip all with word 2.. etc... and even mix these methods reasonably getting 6 "ands" in 2 passes.) So it is do-able but it would be nice to have — Preceding unsigned comment added by Rich Farmbrough (talk • contribs) 15:50, 3 May 2010 (UTC)
John(George(PaulRingo|RingoPaul)|Paul(GeorgeRingo|RingoGeorge)|Ringo(GeorgePaul|PaulGeorge))|George(John(PaulRingo|RingoPaul)|Paul(JohnRingo|RingoJohn)|Ringo(JohnPaul|PaulJohn))|Paul(John(GeorgeRingo|RingoGeorge)|George(JohnRingo|RingoJohn)|Ringo(JohnGeorge|GeorgeJohn))|Ringo(John(GeorgePaul|PaulGeorge)|George(JohnPaul|PaulJohn)|Paul(JohnGeorge|GeorgeJohn)) - testing... takes up to 6 paramters. You really don't want to go to 7. RichFarmbrough, 05:49, 28 September 2010 (UTC).
It would be extremely useful for me if AWB has the option to skip pages that do not contain redlinks, redirects, disambiguation pages, etc. It would greatly facilitate the maintenance of Wikipedia books and of several lists. Of course each option should be individually selectable. Let me know if I'm not being clear with what I mean here. Headbomb {talk / contribs / physics / books}07:11, 13 January 2011 (UTC)
Newsletter delivery with AWB often leaves messages for indef blocked users, which are then usually reverted. A skip option "Page of indef blocked user" would allow the operator to avoid such edits in the first place. Useful in particular for implicit lists of users where it's not a matter of removing a category or a name from a list. Amalthea10:05, 4 May 2011 (UTC)
Either have the option to, or automatically skip pages that have no content. At the moment when hitting a blank page AWB will add an orphan, dead-end, uncategorized and stub tag, this is obviously not appropriate.
Offer a way to skip if only change is unicodification. I realize that "only cosmetic" will do this, however, based on my understanding of what will be skipped if that is checked, removal of an element class would be skipped as it doesn't change the appearance of the page at all. So, if unicodification could be added as a separate skip if only option or rolled into genfixes somehow, that would greatly be appreciated. Thanks. — {{U|Technical 13}} (e • t • c)23:12, 8 July 2014 (UTC)
Added in revision
Find and replace improvements
Allow advanced find and replace entries to be marked minor
Allow advanced find and replace entries to be major or minor to work with the new option to "skip if only minor replacement made". –xenotalk22:24, 7 January 2010 (UTC)
The ability to apply the four generic settings (ignore links, ignore templates, add to summary and apply after general fixes) individually to the different rules. mattbr17:15, 26 January 2008 (UTC)
Added in revision
is that so hard to implement? i mean't split "ignore templates, refs, link targets, and heading" into separate options, and improve "ignore images" to ignore "images target" not the whole image section --84.234.42.68 (talk) 17:34, 14 March 2008 (UTC)
For now, how about just the ability to selectively add replacements to the edit summary? That seems like it would be the easiest to implement, and would be very useful to me. I often have intermediate, temporary replacements which should definitely not be in the edit summary, but other replacements which should. MANdARAX•XAЯAbИAM21:47, 21 October 2013 (UTC)
A workaround I use is to set up Find & Replace rules for replacements I want in the edit summary, and Advanced Find & Replace rules for replacements I don't want in the edit summary. GoingBatty (talk) 00:40, 23 October 2013 (UTC)
Please move the ignore options that are in the find and replace options (templates, wikilinks, external links, HTML comments, files, excreta) so that these rules can be applied on a per item basis, often I have find/replace that I want to do on external links, while others that I don't want on external or internal link. Having the ability to set these on a per item basis would make it easier. ΔT The only constant13:44, 15 October 2010 (UTC)
I've run into a need for this myself today. I have a rule that I don't want run on links/refs, but I don't want these ignore options applied to everything else. This is very frustrating as I want all my general cleanup F&R's in one settings file. Stevie is the man!Talk • Work15:08, 19 October 2014 (UTC)
In the Advanced "Find and replace" rule list, allow the easier replacement or removal of HTML attributes on tables. Currently, I'm running all code through a rule to quote all the unquoted attributes and then doing the processing that I want, but every so often a flaw comes up as it quote non-html text. —Dispenser03:44, 27 February 2007 (UTC)
For some find and replace tasks it's useful to ignore certain parts of an article, like <math>...</math> or [[Image:...| (so the name of the image can't get changed). A new tab in Replace Special, in addition to "Replace" and "If", would be nice, so that it's possible to define parts of the articles in which these rule doesn't work. – 84.179.33.6523:21, 7 March 2007 (UTC)
Add in the subrule a new type, like "Entire text" and "Inside template call {{..}}", but it only does the regex on the match from the parent rule. It call it something like "Matched from parent in $1". Implementing this would likely knock out some other feature requests, like my HTML substitution —Dispenser02:26, 8 June 2007 (UTC)
Rule: Find the body
RE find: (<body>.*?</body>)
Sub-rule: Use what was captured in parent
RE find: String1
RE replace: RE1
While the example is a little simplistic, it allow greater flexibly. The String1 in the title tags will never be parsed and there can be many String1 in the body without turning the recursion as with (<body>.*?)String1(.*?</body>). Hopefully that simplifies things. —Dispenser (talk) 04:16, 19 December 2007 (UTC)
I'd appreciate AWB remembering the size/position of windows such as "Find & Replace" between sessions (and also, if possible, the widths of columns within). I guess it could be something saved along with the other settings..? Hope it's straightforward to implement and sorry if it's already on the to-do list. Sardanaphalus (talk) 01:26, 14 October 2008 (UTC) PS Thanks for the recent upgrade.
Wholeheartedly agree. Yesterday I went to resize a column after opening AWB again and accidentally clicked on the heading instead of dragging the separator, thus inadvertently sorting the rules (that's an existing feature request) and forcing me to scrap the F&R rules. Christhe spelleryack13:55, 18 November 2012 (UTC)
I'd need the window/column sizes to be saved in the AWB folder or registry somewhere, not in the main settings file. I keep my settings files on a data stick and use the same files on a big-screen desktop and a small-screen laptop. As an aside, why is the "After fixes" column so wide? I never use it. -- John of Reading (talk) 17:18, 18 November 2012 (UTC)
When entering a regex in the Find/Replace dialog, it's easy (ahem) to forget to tick the Regex checkbox. Could a test be made when clicking Done for [, * and ?, and if Regex is not checked then prompt the user with something like "This appears to be a regular expression. Did you mean to tick the Regex box?".
Too many false positives would get irritating, so deliberately keeping the list simple like that means that newer users, who are more likely to restrict themselves to simpler regular expressions and also more likely to forget, will be the ones who will more often be prompted.
ClickRick (talk) 10:12, 29 June 2009 (UTC)
Added in revision
Any false positives would be extremely annoying. This would have to be explicitly turned on somewhere, which would probably make it of little value for new users. Rjwilmsi15:29, 29 June 2009 (UTC)
"a*" could be a regex or a literal search. Any warnings would have to be explicitly turned on somewhere, which would probably make it of little value for new users. Rjwilmsi09:43, 30 June 2009 (UTC)
I would personally take Jarry's suggestion with a pinch of salt for exactly that reason. To your earlier concern, though, I would say that the prompt should be turned on by default but include a "Do not show this prompt again" checkbox, which should be remembered for that user, so an advanced user will only be prompted the once. An option somewhere else (Options|Preferences..., perhaps?) to turn it back on might be appreciated, too. ClickRick (talk) 10:20, 30 June 2009 (UTC)
Maybe make the default for the regex box, checked? Most non-regex (for a loose value of the word most) also wokr aas regexen, e.g. "Horse" RichFarmbrough, 21:55, 21 April 2010 (UTC).
Copy regex find and replace from the "test regex page" to the active page
It is quite laborious copying the regex to the active screen from the test regex page. I would like a button on the testing page that will put the regex into the active find and replace page on the next available row, and automatically open the active active find and replace page ready for approval by clicking "done". Snowman (talk) 21:15, 10 January 2010 (UTC)
Added in revision
You can do that in reverse. If you go via the FaR, and right click and move to regex tester (not sure on the specifics), you can move it to the tester, edit it, and move it back... —Reedy21:18, 10 January 2010 (UTC)
I will look for that - I have not been looking at what was available in the right click menu. When I have needed to test a complicated regex I have started writing it on the tester, and there is no indication of how to transfer it easily to find and replace. Snowman (talk) 21:35, 10 January 2010 (UTC)
On the find & replace panel, you can mark a row and use the right mouse button to bring up an options panel. On this panel, you can insert a new row. It would be helpful to have a field here to key a 'number' which would insert the indicated number of rows, not just one row. At the same location, it would be helpful to have a 'copy' function, not just an insert function, again with a 'number' so you can copy the marked row the number of times. This would help reduce the amount of repetitive keying now necessary when slight variations in a row entry are needed to handle all the desired changes. Hmains (talk) 19:20, 28 August 2010 (UTC)
On the start pulldown, at the bottom right, there is a place where you can do a string find against material in the edit box. When the find is made, the top line of the edit box shows the line on which the string was found--the 'find line'. As often as not, I have to scroll up one line to get the context of the 'find line' It would be helpful if the find command itself would always point to the line above the 'find line' for context. This is pretty standard on other 'finds' that I have seen. Hmains (talk) 02:37, 2 September 2010 (UTC)
There are two types of advanced "find and replace" feature in the AWB: (1) find and replace works for whole text; (2) find and replace works only inside templates call. Maybe is is possible to make find and replace work only in specified range of article lines, for example, for lines from 1th to 100th, or for lines from 50th to 75th. I wrote some settings for AWB, which make some useful edits, but end of article (roughly 20 or 30 lines) must be unaffected. --Mitas57 (talk) 18:19, 17 January 2013 (UTC)
In the Find & Replace screen, Find column, there is a box that when touched sorts the list of finds. In many cases, the usefulness of the list is destroyed if it is sorted; entries must all be re-keyed. It would be very helpful to have an option the AWB user could set so that the sort box would not be presented for a particular list Hmains (talk) 19:09, 28 August 2010 (UTC)
Actually, this little sort box in the find column header is not always present: sometimes there, sometimes not. ?? Hmains (talk) 03:16, 31 August 2010 (UTC)
I noticed that new button and it is comforting to have. Thanks. Still, it would not help when I have just added/changed many entries in the table and am checking that they are ok and then mistakenly sort everything and destroy my work. Now I can avoid saving and thus destroying my previous-to-this-session work, but I cannot unsort what I just did. If unsort cannot be done (or more generally undo whatever I just did), even a pop-up confirmation message asking 'do you really want to sort this table?' allowing escape before doing the sort would be helpful. Hmains (talk) 01:12, 18 November 2013 (UTC)
I have not tried AWB and I don't really want to set myself up on it since I am on semi-wikibreak. :-) But based on what I have read about it, I have an idea. It seems AWB makes users wait while it loads articles. Why not preload several dozen articles in the background? The server impact would be minimal. (Please feel free to close this or my other requests if they turn out to be based on invalid assumptions.) Cheers, ----unforgettableid
I presume AWB makes users wait while it saves articles. Why not save them in the background? This would make AWB more pleasant to use. ----unforgettableid
Added in revision
Migrate to API? will allow us to complete this much more easily. With the current way, we'd have to be faffing about with multiple browsers (which we already do do...) —ReedyBoy20:51, 22 January 2008 (UTC)
How about a visualization like the pre-parse mode? Pages that are saved without problems are just removed from the list and pages with errors, conflicts and so on are just marked with orange or red background and kept in the list. This way errors are handled gracefully and the fast workflow without delays is still preserved. Penguin (talk) 16:41, 21 May 2009 (UTC)
It will happen, just i've been busy for the last few months. Its on my summer todo list, most of the stuff IS implemented, its just making AWB fully use the API for editing.. —Reedy16:28, 20 June 2009 (UTC)
What I suggest is to bunch up this request and the top request into a server request queue sort of thing. As in, let the user queue up some pages for processing and generating diffs. The user then can do something else, while it queues. Then, the user comes back, and queues up the saves after checking each page. The queue will empty itself as the user is checking the page. That way, your diffs preload, and your saves get executed a bit later, but then you get a continuous stream of pages, instead of a broken one. ManishEarthTalk • Stalk18:37, 19 January 2011 (UTC)
I could still use this feature. When doing typo work with hundreds of articles (which I do at dawiki on a regular basis) the small delays between every save really slows down the process. There are three methods that I want (maybe in combination): 1. Preload the next page (so I don't have to wait for it to load). 2. Background save of article (so I don't have to wait for it to save), 3. Create some kind of bulk save in the end. The revision conflict risk is already there even if we don't pre-load/pre-save. It should just be marked in a sensible way (for bulk saves an option to add conflicts to the could be fine - for background saves a conflict save or otherwise server error could just keep the article in the list with a "notice/warning" background color, just as with pre-parsed links) --Penguin (talk) 20:23, 24 October 2011 (UTC)
I was about to add this suggestion and found it here already. Sometimes the saves are just brutal, taking 5-10 seconds each. If they could be queued, that would be a huge time-saver. I think though that there should be a (1 sec?) time delay built into the save (move to save queue) to help prevent machine-gun/asleep-at-the-wheel saves. And the user should be able to stop saves in the queue before they are executed. Perhaps also have an option to "Stop and move back into the list" where the user wants to re-do their work. Stevie is the man!Talk • Work14:18, 30 August 2014 (UTC)
Another idea. Allow the "make page" button to some how automatically reload. For example, if you use Special:Recentchanges to make a list, have an optional timer that you could use to automatically reload the list. ~Wikihermit01:19, 28 June 2007 (UTC)
Added in revision
And like this idea. Sometime we need make a list from many source. Like "Links on page" from 70 page or/and "from text file" from 10 texts files in one folder. That's hard by hand. --OsamaKBOT16:35, 30 June 2007 (UTC)
As a point for any that are just allowing you to type things into the text box, you can type "Page1|Page2|Page3.....Page55|Page56"
Automatically have AWB do multiple things for bots
I'm guessing this is way too hard and the reason people write their own scripts, but here goes anyway: To start my bot, I must make a list from one place, filter it, make from transclusions and filter. Though it doesn't take long, I'm requesting a way to automate this and put it on a timer. For my other bot, it does take a little longer What I mean is a timer that automatically runs a bot every x minutes, and then performs the things you set it to do (ie. filter, then do this, then filter, then start with appending y). Basically, just tell AWB exactly what you normally do by hand, and have it do everything at the same time with just one click or automatically.. Like I said, probably pretty hard, but mines well request it. --(Review Me) RParlateContribs@(Let's Go Yankees!) 02:04, 15 June 2007 (UTC)
The ability to set up multiple tasks (for example, several template replacement runs), do a test edit with each task to make sure they do what you're expecting. Then hit "run tasks in sequence" and have AWB go through each task in turn. – Mike.lifeguard | talk18:47, 5 October 2007 (UTC)
Added in revision
Well, in theory there isnt... If we could build some form of an instruction set that awb can follow... and all the functions are software call-able, its viable... Reedy Boy08:54, 15 June 2007 (UTC)
Sounds like you're talking about a sort of basic scripting language. If it is an itch someone wants to scratch make sure there are commands/points where you can prompt for user input. --Brianmc12:12, 15 June 2007 (UTC)
Agree, this is very convenient to run two scripts subsequently each with its own settings file and its own input data file. A command line parameter will be fine for script name definition. Mashiah23:29, 13 July 2007 (UTC)
Things like this can in theory already be done with plugins or modules. However, I think that AWB does need this as a standard feature. I'd like it to be able to automatically check certain categories periodically and run a bot job (with seperate settings for each category) too. I was thinking more of keeping it all in the UI and using XML to define tasks myself. Anyway, if at some point I'm looking for something to do I might investigate this further. --kingboyk (talk) 23:22, 26 January 2008 (UTC)
Now we also have the "call external script" functionality, just adding enough to allow AWB to start log on (log on used to work with profiles?) and run would do the trick. RichFarmbrough, 11:03, 11 August 2010 (UTC).
Possible plugins
Allow easy way to add missing parameters to a template, and also a way to reorder parameters
I'm in the process of doing some massive infobox and other template conversion for standardization purposes. The feature to rename template parameters is extremely useful in this process, but it would another big advantage to have a way to add in missing parameters (and set a default value to them if they are missing). It would also be convenient to have a way to reorder all of the parameters so that similar ones can be grouped together, or so that every template instance can have the same ordering as every other one. I guess I'm envisioning a dialog with a grid with three columns: in the first column goes the parameter name, the second column gets the default value, and the third column a checkbox signalling whether to add the parameter (along with its default value) if it is missing. The parameters would then be automatically ordered according to their order in the grid, and added in if missing. Buttons would include "Move up", "Move down", "Delete", and possibly "Don't reorder" for cases when adding missing parameters but no need to reorder. The dialog would be perfect as a new type of rule in the Advanced Find and Replace dialog. Note: this would be used to add infoboxes to every city/town in the US, and add missing data to existing ones. Thanks, --CapitalR12:07, 5 March 2007 (UTC)
Added in revision
I want to encourage this development as well. It's been seven years, and it is a sorely needed functionality not found in AWB. VanIsaacWScont05:06, 18 May 2014 (UTC)
Just to check: Are you aware that everything requested above is possible to do with regular expressions in the find and replace dialog? Admittedly, doing so is much less convenient than what is requested above. It would also be somewhat less robust than would be possible if specifically programmed. — Makyen (talk) 07:02, 20 May 2014 (UTC)
Display more info for images and coordinating image tags added to menu
When a page is in the Image namespace it should check to see what other pages are using that file and also include the uploader history for ease in contact (boxes similar to Alerts->Multiple wikilinks). Would help for determining if Fair-use images are being used outside the main namespace, if an image is orphaned, if image is being used in articles it is not intended for. Tags to be added under an Image menu would be {{redundant}}, {{no license}}, {{no source}}, {{notorphan}}, and {{or-fu}} (with date auto-inserted) -ΖαππερΝαππερBabelAlexandria05:36, 13 March 2007 (UTC)
There are currently 55 other AWB feature requests, 3 dev's inactive, and a v4 to sort. This would be a major new feature, requiring quite a lot of work, as the actual html of the page would probably be needed to be loaded to get the file links, as with images being here and on commons, it doesnt make life easy. Requests that questions are needed to be asked, or input gained, have had replies...Reedy Boy20:22, 25 July 2007 (UTC)
sorry if i sounded antsy... i wasn't aware that four of you were currently inactive during the newest upgrade. I actually had posted this a while ago when there were relatively few features being asked for so I was hoping for some sort of a response. Letting me know that loading the file links by html rather than query.php (i think that's how you typically retrieve the data, right?) presents a large issue helps me put my request into perspective and I appreciate your response :) On the other hand, tacking on a couple more template options to the right-click menu is likely a relatively simple addition. If you guys need the help, i do know c++ and can wade through c# pretty well, but i can't do much that's complicated. I never offered to help before b/c it seemed you guys had things well covered and only recently ported the code into OOP format, right? I'm confident i could actually implement my changes myself (the menu additions) if i knew exactly how to go about it... never worked on a sourceforge project before. -ΖαππερΝαππερBabelAlexandria00:07, 27 July 2007 (UTC)
It was sort of OOPs before, but the code wasnt the best. Kingboyk has done a lot of reworking of the code to add functionality, speed it up, and make it generally better. I have helped with this also, adding some major new features and so on.
If you wanted to add this yourself, i mean, if you want to do it, or at least make a start, and we can help out as and when, that would be fine. It may be worth looking and having a play with query.php and api.php to check and see if they do what you need to. We do use both the query.php, api.php and loading the actual edit pages to pull off the text, we tend not to really load the actual user view of pages too often for pulling off data. If you checkout a copy of the SVN version, have a play and see where you get.
As for the developer side, our "main/lead" developer, MaxSem, has been away from wikipedia for nearly a month now. Feature requests and bugs tend to really just get done by whoever knows how to do it, or wants to attempt to do it, ie personal preference, not really priority/importance of them.
v4 Beta (Alpha + a few changes) should be pushed out to most users this weekend... (Force upgrade... :D) so we'll see how that comes about. Reedy Boy09:32, 27 July 2007 (UTC)
When there is more than one word is highlighted in yellow on a line in AWB, and one of the highlighted is not a typo (i.e Cristian vs Christian both are correct but AWB recognizes it as a typo) when the non-typo is double clicked, it removes everything in that yellow box instead of the specified word, maybe have it where it only removes the highlight because more than one typo could exists on a line or paragraph dputig0700:54, 12 September 2007 (UTC)
Hi, I am using AWB in ml.wikipedia. The Find & Replace option is not displaying Unicode characters. Edit box is working fine. If Someone can add this functionality in next update will be appreciated. --Sadik-khalid (talk) 10:12, 20 November 2007 (UTC)
There are lots of them that don't display on different machines (especially running XP, I suppose). The problem is that there is no standard Unicode font that every user has. Arial Unicode MS comes only with M$ Office (and is too wide to be simply used w/o other changes). Other variants are even less standard. Also, many of them are not suitable to be used in interface, cf Code2000. MaxSem(Han shot first!)19:22, 17 December 2007 (UTC)
I believe AnjaliOldLipi is the most popular font in Malayalam. Here is the AWB screenshot from a win2k system. In XP, it works fine. If there is an option for adjusting font size will be appreciated. Some characters are difficult to read. --Sadik-khalid (talk) 09:19, 18 December 2007 (UTC)
Note, this image was tagged for deletion by the bots. I extended it for a week, but someone may want to declare it as free if that is true, or make sure that they look at it before it is actually deleted by someone else. --After Midnight000113:42, 25 December 2007 (UTC)
I've just started using NewPageWatcher and really like the auto-prod and notify and auto-afd and notify buttons. Could they be made an extra module in AWB when I scan categories like OR or essay it would be most useful. MBisanztalk17:13, 2 January 2008 (UTC)
Added in revision
Sounds like a good idea. It would be easy enough to do for Wikipedia EN. I guess it could be a feature which is invisible if the settings say we're on another wiki. --kingboyk (talk) 23:33, 26 January 2008 (UTC)
I presume users must make multiple mouse clicks to operate the disambiguation repair dialog. Why not allow them to use just the numeric keypad on their keyboard to make their choices from 1 through to 9? Then you could mention this fact in the manual and perhaps onscreen. ----unforgettableid
Just by numbering the first nine options 1 to 9... Typing 5 will use the fifth option. Now you can use home/end and the first letter, which can be annoying if multiple options have the same starting letter. My request on disambiguation: the [done] key has no hot-key likt alt-D. Now I am forced to use the mouse (the tab key will only activate this button after many many button presses). It would be cool to have alt-Done and alt-Cancel available, so you can operate AWB to solve disambigs without the mouse. Yeah, sorry, I hate that mouse and I want to keep RSI/CTS away by keeping on hating it. Edoderoo (talk) 10:58, 11 September 2008 (UTC)
Do you program in C#? Or are you capable at least of designing Windows Forms* in Visual Studio? If you are, perhaps you should join the team. Judging by our efforts up to now we don't have any UI artists aboard. *Or we could go with WPF, which would necessitate a move to .net 3.5, a jump I personally feel we will have to make at some stage... --kingboyk (talk) 12:17, 5 March 2008 (UTC)
I don't know anything about C/C++/C#, yet. But I'd be willing to give the GUI thing as free time becomes available. I assume its part of the Visual Studio package I use to compile AWB.—Dispenser22:28, 5 March 2008 (UTC)
Yes it is. Windows Forms in Visual Studio is fairly simple if you're sticking to design, don't really need to be able to code much to do the visual part.
The .Net Framework 3.5 introduced design/code seperation I believe, a replacement for Windows Forms called Windows Presentation Foundation and a new markup language (XAML?). I've not tried any of these features yet and we're currently using .Net 2.0 with AWB so we're stuck with Windows Forms unless there's a compelling reason to "upgrade". Visual Studio can do both types of design anyway. --kingboyk (talk) 19:17, 6 March 2008 (UTC)
If you can do the graphical side.. It wouldnt be too hard to tie the backend code into it all... I suspect, looking at that, we're gonna need to make a custom control, probably inheriting from listview/similar... As a list view can do the different types of view like you see in Windows Explorer - Large Icon, Small Icon, Details... I may have a play later on —ReedyBoy19:33, 6 March 2008 (UTC)
Bit of googling/looking about. We are certainly going to need either a custom control, or to write our own. As list views by default do not accept images in the sub items (ie columns other than the first) —ReedyBoy20:56, 6 March 2008 (UTC)
There are quite a lot of typos that have had to be rejected for the RETF page because either the correction isn't unambiguous (e.g. 'distict' could be a typo for 'district' or 'distinct', or because it's valid in one context, but not in another e.g. 'Valparaiso' is correct when referring to Valparaiso, Florida, but should be corrected to Valparaíso when referring to the city in Chile.
I'd like suggest an enhancement to AWB to help with situations like those. There would be a new 'Ambiguous Typos' list, much like the current 'Typos' list, with entries along the lines of
AWB would read this list and, on finding the RegEx value in an article, would present a panel much like the current link disambiguation panel, for the AWB user to select from the listed replace options. Colonies Chris08:22, 19 September 2007 (UTC)
This would be a useful feature, provided that users had an option to 'ignore ambiguous typos' i.e. AWB would not change a word matching an ambiguous typo and would not prompt the user for the correct correction. Otherwise I could envisage users being regularly pestered by message boxes ;) Rjwilmsi (talk) 11:15, 30 May 2008 (UTC)
The ability to view block logs and page logs without leaving AWB, for example, if I'm viewing a User_talk page, it would be useful to be able to see whether the user is blocked or the page protected. Even better if this could be done as part of a search parameter (e.g. "skip if indef blocked") but that would probably be pretty complicated. Right now, I have to open my regular browser to view the logs then return to AWB to complete my edit.--Doug.(talk • contribs)19:19, 10 April 2008 (UTC)
It would be extremely helpful if I could just keep the find and replace dialog box open all the time, and have it stay on top of the main AWB window. Right now, however, if I want to scroll down in the edit box when the find and replace dialog is open, I have to close the dialog, scroll down, and then reopen the dialog to continue editing my regex statements. Thanks, --CapitalR (talk) 01:56, 16 April 2008 (UTC)
Added in revision
It isn't modal, see Feature Request Modeless Dialog "Text regex" (Couldn't spell). But I have the main window on my first screen and AFAR on the second. Perhaps, you want an always on top feature or a taskbar button?—Dispenser03:23, 16 April 2008 (UTC)
It would be very useful to be able to see both the preview and the diff at the same time using a split screen setup. This is actually so useful that I recently hacked AWB to allow it (using one of those splitContainer controls), but I think it would be a great option to have available to everyone. Even better than that would be to allow seeing the original page, the new preview, and the diff (or any combination of the two) all at the same time (which would probably only be possible on wide screen monitors, but it would be quite useful). --CapitalR (talk) 02:00, 16 April 2008 (UTC)
Do you think you could make it so that you can scroll through the entire window? I use a laptop with a 1024 * 600 resolution, so some of the window gets cut off. I really like AWB and I just want to be able to use all the features. Oracle Techie 16:40, 1 March 2009 (UTC)
Added in revision
To be honest, im not sure about making the whole window scrollable.. Should be doable, would need to try having a play... That a 9/10" screen by any chance? —Reedy22:37, 2 March 2009 (UTC)
Right now, it is only possible to force Find&Replace to be before/after General Fixes. But it's not possible to set when the External Processing will be executed, Add/Replace/Remove Category, Template Substing etc. I guess it'll be hard to implement, but I believe in your skills.
(rest below, {{AWB feature}} breaks pipes in preformatted box)
Added in revision
How I imagine it to look like: it would be a little window like F&R one, with rows like there, and easier possibility to move row up or down (=make it be executed earlier or later), delete or temp disable operation. Also, a button "Add operation...", which will allow to add custom bunches of replaces or, if they were deleted form list earlier, General Fixes, External Proc. and others from another list.
#Options_list, its a good idea.. but there are some that are supposed to be run before others (but this could be catered for).. All that would technically be necessary would be to turn most of the stuff into individual modules that could be added to a list for processing or similar... —Reedy10:54, 10 January 2009 (UTC)
Aw, I was looking for request like mine, but I couldnt find it (I suggest to archive requests more often ;)). What is supposed to run before others? I have no idead what it can be. And, well, I said it'll be tought, but - I repeat it - I believe in you, developers. Matma Rexpl.wiki talk11:11, 10 January 2009 (UTC)
Haha. People get annoyed if we archive non finished ones ;). I'll combine them later. However, i think this is more a v5+ request.. Hopefully where we'll have *maybe* changed to .NET v3... and can look at improving the gui with the new toys it gives us! —Reedy11:27, 10 January 2009 (UTC)
Move 'Make list' to new dialog. The model that I have is that an AWB list is like a file in many applications. I can create a new list, open an existing list, work with it, and/or save it. In Microsoft Word, I can create a new file, open an existing file, work with it, and/or save it. In Microsoft Word, I use items from the 'File' menu: 'New', 'Open', 'Save'. I suggest that AWB has similar items in the 'File' menu, although they might have to be appended with 'list' e.g. 'New list'.
Description: 'New list'. Produces a dialog. All the items in the current 'Make list' group would be moved into this dialog. The list itself would be duplicated so that it is in the 'New list' dialog and the main AWB interface as it is now. An advantage is that the main AWB interface could have a longer list due to the removal of six elements (the two fields for source and the two buttons).
Description: 'Open list' or simply 'Open'. Produces the usual dialog for opening a text file.
Description: 'Save list' or simply 'Save'. Produces the usual dialog for saving a text file .
I defined 'New list' but not 'Edit list'. An 'Edit list' function needs an 'Edit list' button beneath the list where the 'Filter' button is now.
'New list' and 'Edit list' should produce the same dialog. A working name for the dialog should be 'List'.
The 'List' dialog is the main place for editing lists but removal of articles should be possible in the main AWB interface.
The right-mouse menus throughout AWB could do with a review. The right-mouse menu in the current list has 'Filter', 'Save list', and 'Sort alphabetically'. If this design idea goes further, I suggest eliminating those three options.
Sometimes its useful to be able to tailor the messages you leave at user talk pages based on when that editor last edited. The tab would only need to display when working in the User: or User talk: namespace. I've drafted a possible patch that would do this. - Jarry1250(t, c)15:54, 8 June 2009 (UTC)
Added in revision
I wouldn't want to use a separate browser for this... Im wondering about combining the history and WLH here ones and having like radio buttons to select or something... —Reedy17:06, 8 June 2009 (UTC)
Yes, having one browser would be simpler. You could then disable the user contributions option for non-user (talk) space. It's not urgent, I can use my local version to do the few I'm interested in at the moment. - Jarry1250(t, c)17:17, 8 June 2009 (UTC)
Over on the right set of tabs, after History I could really benefit from an "actual" view of the current page. While the current Preview method works fine for articles, it's completely useless for categorizing images - which don't display in any diff/preview page view. I'm not sure how hard this would be (I'm aware that AWB briefly performs an "actual" load immediately after saving), but if possible it would save me from having to manually "Open in browser" every single image i look at (plus it would be a nice workaround for the lack of "Articles using this image" list). To reduce server strain it should operate like the History tab, only loading when the tab is active. -ΖαππερΝαππερBabelAlexandria20:22, 7 August 2008 (UTC)
Added in revision
There was a request for something like this using something that could potentially done internally.. —Reedy11:39, 10 August 2008 (UTC)
Search within 'Make module'. I sometimes want to change a small detail but can't find the text. Would it be possible to use 'Control-F' to search within the module? Lightmouse (talk) 17:46, 12 June 2008 (UTC)
Added in revision
Code already exists to do this (just not on the custom module).. Any suggestions how to do it designer wise? Do we need a Find form creating..? —Reedy19:28, 12 June 2008 (UTC)
Interesting predecessors include
the method that AWB already has below the 'Save' button. Perhaps you could just copy that or update it with ideas from the other predecessors below.
MS Word is interesting in that 'Find' and 'Replace' are just tabs combined into one dialog. So it is easy to start with Find and then decide to use Replace.
MS Notepad is very simple but can be irritating - it does not search the whole thing like MS Word does. You have to search forward from where you are, or then backward from where you are.
Firefox is instant search and has a very small field tucked away
You could add some menus to the 'Make module' page and then have the 'Find' menu item in one of them. I instinctively use 'Ctrl-F' to get 'Find' or 'Ctrl-H' to get 'Replace'. Lightmouse (talk) 20:09, 12 June 2008 (UTC)
I think a dialog like the MS Word dialog would be good. You could enable it on Ctrl-F even without menus for now. You could use the same dialog in both places. This would allow you to remove the field from below the Save button. Lightmouse (talk) 20:48, 12 June 2008 (UTC)
Dialogs are so passé. Firefox's find toolbar is what we should be aiming at, but safari has such cool effect (it dims all non highlighted words). Visual Studio has some cool options, but isn't as powerful as notepad++. I believe I had add Ctrl-F and a few other shortcuts in one of my uncommitted patches.—Dispenser04:44, 15 September 2008 (UTC)
You are right, Firefox is better, although MSWord integrates 'Find' and 'Replace' very well. It would be good if the AWB solution could do that. Lightmouse (talk) 12:50, 15 September 2008 (UTC)
What i dont get, is why the .NET framework doesnt have a standard FaR contorl that can be reused. I suppose, its probably worth trying to make up a decent one we can reuse throughout AWB... Hmmm —Reedy12:55, 15 September 2008 (UTC)
AWB is not an IDE for development of plug-ins or custom modules, therefore I don't think that this feature should be a high priority. Anybody who wishes to improve their custom module can use their own text editor or IDE to work on it. Rjwilmsi19:33, 27 January 2009 (UTC)
We're waiting on [11]. As said before, i am not HTML scraping (not worth the hassle). So this will be resolve when the bug gets sorted on the MW side —Reedy12:23, 8 April 2009 (UTC)
The ability to use it on a localhost wiki would be really nice. I would use it quite often. I know that it's designed for Wikipedia, but... Thanks, Genius101Guestbook12:18, 9 April 2009 (UTC)
Added in revision
I'm presuming it whinges about the domain name/similar? (Tbh, never tried a localhost server... Have you tried 127.0.0.1 also?) —Reedy19:01, 9 April 2009 (UTC)
I've tried both localhost and 127.0.0.1, and the error message it gives is: "An error occured while connecting to the server or loading project information from it. Please make sure your Internet connection works and that combination of project/language exist. Enter the URL in the format wiki.riteme.site/w (including the directory where index.php and api.pho reside)." Thanks, Genius101Guestbook20:25, 9 April 2009 (UTC)
AWB is probably the best tool already set up to handle this sorely needed ability. it already sits on a user's computer and can directly access the relevant pages and templates. the only problem i foresee is updating articles with the new name automatically would be unallowed for non-bot accounts (technically). -- ΖαππερΝαππερBabelAlexandria17:58, 7 June 2009 (UTC)
well i figured that since AWB is running on the client side, it could work around MW's constraints. The image and info would be stored temporarily on the user's hard drive. of course maybe this is better handled by a script like twinkle or friendly. --ΖαππερΝαππερBabelAlexandria23:29, 21 April 2011 (UTC)
Provide a separate checkbox for "Incorrect phrases" during Regex typo fixing
Fronds ofc does something like this.. Its whether you wanted the audit trail and it kept onwiki. —Reedy06:14, 14 August 2009 (UTC)
Well there's some disagreement as to whether phrases should be included with typo fixing. I'm leaning towards the no side, but some people like them. So I figured the best of both worlds could be included with a further check box. I would say suggest default to off. –xenotalk12:39, 14 August 2009 (UTC)
FYI, someone at the discussion was concerned that Fronds was in its infancy and thus wouldn't be as effective as it being built into the typo matrix. –xenotalk13:17, 17 August 2009 (UTC)
Rather than terminate the operation, AWB should just wait for, say, half an hour, when maxlag is exceeded 10 times in a row and give it another shot. –xenotalk21:32, 29 August 2009 (UTC)
I'm not sure if you guys killed the maxlag features with fire or whatever, but I rather liked being able to "let my bot loose" as it were and let maxlag act as a leash. So maybe having maxlag as an option to enable would be cool. –xenotalk15:00, 16 March 2010 (UTC)
When working on wikisource pages, especially in the Page: namespace, the page names are very long, and only vary in the last few characters (looooong filename.djvu/\d{1,3}). Ability to be able to expand the size of one of the list screens to see the whole pagename, or even to be able to have a scroll bar. This is necessary as pages don't sort well due to /1, /10, /100, /101 ... /199, /2, /20, /200, /201 ... and can be a selection of pages upon which to be worked. billinghurstsDrewth14:32, 16 January 2010 (UTC)
Added in revision
I was about to post a request regarding the same issue. In the list comparer, one can hover over a long page title in either input list and see the full page name on a mouseover. The Results lists do not have this feature; the full title of a page in one of these lists can therefore only be seen by saving the list to a text file. Can a mouseover text (or alternatively a scroll bar, as requested above) be added to the three "Results" lists in the list comparer? SiBr4 (talk)19:58, 25 November 2014 (UTC)
Also ctrl+c to the diff window. I don't know how many times I've highlighted some text in the diff window (it's easy because what I want is usually already highlighted in part by the differencing code so I don't have to try to find the same thing in the edit window), ctrl+c then go to the regex tester or wherever then ctrl+v the wrong stuff. Argh!
Current status Editor launches AWB and get informed that there is that their version is disabled and has to download new version. They are asked to press Yes to update, No to do it manually or Cancel. Pressing yes results AWBUpdater to launch. AWPUpdater asks (again) if editor wants to update with a Yes/No.
Suggestion First box changes to Automatic/Manually/Cancel. If "Automatic" AWBUpdater is called with a parameter and no new window pops-up. Update starts. -- Magioladitis (talk) 18:41, 10 March 2010 (UTC)
The idea would be that if it contain a template (or any specific phrase) like {{Main word|Liking}}, where I have specified it is {{Main word|X}} it looks for, where X can be any set of string, it could open up the page of "Liking" and then add whatever I wish it to add to that page, typicly a link in a specific manner that I have already pretold it
The ability to save and load AWB settings to/from a wiki file page would be helpful for publishing source code and collaborative development. –xenotalk04:11, 5 May 2010 (UTC)
Added in revision
Do you mean a physical [[File:]] - Of course, using the relative file location will let you do it, with MaxSems warning... —Reedy14:47, 7 May 2010 (UTC)
Well, the real reason is so that I don't have to email myself settings file back and forth between locations when I'm in the middle of a task. –xenotalk13:10, 4 June 2010 (UTC)
This or a similar feature would be very useful. For example me and some other users maintain a list of regular expressions at Commons:File description page regular expressions (XML) and it would be really cool if the finds and replaces in it could automatically be loaded by all our bots at startup, kind of like what's done for typos. - EdoDodotalk09:52, 3 September 2010 (UTC)
Looking to know whether there was an efficiency to be able to filter (pre-parse) pages on their last edited date, eg. pre-parse can skip files edited before DATE-A
Background: On a project at enWS we are looking to update the style used on transcriptions, and after we have updated pages, we would only want to run the search on pages edited after the last cleanup. It is a link style cleanup so there is no ready other means to filter. billinghurstsDrewth14:57, 13 May 2010 (UTC)
Added in revision
Trivial, yup. PageInfo gives us the last edited timestamp. If someone can suggest a good place to add this, and what to do about date formats etc, I'll get this added. —Reedy06:31, 24 June 2010 (UTC)
In the 'Filter' there is a tick box option called 'Sort alphabetically'. I suggest that we turn that into three option buttons labelled: 'Do not sort', 'Sort alphabetically', and 'Sort by date of last edit'.
It would also be useful to have date fields labelled 'Remove articles with last edit after' and 'Remove articles with last edit before'
At the same time, I'd review the 'Keep alphabetized' option in the 'List menu'. I'm not sure what it does and perhaps it could be removed or have another option relating to the new chronological sort.
AWB should support using more than one typo list on third-party wikis. Being able to specify from within AWB which lists are used (e.g. separate checkboxes for individual lists, and one "all lists" checkbox) may also be useful, but isn't central to this request. 「ダイノガイ 千？！」? · Talk⇒Dinoguy100021:02, 7 June 2010 (UTC)
When making a list, it would be neat if by adding a code, like "&from=fe", AWB would start the list from the letters "fe". Fore example "Category:Living people&from=fe" starts the list from "fe" then "fc" etc. This can help in very large lists. Sole Soul (talk) 13:25, 28 June 2010 (UTC)
Added in revision
Prevent the article from being saved if a condition exists
An option that would prevent an article from being saved if a rule is violated. So if the Save button is pressed and a condition exists that violates a predefined rule then the article is not saved and an error dialog is displayed (or the article is skipped and the occurrence is logged).
Added in revision
This could be a plug-in I guess. It would create a additional rules dialog analogous to advanced settings. The dialog would allow the user to enter some number of rules (regular expressions) paired with an error messages. The text would be evaluated when the Save button is pressed. If a condition exists in the text which violates one of the rules then:
The save is aborted.
The first instance of violating text is highlighted as with the highlight errors option in the Options Menu.
An error dialog is displayed containing the paired error message.
If in bot mode, or at the users discretion, the page is skipped and the error message entered in the skipped log.
This process is repeated until the text does not violate any of the rules.
Perhaps a check box could be added to the rules dialog that would prevent saving when an error exists of the type highlighted by the existing highlight errors option. This might also include the option of skipping and logging the error type.
The highlight color should probably be different than the red used by the highlight errors option. Some striking color that could not be confused with red but does not obscure the text.
The motivation is that it is possible to detect situations that are ambiguous and need human attention. In the past I've created rules that inserted brackets that created an open bracket error. On re-parsing the bracket errors where highlighted and allowed me to handle the case manually. It's a laborious process that is prone to failure.
The interface might not have to change much. Maybe add a rule type to advanced settings that implements the functionality. –droll[chat]03:29, 15 July 2010 (UTC)
Would it be possible to improve the "wiki text search" facility by adding an option to search by namespace? When searching for phrases such as "July 3", at least 10% of the 1000 pages selected tend to be non-mainspace, for example. BencherliteTalk08:37, 28 July 2010 (UTC)
I know that feature exists, but If I'm looking for all the particular instances of phrase in article space, for instance, wouldn't it be easier to specify that I want only article space instances to start with? Otherwise I hit the 1,000 limit with irrelevant namespaces. E.g. I search for "March 3" and after filtering only 824 - not far off 20% - of the hits are in article space. BencherliteTalk13:08, 26 August 2013 (UTC)
Handle the initial directory for "Open settings" and "Save settings as". Either allow users to specify a default directory, or start in AWB's directory. MANdARAX•XAЯAbИAM21:41, 30 July 2010 (UTC)
Added in revision
I use many different settings files, so this would be very useful to save a little directory navigation. (It would be especially helpful when running more than one instance of AWB simultaneously, as "Recent settings" only works initially for the first. This specific issue may only be relevant with snapshots which are debug builds; I'm currently using SVN 6872.) MANdARAX•XAЯAbИAM21:41, 30 July 2010 (UTC)
I concur with this request. Every time I go to open or save settings, it doesn't remember what directory I last pulled from. This is unlike most other Windows apps I use. Stevie is the man!Talk • Work15:40, 30 August 2014 (UTC)
It would be very helpful to be able to select and undo any given change as shown in the diff window; this would be a big time saver over manually undoing a change, and would be less prone to manual error. (Thanks for great work!) David Hollman (Talk) 07:35, 15 September 2010 (UTC)
...and as long as you realize that double-clicking on any change in the diff window (or making any change in the edit box) wipes out all the automated messages that get added to the edit summary (e.g. typo fixes, replacements, addition/removal of templates), which is a bummer. GoingBatty (talk) 01:22, 16 September 2010 (UTC)
It is line by line, but in the code, one line usually contains the whole paragraph. Do you mean sentence by sentence? Svick (talk) 18:31, 18 September 2010 (UTC)
It doesn't just show the paragraph containing the diff. It also puts paragraphs either side that have no diff. These provide some context but use up a lot of space. Can these non-diff paragraphs be reduced in size?
It might also be helpful to have an "undo like" capability that would allow the user to deselect all changes like on an article..(for example if AWB suggested a bunch of typo changes the user could select the undo like option to uncheck all the top changes). --Kumioko (talk) 17:25, 25 September 2010 (UTC)
Regex tester is a very useful tool, but it is inconvenient that one should type all expressions again and again every time one opens the tester window during thorough debugging of a complex regexp. It would be great if the last content of the window was restored when it is open. --Vladimir Ivanov (talk) 12:18, 3 October 2010 (UTC)
Added in revision
Request of CheckPage feature to make it more easily maintainable (generall access option)
AWB has WP:AWB/CP page for access control. I think this is good feature. But this page needs few bureaucratic works for maintainance (user needs to write request, admin must check applicant's contributions and judge it. And after it, adding name to the page, or decline request.) Although maintainance needs few works, "no-CheckPage" could cause troubles. So if CheckPage supports general option, like Wikipedia:Huggle/Config#Access_control or WP:Twinkle(User:AzaToth/morebits.js), it would be very helpful for many wikis. Especially wikis in which admins are not so many. For example, JAWP where I'm from has only about 60 admins[12]. Active admins are more fewer than it. Active admin who can use AWB is almost nearly zero. If some general options (e.g. "number of edit", "account-age", "all bots are yes") are available, I think it is very helpful. Thank you. --Was a bee (talk) 06:38, 10 October 2010 (UTC)
AWB should allow sending a list to Special:Export, and automatically loading the resultant xml file in the database scanner. –xenotalk03:20, 24 October 2010 (UTC)
When you click "save", the edit box becomes grayed out. If you get a timeout error (or whatever causes AWB to restart), AWB just re-processes the article from scratch automatically, and you lose all your manual changes. It would be IMMENSELY useful if, when you click "stop" to interrupt the re-processing, that the edit-window would become editable again so you could keep your changes (via crtl+c), or that AWB would prompt you before reprocessing the page, giving you the option of keeping your changes rather than start from scratch.
This happens to me too and I find it very discouraging. I spend considerable time on an article only to have the changes disappear. It is not always related a long period of inactivity on my part. Is the edit box content saved anywhere on my computer where it can be retrieved after restart? –droll[chat]06:24, 30 June 2011 (UTC)
Have Filter window remember from one time to the next if it was set to "Intersection"
Have the "Filter" window remember from one time to the next if it should be set to "Symetric difference" or to "Intersection" - usually when I've used "Intersection", and go to there again in the same session, I want to use "Intersection" again, and I imagine that so would most other AWB users. עוד מישהוOd Mishehu
Please add option 'Open diff in browser' for log. The options 'Open page in browser' and 'Open history in browser' are very useful. I'd like to be able to see one or more items in the log and select 'Open diff in browser'. The use case is: check multiple edits. Thanks. Lightmouse (talk) 13:27, 14 April 2011 (UTC)
Please add option to increase visible size of log. The log only shows 6 lines at once, sometimes I'd like to be able to see many more lines. I know that I can save the log and review it that way, but Would it be possible to have an option that enlarges the log? It currently shows 'Successfully saved' and 'Skipped' by default. Perhaps you might be able to allow the user to choose which one to show and that would allow more vertical space for possibly 30 lines. Lightmouse (talk) 13:37, 14 April 2011 (UTC)
Not infrequently, I use a toolserver tool to create a list of all pages where certain edits would probably be needed using AWB (category names by patern using this tool, category scans using this tool), where creating the list inside AWB is either impossible (such as my first example) or difficult (the second example - I still would need to filter out all non-mainspace, for example, but that would be simple AWB filtering). I think that we should be able to make a list from all incoming links from an external web page, given by URL. That would make these tasks much easier. עוד מישהוOd Mishehu10:38, 5 September 2011 (UTC)
I'm using AWB on frwiki since a few years. I’d like to be able to use some of its functions on the others wikipedias without having to log-in. The pre-parse mode doesn't modify the articles so it shouldn't be access protected. It would allow to use all the skip tab functions to filter (regexp) a list of articles with what their text contains and not only their titles.
In fact, I'm trying to find out what articles of a category exist on other wikis and doesn't on frwiki (ie. what could be traduced). And I want to do it for a lot of languages. The two solutions I have for the moment is either download all dumps then use the database scanner, either request an AWB authorization on each WP to access the pre-parse mode and use it as a little scanner. Downloading full dumps and scanning them to check a few hundred articles is absurd. Asking an authorization to use AWB on all WP could be very long, especially that I don't ask it to edit there, only to read the wikis (ambiguous request, and biaised for me).
I can't use the list comparer either because I don't know the traduction of the titles of the articles.
If there's another way/tool to do what I want, I'd be glad to know.
If the way users access are handle in AWB could be revised, I'd be glad to continue using it to work!—A2 (talk) 00:23, 26 February 2012 (UTC)
I also requested this some time ago as well. If I remember correctly it can't be done because of the Wikipedia API. In order for AWB to connect to WP it needs to use a login through the API. I can't remember exactly but I think thats how it was explained to me a while back. --Kumioko (talk) 04:03, 26 February 2012 (UTC)
So perhaps a slightly revised feature request would solve this: to allow pre-parse mode (only) if logged in with an account that has no AWB permission? -- John of Reading (talk) 08:13, 22 March 2012 (UTC)
Minimialise the config stuff on bottom half of screen
What matters for typos checking is the top view pane, have a toggle to reduce the bottom part down to 'the current article name', the existing four buttons, 'start', 'stop', 'skip', 'save' with the edit summary beside it, plus a way to re-expand to normal. Regards, Sun Creator(talk)23:50, 4 September 2012 (UTC)
Added in revision
Built in AWB functions
Edit summary
Provide ability for edit summary to change dynamically based on what custom Find & Replace changes are made
Allow a field in an Advanced "Find and replace" entry to modify the edit summary. The edit summary would allow for numerous variables and these would be filled in by the Find/replace entry. –xenotalk14:12, 15 December 2009 (UTC)
Added in revision
Example: "(Bot) Tagging for WP:FOO - Inherited $1-class assessment from other projects"
Only in the normal Find and replace though and in any case this wouldn't allow for what I've shown above as an example. –xenotalk16:50, 15 December 2009 (UTC)
This would be very useful to me too. Imagine you have a list of pages (from a cat or something) and you want to do task A to pages matching condition A and Task B to pages matching condition B. At the moment you can do this easily using advance find and replace but you have to use one edit summary for all the edits, or continually change it. Jamesmcmahon0 (talk) 08:47, 13 March 2015 (UTC)
My longest edit summary is something like WP:GenFixes, refcleanup, parse authN/edN, cmn/reflist|2/3->30/20em, {{LMR☉⊕J/up}}. I (and others) would find it useful to automatically select only the relevant changes. Also, the edit summary dropdown is so short that it makes selecting the most appropriate (long) edit summary tedious. ~Tom.Reding (talk ⋅contribs ⋅dgaf) 13:34, 26 March 2015 (UTC)
Dynamically change edit summary based on its length
Use normal, prettier edit summaries when space allows, and use space-saving measures when the summary is so long that it would be truncated. MANdARAX • XAЯAbИAM22:05, 17 December 2009 (UTC)
Added in revision
One space-saving measure would be to remove the spaces surrounding the arrow in typo replacements: "againt→against" instead of "againt → against".
Some (most?) people prefer the visual appeal of "using [[Project:AWB|AWB]]" to "using [[Project:AWB]]". With this feature, the prettier version could be used most of the time while not wasting those valuable characters when necessary, and the space-saving version could be even further shortened by removing "using". Make it simply " ([[Project:AWB]])".
I already do something similar. When I'm fixing typos, my normal edit summary is [[WP:AWB/T|Typo patrol]]. When there are so many typos that the list is truncated, I clear my edit summary to allow more of the list to show. Note a bug in my example diff: the "[[Project:AWB]]" got lopped off. MANdARAX • XAЯAbИAM22:05, 17 December 2009 (UTC)
If I understand how edit summaries work currently in AWB, if I change something either by directly editing the text of the article or by double clicking a line in diff, AWB uses just the default edit summary, not specific changes, because the user might have undid them. I think it would be better, if the edit summary would be smarter, so if AWB suggest two fixes and I undo one of them, the edit summary would still contain the other. Svick (talk) 23:35, 21 December 2009 (UTC)
Sometimes when I have undone a change, I forget to revise the edit summary accordingly. It sure would be nice if it were updated automatically. It would probably be easier to implement if these automatic updates would work only if the user hasn't already altered the edit summary for a specific list item. Stevie is the man!Talk • Work15:48, 30 August 2014 (UTC)
Edit summary when I use find and replace on ar.wiki
I think this would be worth doing for all the wiki's where we have the different namesapces. It would only require a few code changes, ie instead f Replaced, use Variables.Replace (or whatever), and then have the local word for each... Reedy Boy08:46, 10 July 2007 (UTC)
My goal isn't to change the existing rules, but to have the edit summary that is automatically generated better reflect the changes that are made. Took a quick look at Custom Modules, but didn't see an example of one that with logic such as: if (something) is added, then include (some text) in the edit summary. GoingBatty (talk) 17:54, 30 November 2010 (UTC)
Use something like this to begin:
Summary = "StartSummary"
and call a function inside an if..then statement adding:
Summary += MoreSummary + @", ";
I guess Rjw can give you more specific instructions.
The first one is good because adding persondata is an improvement and not a fix. The others go a bit off our philosophy, because generating lentghy edit summaries can be time consuming and endless. -- Magioladitis (talk) 18:11, 30 November 2010 (UTC)
Automatically change edit summary based on rule applied
Allow a new field for every rule defined in Advanced "Find and replace" with the description for the "edit summary" if that rule is applied. So edit summary can automatically change based on the rule actually applied.
Added in revision
I think it is not hard to do, but very useful (at least for me). For example if changes are made using rule "1" the edit summary is "Bot - links to years removed", if changes are made using rule "2" it is "Bot - original title removed from the table of the movie". So my bot has to scan the "Category Film" just once, and it can do automatically more different replacements.
Thank you. User:FRacco --BRacco (bot) (talk) 17:10, 5 April 2014 (UTC)
Hi! I have set up various find/replace rules. One of them is to look for instances of "passed away" and change these (with my oversight) to "died", per WP:EUPHEMISM. What would be useful is if there was another field in the Find & Replace window where I could add "per WP:EUPHEMISM" or any other comment like "straight quotes preferred", but to have these notes appended to the edit summary, so that I can better communicate my rationale. (I thought that was the purpose of the comment parameter, but it doesn't append.) Obviously the edit summary field is limited, so we wouldn't need multiple mentions of the same policy. Such an edit summary might look like:
Clean up, replaced: “ → " (18), ” → " (19), ’ → ' (34), ‘ → ' (3) straight quotes preferred. passed away → died, made love → had sex per WP:EUPHEMISM, typo(s) fixed: forthe → for the using AWB
Feature for an external process to designate an edit summary. Currently, edit summaries are either globally defined, or manually modified. Envision the edit summary passed as a filename similar to the "Input/Output filename" field in the Tools->External processing menu. GreenC20:14, 27 January 2015 (UTC)
Added in revision
@Green Cardamom: I hope you're also aware of the Options->Default Edit Summaries functionality, which allows you to have multiple edit summaries and choose the appropriate one before saving each edit. GoingBatty (talk) 02:55, 28 January 2015 (UTC)
Yes thanks. I'm using an external script (Tools->External processing) and would like the script to choose the edit summary. So I don't have to manually. This could be simply accomplished by adding a second "Input/Output filename" field in the Tools->External processing menu for the edit summary filename. -- GreenC13:17, 28 January 2015 (UTC)
Solution using custom module
The solution I'm using for years now is to have custom C# snippet (Tools/Make module) to call external filters. Here is an example.. It solves multiple common external filters issues in AWB:
provides support for summary file
provides a command line template, to allow sending extra arguments to external script. Most useful are: article title (URL escaped, to allow safe transfer of UTF strings) and article namespace.
allows for custom control of child process creation details (e.g. minimized or no window at all)
That way, any script/languague/framework with CLI interface can be used to write a powerful external filter. Personally, I'm using PHP scripts most often, which allows for MW API calls, caching, functional regex replacements, transformations requiring parsing of wikimarkup and so on.
The only issue I know is, that System.Web.HttpUtility.UrlEncode() method requires System.Web assembly to be loaded. Which fortunately is the case in AWB, after logging into some wiki.
Maybe this kind of solution can be integrated directly into AWB, it's quite trivial, yet considerably more powerful than current external processing support. --Teslaton (talk) 14:02, 1 April 2015 (UTC)
1) The {{saved book}} and these four parameters (in this specific order) should always be present at the top of the book. There can be other parameters, but they should all be after |cover-color=.
{{saved book
|title=
|subtitle=
|cover-image=
|cover-color=
}}
2) If a level 2 header isn't present, add one (=={{SUBPAGENAME}}==)
3) The |title= and |subtitle= of {{saved book}}, if empty, should be filled with the level 2 and 3 headers respectively (==Title== and ===Subtitle===)
4) :''[[Foobar]]'' and :"[[Foobar]]" should be replaced with :[[Foobar|''Foobar'']] and :[[Foobar|"Foobar"]], respectively. Likewise for
:''[[Foobar|Barfoo]]'' and :"[[Foobar|Barfoo]]" which should be replaced with :[[Foobar|''Barfoo'']] and :[[Foobar|"Barfoo"]], respectively.
5) As in articles, :[[Foobar|Foobar]] should be replace with :[[Foobar]]. However, unlike in articles, :[[Foobar (barfoo)|Foobar (barfoo)]] should not be replaced by :[[Foobar (barfoo)]]
6) Articles found in Category:All disambiguation pages should be tagged with {{dn}}, except those with an explicit (disambiguation) in the title, such as :[[Foobar (disambiguation)]]
AWB currently adds {{WikiProjectBannerShell}} if 3 or more WikiProject templates are found, but only for WikiProject templates named "WikiProject foo" and not with their redirects. Could the Template redirect functionality be expanded to fix the redirects before adding the BannerShell? (e.g. something like this edit.) Thanks! GoingBatty (talk) 23:54, 11 July 2012 (UTC)
BTW I have already written written code for most of the Wikiprojects and redirects. If the logic for the AWB/TR page were updated to include talk pages or if another separate page were created for talk pages, I think it would be very beneficial and would be easy to convert the code I wrote over to the new format.Kumioko (talk) 13:38, 18 July 2012 (UTC)
Currently AWB announces that it is skipping an article because it contains PUA characters and can't be edited. There are two problems with this: (1) in pre-parse mode, the article is simply skipped without warning, and (2) in an article like Kabardian language, I'm not able to locate the PUA character that's the problem. Could the article not be skipped in pre-parse mode, so that it can be reviewed manually (I suspect this will result in weeding out more PUA characters), and could the guilty character be displayed in manual edit mode, or at least the line or paragraph highlighted, so we can either replace it with a supported character or format it appropriately for AWB? (E.g. {{PUA|}} for .) — kwami (talk) 00:04, 20 February 2013 (UTC)
In the alert box, list categories that do not exist when the article/etc is first brought up to edit and also after the 'diff' button is pressed or when you preview the article edits in any way. Currently, there is no warning at all that something was already wrong with a category in the article when the article is first presented to you for editing or after you have made any manual edits. This results in unnecessary and undesired category errors in the saved articles. Hmains (talk) 16:59, 4 July 2012 (UTC)
Added in revision
Keep "delink selection" button available even when no multiple wiki-links
In the "Alerts" area, could the "delink selection" be made available when there aren't any multiple wiki-links? I use the "Find" feature with regexes to find things that I'd like to unlink, and being able to use the delink button would be very helpful. --Auntof6 (talk) 01:18, 28 May 2010 (UTC)
FYI, using a regex search of the March 4 database dump, I found 2,292 instances of \[\[Category:[\w\s]+stubs\]\] in mainspace (excluding comments). GoingBatty (talk) 00:01, 13 March 2015 (UTC)
Alert for interwikis (prev. Remove interwiki functionality)
It probably should be removed, but when is the question. Wikidata isn't being pushed to all the other different language Wikipedias until sometime in March. It's going to take at least several months before the bots clean everything up. After that, the bots need to start automatically adding new articles into Wikidata. As long as no harm is being done, probably keep the code around for awhile... better safe than sorry. Bgwhite (talk) 20:43, 28 February 2013 (UTC)
Another thought: would it be worth adding an alert when an article does have interwikis? An editor could choose whether to try to clean it up. --Auntof6 (talk) 05:41, 28 March 2013 (UTC)
We have a winner. An extremely excellent idea. FYI... I think the bots are done removing interwiki links from articles and have moved on to other spaces. I see one doing categories right now. For anything left in article space, the bot either had problems moving the links (two links going to same language for example) or the wikidata entry hasn't been made in Wikidata. Phase 2 of Wikidata went live yesterday to 11 wikipedias. This should cause the demise of Persondata and make life a living hell for infoboxes. Bgwhite (talk) 06:54, 28 March 2013 (UTC)
1) Add Multiple wiki-links to Options → Preferences → Alerts, providing the ability to hide it; it's sometimes very distracting. 2) Provide an option to limit the list to those with N or more occurrences (perhaps with N defaulting to 3); it's generally the higher numbers that merit the effort to fix 3) Increase the width of the displayed list; having to constantly scroll left and right is a major disincentive to action.
These categories should not be hidden categories, maintance categories nor stub categories. Any rule when exactly the article is considered "categorised"? -- Magioladitis (talk) 02:59, 12 September 2010 (UTC)
If it's in at least one true content category, it's considered categorized (though {{cat improve}} may still be appropriate, such as if an article's only content categories are birth/death years.) You're correct that hidden/maintenance/stub categories don't count toward the article being considered categorized, if they're the only categories present. Bearcat (talk) 03:56, 12 September 2010 (UTC)
Add a window to allow editors to choose which part of autotagger they want to enable/disable. A window similar to GF skip options with the following options:
Add/remove the following tags if applicable:
☑ stubs ☑ expansion (remove only)
☐ uncategorised
☐ orphan
☑ deadend
☑ wikify
☐ ibid
Notes:
AWB should not add expansion tags, only remove if article is a stub thus the "(remove only)"
The above list could be expanded with two columns Remove / Append
Support I would like to ask for the enhanced autotagger too. I have been using the AWB on Wikia, and I would like to be able to use the autotagger for the placing/removing of the stub tag only, because the tags such as uncategorised, orphan or deadend aren't mostly used on Wikia. -- Spike144 (talk) 12:25, 21 November 2013 (UTC)
Support Using AWB at Wikia with the auto tag option checked is an all or nothing option. It appears however that many of the options are specific to Wikipedia such as wanting to add {{orphan}} to seemingly every page which makes the feature pretty much useless. I see that there is a tab under preferences that allows the alerts to be selected if desired and I wanted to ask if the same thing could be done to the Auto tagger function so that projects can decide which ones they want to use since it may apply to some (although most don't use it). Reguyla (talk) 00:29, 21 January 2014 (UTC)
Change uncat with Cat improve, if page contains only Living people, births, deaths categories
I would first like to see the template documentation provide a complete list of what is and isn't permitted HTML tagging in Wikipedia. Currently it's unclear. Rjwilmsi10:14, 10 January 2011 (UTC)
It is not required that a ref have a cite template, & quite a number of refs do not, from both new and experienced users. . I , for example, never use it unless there are already refs in the article formatted with it. DGG ( talk ) 16:23, 15 April 2011 (UTC)
Could AWB remove {{Stub}} when a more fine-grained stub tag already exists on the page. In these cases the {{stub}} tag is redundant. I think a Regex search of "\-stub\}\}" should find any 'fine grained' stub tag on the page. Jamesmcmahon0 (talk) 12:50, 1 August 2013 (UTC)
Added in revision
Make List
Add Special:PagesWithProp to "Special Pages" dropdown list.
It appears that AWB does not have a way to easily pull in pages from the Maintenance reports lists under Special pages. I would like to request a new Make list option for the Maintenance reports available under the Special pages link. If some of these cannot be done I completely understand but if possible could the following be added at least?
With CirrusSearch now enabled and running fairly well, it is now possible to do regex searches. For example, To search for <center>, you would use insource:/\<center\>/. It should be able to use search to create lists. There are currently limitations on the number of regex searches that can be performed simultaneously. More info at mw:Help:CirrusSearch. --Gadget850talk17:00, 28 October 2014 (UTC)
If I do an on-wiki search for insource:/\<center\>\<gallery/i I get 38 pages (seems like it always reports one more than actual). If I do the same search in AWB I get 0 pages. If I search wiki or AWB for "gallery" insource:/\<center\>\<gallery/i I get 17 pages. If I search for "e" insource:/\<center\>\<gallery/i I get 22 pages (e is the most common letter in the English alphabet). If I search for "a" insource:/\<center\>\<gallery/i I get 37 pages (a is the second most common letter in the English alphabet). I don't understand why AWB needs this prefixed search. There were issues with insource searches when CS was first introduced, but they seem to have been resolved. --Gadget850talk11:39, 22 January 2015 (UTC)
On exit, check for changes to settings and query whether user wants to save the changes. When I exit most applications, they check if changes have been made. If no changes have been made, they exit immediately. If changes have been made, they ask if I want to save the changes. Look at how MS Word behaves. AWB does not do this. I know that AWB cannot test for everything but I would like it to be able to save my javascript changes as a first priority and my skip options etc as a second priority. Lightmouse (talk) 22:05, 13 May 2008 (UTC)
Added in revision
This is a very useful proposal, however checking for changes the right way (e.g. by serialising settings and comparing against saved ones) is a heavy operation. Needs more thinking. MaxSem(Han shot first!)08:23, 16 September 2008 (UTC)
Thanks. Perhaps we could divide the request into several pieces. For example, changes to the 'Make module' could be one problem piece to solve and changes to checkbox options in the tabs could be another problem piece. I would regard it as a useful advance if you could solve either of those problem pieces. Lightmouse (talk) 08:39, 16 September 2008 (UTC)
Hold on. I often make incremental adjustments to "Find & Replace" while processing a list of articles, and don't want to save new settings until they are tested on a few articles, so I would not have 'auto-save' turned on in that situation. Before exiting AWB or before loading a new settings file, I would like to be reminded that I have unsaved "Find & Replace" changes to the current settings file. This part should not be difficult to implement, especially if the request for "Unwanted saving of Find & Replace changes" is taken care of by dropping the "Done" button and adding "Apply" and "Cancel" buttons. If the "Apply" button is pressed, just make a note that the settings file has changed and may need to be saved on exit; the heavy operation of comparing settings against saved ones would not be necessary. A similar technique could be used to mark changes to skip options, etc., and all the checks would not have to be added at once. Christhe spelleryack14:53, 25 June 2011 (UTC)
It would be very great if AWB will allow us to move many pages the same way (e.g.: moving Goldfinger article to Goldfinger (film)), there is yet 2 pywikipedia scripts (movepages.py and pagerename.py), but movepages.py does not append words to the title and pagerename.py does not convert to UNICODE, especially when using special characters (as French é, è, à... and also Arabic letters). So, please, such a feature on AWB is highly recommanded, thank you. --DrFO.Jr.Tn (talk) 20:05, 15 June 2008 (UTC)
The needed actions are :
Prepend (to titles), i.e.: prepending a word to titles (in a list of articles), e.g.: USS Michigan (for the USS ships). (this is just an example)
Append (to titles), i.e.: appending a word to titles (in a list of articles), e.g.: St. Louis, Michigan. (this is just an example)
Find and replace (in titles), i.e.: finding a word in the titles (in a list of articles) and replacing it by another, e.g.: Goldfinger (cinema) to Goldinger (film). (this is just an example)
Delete (from titles), i.e.: deleting words from the titles (in a list of articles), e.g.: State of North Carolina to North Carolina. (this is just an example)--DrFO.Jr.Tn (talk) 20:18, 16 June 2008 (UTC)
Allow the user or hardcode an agreed distance where by Multiple links are shown
If we had:
Wikipedia:AutoWikiBrowser is a semi-automated Mediawiki editor for Microsoft Windows 2000/XP/Vista designed to make tedious repetitive tasks quicker and easier. It is essentially a browser that automatically opens up a new page when the last is saved. When set to do so, it suggests some changes (typically formatting) that are generally meant to be incidental to the main change.
At present, Wikipedia:AutoWikiBrowser can create a list of pages from single or multiple categories, "what links here", the wiki links on a page, a text file, a Google search, a user's watchlist, or a user's contributions.
Distance =2 lines . This would return false
Distance =10 lines . Would return true
Distance =1 paragraph would return true Gnevin (talk) 14:14, 6 August 2008 (UTC)
Lines depend on resolution... Hm... an idea would that in the Multiple wikilinks box to have not only number of occurrences but minimum distance in terms of words too. -- Magioladitis (talk) 18:25, 8 March 2010 (UTC)
50% of the WMF traffic is not English. If this software is useful for other languages, it should be possible to use the software in this other language.
As an aside. We have thought about it, and it would be a good idea. It just would require a lot of reworking of the code, and the interface changing to cater for the larger text aswell... Getting translators would be easy enough, i suspect... —Reedy11:13, 1 December 2008 (UTC)
We now provide edit summaries in many languages and the set of rules are better localised for the various projects. I call for help in order to expand the localised edit settings to more languages. -- Magioladitis (talk) 18:52, 15 August 2013 (UTC)
Currently, the disambiguation dialog box only has a short excerpt of the text. This is often not enough to determine context for the correct replacement, and you have to open the page in an external browser. It would be nice if you could drop a webbrowser control in the blank space in the bottom of the dialog so we can read the article without opening other windows. Thanks for a great tool Phil153 (talk) 05:48, 16 December 2008 (UTC)
This feature would be very useful for bots to filter out articles that it has already edited so it can avoid editing an article for a second time. For example, in some scenarios the bot makes a false positive edit, then a user reverts the bot, then the bot makes the same edit again. I would like to be able to see if the history contains the bot user name but it would be useful to test for any string in the history. I imagine the interface as similar to the 'article contains', 'article does not contain' code. Is it possible? Lightmouse (talk) 07:38, 19 April 2009 (UTC)
It probably depends on whether the mediawiki API has a feature whereby it will tell us how many times a given user (or bot user) has edited a page. If that's available then we could certainly have an option "skip if I've already edited this page". Searching a history for a string doesn't sound feasible: the history of some pages is 50 edits per day, so AWB could easily spend several minutes working through the last two months of a page history. Rjwilmsi22:27, 19 April 2009 (UTC)
Also, you'd probably want some time constraint on it. You may have not touched it for 2 years, so missing out on those edits would therefore be bad... It may be possible to query the API (ie have to ask Roan nicely to do it), to be able to specify a date, and a username and return whether the user account has touched that page since that date... *May*. As per prior mentions, i'm NOT html scraping. I'll stick a bugzilla request on and see (or ask Roan on IRC).. —Reedy11:25, 20 April 2009 (UTC)
omitting rvstart (prop=revision&rvend=timestamp&rvuser=username&rvlimit=1), will give all (if any) edits by that user to the page since the date. It's possible —Reedy11:28, 20 April 2009 (UTC)
It would be very useful to be able to restrict searches to individual templates. For example, suppose I'm using advanced search for cleaning up an IPA template, including reformatting stress and length marks <ˈ>, <ˌ>, <ː> for apostrophes, commas, and colons. Now, if the IPA template is embedded in an infobox, as it is with the planet infobox, or even if there are both an IPA template and an infobox on the same page, then every apostrophe, comma, and colon in the infobox will also be replaced and need to be deselected, making maintenance very time consuming. It would be nice to only replace things in the template I'm searching for. kwami (talk) 19:26, 1 July 2009 (UTC)
Interestingly, I'm currently working on a plugin which will focus on doing things in specific places within templates, with a UI to help define the template, parameter and values which need attention. Might this be of interest? ClickRick (talk) 23:37, 1 July 2009 (UTC)
Yes, that looks like it might do it. But it would still be nice to have a simple check box to restrict the find-replace to the template that matches the if/not conditions. kwami (talk) 08:53, 2 July 2009 (UTC)
If I want to make changes in a specific template I am doing the following: Find and replace -> Advanced settings -> New Rule -> If contains -> Infobox Example and then I am creating subrules. -- Magioladitis (talk) 08:29, 2 July 2009 (UTC)
But if the page contains a template that contains your search string, then the substitution will be made in all templates on the page, not just the one that contains the string. So if I change apostrophes, I mess up formatting in all of the info boxes. kwami (talk) 08:53, 2 July 2009 (UTC)
Try a regex like \{\{IPA\|([^\}]*)\ˈ([^\}]*)\}\} to $1[whateveryou'rereplacingˈwith]$2. I'm sure there are much more elegant ways. --NE221:07, 6 July 2009 (UTC)
That works! Thanks. A feature option would still be nice, but with this I can at least take care of some of the more grievous problems. kwami (talk) 07:56, 7 September 2009 (UTC)
Namespace-insensitive list sorting or add talk page after article
Add an option to sort a list while disregarding the namespace, so that a mainspace article would appear immediately before the same title in other namespaces. For example, instead of sorting as Article1, Article2, Talk:Article1, Talk:Article2, it would be Article1, Talk:Article1, Article2, Talk:Article2. This could be in the right click menu and/or the list filter. MANdARAX • XAЯAbИAM00:18, 3 November 2009 (UTC)
Another possibility would be to have an "Add talk page" option to take a list of articles and add the corresponding talk page right after each article. MANdARAX • XAЯAbИAM20:22, 14 November 2009 (UTC)
Added in revision
I'm working on sets of pages where I fix/add DEFAULTSORT and also work on the corresponding talk page, adding WPBiography if missing and fixing/adding the listas and living parameters. It would be very convenient to have the pages sorted together so I can manually apply any nonstandard DEFAULTSORT which I find in the article to the listas, and also determine the living status from the article to fill in for WPBiography. Or, if I find that the article's not a bio, I can immediately skip the talk page. MANdARAX • XAЯAbИAM00:18, 3 November 2009 (UTC)
This is such a useful feature for me that I wrote a simple external program to do it and I paste the list into AWB. But it would still be very convenient for AWB to do this. MANdARAX • XAЯAbИAM20:22, 14 November 2009 (UTC)
It would be useful to have a language setting for AWB so you could tell it whether the page is written in British English, American English etc so it could fix, for example, "color" in an article about England as "color" is not a word in British English. This way it could standardise the spelling of an article without having to manually (and tediously) search through the hyperthetical article about England that an American user has edited to find all of his spelling mistakes. McLerristarr (Mclay1) (talk) 03:36, 26 July 2010 (UTC)
In order to standardise redirects, AWB should do things like this. Capitalise 'REDIRECT', have a space before the target and another space before a redirect template.McLerristarr / Mclay110:52, 11 September 2010 (UTC)
Not all changes are visible, such as avoiding redirects to {{Citation needed}}. There are many edits that AWB does that qualify for rule 4 if only done by themselves. However, if all these changes were made to a redirect page, as well as edits to categorisation, interwiki links and other existing changes, it could be beneficial. Removing pointless extra text in a redirect page is useful, as it is completely unnecessary and could possibly be vandalism. McLerristarr / Mclay115:53, 11 September 2010 (UTC)
For some reason AWB does not correct dates when following "the". e.g. "the 1st of January" will not correct to "the 1 January". "The" cannot be removed due to phrases like "the 1 January 2000 elections" but I see no reason why the date should not be corrected. "The" could be removed if the date is followed by a punctuation mark. McLerristarr | Mclay122:06, 2 November 2010 (UTC)
Currently, -- corrects to an emdash. Could it instead be corrected to an endash if it is surrounded by spaces? e.g. "text--text" will correct to "text—text" but "text -- text" will correct to "text – text". McLerristarr | Mclay122:19, 2 November 2010 (UTC)
There's no reason to make that assumption. If the writer has included spaces, the fix should leave them in. It would mean the same thing, so it's best just to leave the writer's stylistic choice alone in an automatic edit. McLerristarr | Mclay107:43, 20 August 2012 (UTC)
Auto-replace direct use of stub categories with stub templates
If an article being edited by AWB has a direct use of a stub category (such as here; a stub category is any category ending with the word "stubs"), then the category should be replaced with a stub tag; finding the appropriate stub tag is relatively easy - in the category, there is a {{Stub Category}} template or a {{Stub category}} template, where the stub category is givenas the parameter called "newstub". עוד מישהוOd Mishehu
I have never seen the code for AWB, but given that AWB has the functionality for parsing page code, and for looking for specific parameters, I would imagine that it's not too hard. עוד מישהוOd Mishehu12:17, 14 December 2010 (UTC)
Rule: If two redirect templates exist and first parameter is the same then merge:
{{Redirect|foo|bar1|bar2}}{{Redirect|foo|bar3|bar4}} →{{Redirect|foo|bar1|bar2|bar3|bar4}} -- Magioladitis (talk) 23:38, 1 June 2013 (UTC)
Replace normal spaces with non-breaking spaces before forward slashes (/) to avoid awkward text wrapping before the slash. Do not add spaces before a slash if no space already exists. McLerristarr | Mclay109:53, 31 January 2011 (UTC)
Enhance the logic of duplicated category removal duplication logic to remove duplicated categories with different sort-keys when the one key is substring of other. i.e.: [[Category:CatName|Foo]] [[Category:CatName|Foo bar]] → [[Category:CatName|Foo bar]]. -- Magioladitis (talk) 22:58, 17 September 2012 (UTC)
I agree, in page named "Foo" category "Foo" should have empty sortkey i.e. [[Category:Foo| ]]. This will enable us fix duplicated categories like here and here. -- Magioladitis (talk) 09:58, 13 April 2013 (UTC)
The 'List filter' dialog has an option for 'Remove titles containing' with a checkbox for 'Regular expressions'. Can you add an option for 'Case sensitive'? That would make it match similar functionality in the 'Skip' tab in the text search of the 'Database scanner'. Thanks. Lightmouse (talk) 08:26, 11 June 2011 (UTC)
Occasionally, I create a list which I want to save on Wikipedia, such as a list of categories I want to nominate in a single CFD disucssion. It would be nice if AWB would allow me to save this directly on Wikipedia (maybe limited to the account's userspace - I always save it to the userspace, anyway), in stead of having to save on a local disk, and then open in a text editor and copy to Wikipedia. עוד מישהוOd Mishehu11:18, 2 August 2011 (UTC)
Added in revision
Doing save to site would be rather trivial... We should probably add a load from page type feature to compliment this.. Though, saved in the correct format, links on page would work. Reedy (talk) 21:21, 7 March 2012 (UTC)
I see no reason to create a new load option here - saving as a wikitext format should allow current load options to be good enough. עוד מישהוOd Mishehu11:51, 15 July 2014 (UTC)
Pretty much what the header says: convert template list entries to /doc subpages, similar to how AWB can convert from articles to talkpages. Just have it take any template:foo entry in the list and replace it with template:foo/doc. This would be handy for changing documentation across a given template category (or even be used to change the category, since it's often set with <includeonly> on the doc subpage.) Avicennasis @ 06:16, 6 Elul 5771 / 06:16, 5 September 2011 (UTC)
List building a query function for API. It would be excellent if API type queries could be built into AWB interface.
Recently I had to undertake a query that was unable to be determined within AWB, and instead needed to have a separate query run through the API. The only way to have that undertaken was to determine the external query within the API space, and then to import the query through the makelist "HTML scraper with advanced regex". Also as the HTML scraper does an anonymous query, rather than that of the logged-in user/bot, it was limited to 500 entries returned per query, rather than the 5000 of a bot; subsequently time-based versions of the query had to be run three times to get the requisite list. Also made the query more complex and as the line for the url to be pasted is not long, it was difficult to navigate through the url to edit it.—billinghurstsDrewth06:05, 19 October 2011 (UTC)
Added in revision
Create a list of all pages in the last # of day(s) or since last run.
It would be nice to have the ability to create a list of say all New pages created in the last day, or since I was last on.. The host of one of the wikis I am using is having all kinds of troubles with filtering out spambots, refuses at this point to give anyone admin group permissions, and I have been using AWB to mark the pages as spam. It would be nice if I could created a list of all pages since I last ran the check. Similar to using the API "&days=1" or "&since=20120517080000" (since {{#time:YmdHis|08:00:00 may 17, 2012}}. — T13 ( C • M • Click to learn how to view this signature as intended )14:34, 19 April 2012 (UTC)
The filter button currently applies the selected filters on the artice list. Can it be setup so that the user can select filters ahead of time and then click "make list". For example, I would like to load "only categories" from when I click "Make list" using "Category" make from option. Right now, I having to load a bulk of articles first and then select the filters (this takes a lot of time). Try pulling the categories under Category:Unassessed-Class India articles. Thanks, Ganeshk (talk) 22:26, 31 May 2008 (UTC)
Added in revision
Hmm. Something like List --> Filter out non mainspace and Filter duplicates (they run when new stuff has been added).. —Reedy22:35, 31 May 2008 (UTC)
Tried that...when I selected "filter non-mainspace", it returned nothing. I wanted the categories alone to pull up (select category check on the List - filter option). Regards, Ganeshk (talk) 22:47, 31 May 2008 (UTC)
Perhaps it might be worth considering have a separate list maker dialog. This would be equivalent to the 'Open' file dialog in other applications. Then you could have more room for options such as this request for pre-filtration. I would suggest merging it with the similar functionality of the 'List comparer'. I am always looking for improvements that will firstly make the terminology and interaction similar to other applications and secondly give more space for the working areas. Lightmouse (talk) 11:01, 1 June 2008 (UTC)
It would be easy enough to add on the special filter a checkbox to say, apply this filter whenever new articles are added to the List Maker... That saves other real-estate, and puts it in a relevant place? —Reedy22:57, 1 June 2008 (UTC)
When a page can't be saved right away due to timeout, connection problems, or whatever, the page is reloaded. This is a problem because any manual edits which were made are discarded when the page reloads. If I make more than a few trivial edits, I've gotten in the habit of copying the page text so I can paste it in the event this happens. I believe this only applies to the API versions. I'm currently using SVN 4973.
On the plus side, this version does seem to solve a problem in older versions; if you tried to save when it had timed out, AWB would just move along to the next page as though the page had been saved, without any indication that it really hadn't been. Hopefully a solution can be found which fixes the current problem while keeping the old one solved too. MANdARAX • XAЯAbИAM01:01, 25 July 2009 (UTC)
Added in revision
I don't really consider AWB a tool for making significant manual edits to a page – rather to make the same or very similar edits to a number of pages. Still, others could use it this way. Rjwilmsi09:37, 15 August 2009 (UTC)
Well, I want this feature badly ;-). I frequently encounter this problem. Rjwilmsi, talking about significant edits, edits like balancing unbalanced brackets, correcting a unique typo, etc. could still be saved. Thanks. --Siddhant (talk) 20:19, 4 January 2010 (UTC)
I have also had this problem often. It only happens when it has the worst possible effect – after I've worked on something for about an hour. The way I use AWB is explained here. Art LaPella (talk) 23:02, 30 July 2010 (UTC)
I've been able to avoid losing my edits by quickly copying the entire text of the page when I see that AWB is restarting. With the cursor in the edit box, Ctrl-A, Ctrl-C will select the entire text and copy it. After the reload, Ctrl-A, Ctrl-V will select the entire (reloaded) text and replace it with what you copied. HTH. --Auntof6 (talk) 03:10, 31 July 2010 (UTC)
I agree that AWB isn't for major edits, but even making a few minor edits takes a minute or two. And it really pisses me of when I don't make it in the 30 seconds (or whatever) and my work gets erased. Why is this reloading needed at all? Debresser (talk) 22:01, 26 July 2011 (UTC)
I agree - every now and then, in the course of performing some routine edits, I'll come across an article with a lot more problems and spend some time fixing them, and it's infuriating to see all that hard work lost because of a timeout/reload. Like Auntof6, I've got into the habit of manually saving any large article before I hit the Save button, so that I can recover my work if necessary - but it would be much better if AWB could remember my changes and reinstate them, rather than just reload the original page. Colonies Chris (talk) 23:55, 1 August 2011 (UTC)
This has had a solution for a while. After the page reloads, right click in the Edit box and select "Replace text with last edit". I don't know if others still want a completely automatic solution, but as far as I'm concerned, this can be marked as implemented, and can be archived. MANdARAX•XAЯAbИAM07:02, 11 August 2012 (UTC)
WP:AWB/T in section "Usage" says "typo-fixing is automatically prevented on image names, templates, wikilink targets and quotes (including indented paragraphs)". Fine for new editors and/or new AWB users, but it means many errors will probably not get fixed for many years. Not all misspellings in quotes are in the original source; many, many are typos by WP editors. Those experienced AWB users who care to take the additional time to check external sources and take on the additional risk should be allowed to do so. Of course, there is no point in messing with image names, but not all indented paragraphs contain only quoted text. If a misspelling is fixed in open text, it would seem to make sense to fix the same misspelling in a wikilink, which is probably either a dead link or a redirect from a misspelling. I can set up a F&R rule to check deeply for one or two misspellings, but I can't check all 3,000+ typos deeply at once. WP has plenty of need for fixes in these places that are now off-limits to AWB Typos, and quite a few AWB users who could and would take the training wheels off their bicycles. Christhe spelleryack15:22, 6 August 2012 (UTC)
The "→" right arrow is pretty hard to input; if there's already one in the summary, then cut and paste will work, though even that's a little tedious, especially if the clipboard is already loaded with a text selection. Other boilerplate stuff, such as ", also rm 'Dr.' per [[WP:CREDENTIAL]]" could be ready to go when needed. Perhaps Ctrl+number-key could be used, as in the Edit box. If the "Paste more" from the Edit box could be used for the Edit Summary box as well, I would be very happy with that. Christhe spelleryack19:12, 7 September 2012 (UTC)
Added in revision
I have implemented hotkey insertion (I hadn't seen this request, but it seems like a pretty obvious enhancement). In my current project I often have to type some boilerplate text. I've added a feature that lets you define text to insert using 11 fixed hotkeys (tab, and ctrl/numpad/0-9). The selections save with prefs. I have a patch file against the current version - does someone on the core team want a look at it? David Brooks (talk) 01:50, 6 October 2014 (UTC)
Add an option to not make any 'find and replace' changes to any text which is inside quotes "xx" of any kind anywhere in the article text, including not making changes anyplace within any type of block quotations--independent of any other 'find and replace' options chosen. There is currently no way to do this, which results in a great deal of unnecesary manual checking and editing (and mistakes) of the results of the 'find and replace' work. Hmains (talk) 16:39, 16 September 2012 (UTC)
The prospect of automatic but improper "correction" of quotations makes me wonder whether excluding quotations from F&R should be the default, with an option to override. ~ J. Johnson (JJ) (talk) 21:34, 13 October 2013 (UTC)
When I use the search and replace feature of AWB, I *always* want to avoid replacing text inside quotations. Obviously it is very undesirable to change a quotation, and when doing massive word replacements this adds significant burden on the editor using AWB. Marokwitz (talk) 16:18, 1 March 2014 (UTC)
@Magioladitis: I think the point is that if AWB users choose to turn off the two "Ignore" checkboxes on the F&R dialog, then they have to be careful. They have to notice that they are about to edit a quote, and either skip the edit, or go and verify the source, add a {{sic}} tag, or whatever. If an AWB user wants to use F&R in the whole page except for quotes then there is no easy way to do it. -- John of Reading (talk) 16:20, 1 March 2014 (UTC)
This becomes even more of a problem when trying to use AWB within a bot. There is work that cannot be done this way because of this quotation anomaly. Hmains (talk) 18:45, 1 March 2014 (UTC)
For me, it should no altering of text within any of the several ways in which quotations can be indicated in WP text. Whether is it by means of "xxx" or by means of any kind of template such as quote, blockquote, and so on. And also within quote= in templates. Hmains (talk) 18:55, 1 March 2014 (UTC)
I can't tell if that last mentioned request amounts to a subset of this request, or is just about use of non-word characters. But it should be a general principle that that there is absolutely no automatic "correction" of any kind of quoted material. The sole and absolute criterion of correctness for any quoted material is the original material, per the original source. Even if such material is "wrong" by any other standard, when material is quoted it implies an exact copy of the source, warts and all, and the only basis for changing a quote is if it varies from the source. AWB should protect quotes from mass changes, not encourage it. ~ J. Johnson (JJ) (talk) 20:33, 1 March 2014 (UTC)
Your statement here is at variance with the guidance offered in WP:QUOTE, which says "Exceptions are trivial spelling or typographical errors that obviously do not affect the intended meaning; these may be silently corrected ...". Editors who have good sense and well-developed judgment are capable of fixing sloppy transcriptions of spoken words and careless translations of foreign-language texts even when those appear as quotations; they know not to change "PENSYLVANIA" in the inscription on the Liberty Bell, but they know that if French "merde" in a quotation is translated as "shiit" it needs to be fixed. Same thing goes for bad punctuation in an interview with a rock musician; certainly the transcriber of the interview cannot tell that the speaker omitted a hyphen or added an unneeded comma. How are Wikipedia and its readers helped by saying that "Julius Caesar put down the rebelllion" even if that extra 'l' originated hundreds of years ago? Let me remind you that AWB is used for semiautomatic spelling corrections, not automatic ones; every change should be reviewed, and skipped if if there is doubt. Christhe spelleryack02:17, 2 March 2014 (UTC)
I completely agree with J. Johnson (JJ): AWB should not be making any changes to quoted material. It just creates trouble. This is NOT a discussion regarding manual changes to quotes so I will not discuss that. Hmains (talk) 02:37, 2 March 2014 (UTC)
Note that the suggestion is to make an option that quotations are hands-off by default with the ability to apply changes to quotations, if the user specifically selects to do so. Thus, any questions about being able to have AWB perform edits in this situation, or that situation, are specious.
My opinion is that the suggestion should be implemented (hands-off by default, with option to permit changes). I would further suggest that a popup alert be generated when quotation editing is enabled similar to the popup alert when the "Regex typo fixing" is checked.
The point is that quotations are special. They deserve handling by a human who is specifically intending to put in the extra diligence that should be applied when editing them. The reality is that AWB is used in a large number of situations where the special diligence by a human which quotations should receive is just not available, applied, or known by users that it should be applied to a particular edit.
I expect that the vast majority of edits using AWB receive somewhat cursory inspection by a human. Given the large number of edits which are performed by bots via AWB this is almost certainly true. Having the default be that changes are not made to quotations allows the best of multiple worlds: editors do not need to worry that any specific edit is in a quotation (if they are even thinking about that most of the time), bots can be run without the concern that a change which would be desireable in general text might be a bit over-aggressive for a quotation, and humans who specifically intend to be making AWB based semi-automated changes to quotations can do so.
Personally, I would find it helpful to have yet an additional option of applying changes only to quotations. That would permit dedicating a session to making only such changes which require greater diligence. Editing quotations is a different mindset than editing general text. However if the option existed to exclude editing quotations, it would be possible to approximate this by running a first pass making changes on non-quotation text, then a second pass which would be on quotations (and any edits performed by others in the interval between passes).
Having automated (find and replace, regex, typo, gen fixes, etc.) editing be excluded from quotations by default with the option to enable such changes is a good solution. — Makyen (talk) 06:07, 2 March 2014 (UTC)
I would add 2 checkboxes: "Perform replacements in quotations", "Perform replacements in main text". This way the user could target either quotations, "Wikipedia voice" text, or both. I would also recommend this feature avoids editing the article names and publication names in references. Marokwitz (talk) 07:05, 2 March 2014 (UTC)
@John of Reading:, the fact that this capability is already built into AWB is amazing - this would make it so much easier to implement the requested enhancement. Any of the AWB developers willing to take this challenge ? Marokwitz (talk) 06:47, 3 March 2014 (UTC)
Indeed, WP:QUOTE does put a qualification on my statement re "trivial" changes. Though I am not fully confident in all Wikipedians' notions of "trivial". As Mayken said, editing quotations is (and should be!) a different mindset, requiring greater diligence, but often done with rather cursory care. I say that quotations and general text be strictly segregated in regard of any automatic and even semi-automatic editing. I doubt that that we should allow even semi-automatic editing of quotations, as I doubt that any quote is so pervasive that itany correction cannot reasonably be done manually in every instance. ~ J. Johnson (JJ) (talk) 22:38, 3 March 2014 (UTC)
@John of Reading: true. This was done by general fixes. So we need two things: An option to exclude general fixes inside quotes and an option to exclude F&R changes inside quotes. -- Magioladitis (talk) 16:12, 8 March 2014 (UTC)
@Magioladitis: I don't think it should be an option for the general fixes not to edit quotes, as the general fixes shouldn't ever be doing anything contentious. This edit looks like an ordinary bug to me. -- John of Reading (talk) 16:17, 8 March 2014 (UTC)
To Implement this I have to rewrite a medium size of code because the function is enwiki only. We have two cases: If the references section already exists and if it doesn't. So I think the best would be that I make a custom module for you, you test it and you tell me if there are any more changes to be made. I could also provide some examples. -- Magioladitis (talk) 11:38, 20 September 2012 (UTC)
We would need ==Kaynaklar== and {{kaynakça}}, but {{reflist}} also works (redirected to {{kaynakça}}. We'd love to have the custom module if it's not easy to implement directly into AWB.--Khutuck (talk) 15:10, 23 September 2012 (UTC)
There is a standard block of text that is added to all articles at AFC. Articles that are accepted and moved to the main article space retain it. It would be useful for AFC and new article editors to have AWB automatically find and remove the following comment: "This will add a notice to the bottom of the page and won't blank it! The new template which says that your draft is waiting for a review will appear at the bottom; simply ignore the old (grey) drafted templates and the old (red) decline templates. A bot will update your article submission. Until then, please don't change anything in this text box and press "Save page"."
What template does this text come from so we can monitor the template for changes to the wording (or consider loading the text wording via API)? Rjwilmsi18:47, 1 April 2013 (UTC)
Actually I'd !vote for a separate page where a list of unwanted comments can be maintained. Then when someone comes along later to ask for the automatic removal of some other comment, it can be done without modifying the software again. -- John of Reading (talk) 16:43, 4 April 2013 (UTC)
Changes to See also logic for Simple English Wikipedia
AWB performs actions in Simple English Wikipedia that are not applicable to that Wiki. Below are some problems that have been identified so far:
Update logic to use Related pages instead of See also/Other pages
Currently changes some things to See also but See also isn’t used in Simple, that section is called Related pages. This should also change See also and Other pages to Related pages
AWB has logic to move some things under See also. Logic should be changed to apply the changes to the appropriate locations for Simple (ie. Related pages instead of See also)
AWB typo fixing recognises "a" or "A" before a vowel and corrects it to "an" or "An". All good. However it misses cases where the vowel is separated from the article by link syntax eg "A [[iron..." . An enhancement to recognise this would be good. -Arb. (talk) 18:21, 2 January 2013 (UTC)
Thanks for pointing that out. Just had a look. If we're on the same page it's all RegX. Not something I've got round to learning yet so someone else will have to code this. -Arb. (talk) 03:37, 3 January 2013 (UTC)
I've fixed 6,700 of these this month using a database scan and custom find+replace rules. I agree that it would be nice if the standard typo-fixer could do this fix, but I can see that it would be hard to slot into the existing scheme. -- John of Reading (talk) 21:35, 17 March 2013 (UTC)
I don't want the current Append/Prepend text after General Fixes feature to be replaced. I want another option to be allowed. -- Magioladitis (talk) 20:14, 13 October 2013 (UTC)
AWB currently features generating a list of pages 'using' Special:AllPages, but is it possible to add a function that requests a list of pages without redirects? [16] returns a list of pages without redirects, compare with [17] for the unfiltered version. I don't edit Wikipedia, but use AWB on other projects where this would be a useful tool for checking for content that can't be detected through any other method, deprecated HTML for instance. 89.168.132.45 (talk) 21:51, 20 January 2013 (UTC)
It's similar, but not really. I'm using it to scan pages for common errors, lets say spelling mistakes for the sake of argument. There's no way I can check every page for spelling mistakes just by using what links here for random pages without missing a fair few. Also, my apologies for the late reply89.168.137.160 (talk) 16:29, 15 February 2013 (UTC)
What if you generated the list of all articles (including redirects), and then added conditions to skip articles with no changes and/or articles that contain "#REDIRECT"? Hope this helps! GoingBatty (talk) 18:41, 15 February 2013 (UTC)
I would like to join the request. I have often been using AWB to scan for common errors on all pages too. Our project has around 1000 pages without redirects and around 2500 pages with redirects. Checking 1 page takes around 1 second, so when I use the method of the skipping all the pages containing the "#REDIRECT" text, it takes another 25 minutes to "check & skip" all the 1500 redirects without purpose. -- Spike144 (talk) 12:45, 21 November 2013 (UTC)
One to think about: a configuration page could tell RegExpTypoFix to work on specific template parameters. For example, the "trans_title" parameter of various "cite" templates is supposed to be correctly-spelled English; and the "ShortSummary" parameter of {{Episode list}} often contains typos. -- John of Reading (talk) 12:38, 3 February 2013 (UTC)
I was wondering about why "p/pp" would be done without a period, and what else that might imply. Looking at the example provided (Negombo Tamil dialect) I see that there are two under-done references; fixing the periods alone seems lame. It seems to me that in some cases it might be better to leave a mess alone rather than smoothing over a few esoteric details. If AWB is esoteric enough to find munged page numbers, but not enough for other problems that may exist, perhaps these findable problems ought to be flagged but not fixed, to indicate a more comprehensive review is needed. ~ J. Johnson (JJ) (talk) 02:08, 8 September 2013 (UTC)
When removing empty gallery tags, also remove Gallery section header
Sounds sane. Make sure the heading is not removed if the section is not empty, e.g. if it containes something else after the empty gallery. --TMg12:26, 11 May 2014 (UTC)
How about another tab (beside the "Start" tab?) displaying the current page's TOC as links, which, when one is clicked, acts like the TOC links in Wikipedia itself (i.e. moves the text in the Edit box to the start of that particular section)..? Apologies if this already on a list somewhere and thanks for a great tool. CsDix (talk) 15:24, 16 February 2013 (UTC)
Added in revision
Remember size and position of "Find & Replace" window
Remember size and position of "Find & Replace" window between sessions. (Apologies if I've missed a previous request.) CsDix (talk) 01:47, 2 March 2013 (UTC)
Display a counter for each output log, as is done for pages remaining in the input list. This would allow for an easy progress check (if 50 pages left in the input list, 20 pages in the Successfully saved log, 30 pages in the Skipped log, then we are halfway done that task), provided the logs were cleared at the start of the task. At the end of a task, we could quickly see whether there enough corrections to warrant a new Typo rule. Christhe spelleryack16:43, 18 March 2013 (UTC)
Added in revision
Hi Chris! Here's the workaround I would use:
Make the list and process each article in the list
If an article doesn't have a typo (e.g. it's valid text in a quote), click the Skip button to remove the article from the list
If the article has a typo, leave it unprocessed in my list and start the next article.
After looking at each article in the list, you should be left with those with a typo.
If there are enough articles, create the typo rule, refresh the typo list, and process the list to test the new rule.
If there aren't enough articles to create the typo rule, reprocess the list and fix the articles manually.
I would like the ability to process protected pages. Currently, when AWB comes across a protected page, it's automatically skipped. I would like AWB to indicate that it's protected, but go ahead and show what changes should be made; then the user can submit an {{Edit protected}} request or ask a friendly admin to take care of it. There's no reason that typos and other issues should go unfixed when a tool as useful as AWB can help to find them. MANdARAX•XAЯAbИAM02:29, 31 March 2013 (UTC)
Added in revision
I support this request but would go a little further to say that it should open a browser tab to a new section of the talk page of the protected page and prefill in an edit request (it should be able to give a detailed description of changes to be made) so that the AWB user only needs to review the edit request and submit it. AWB should once opening this tab, continue on its task to complete (which would leave the tab in the browser unaffected). As I think about implementation for this, I'm thinking it might be best served as an add-on module. :) Technical 13 (talk) 16:49, 28 August 2013 (UTC)
Making a list from "Links on page" should not alphabetize the links
The links should be shown in the sequence they (first) show up on the page. They should only be sorted if "keep list alphabetized" is set. Now there's nothing but save the wiki page source text in a text file and import this as UTF8 file to circumvent this irritating behaviour. --FA2010 (talk) 14:26, 5 April 2013 (UTC)
I don't think we can do this for you. We use the API to get links on pages since links may come from templates on a page and not just plain wikilinks. The API sorts the returned links, for example, links on page A are these from the API. So we could only provide links in the order on the page if we stopped showing links from within templates etc. Rjwilmsi16:30, 5 April 2013 (UTC)
While the new, huge edit summary box is a nice addition, could there be an option to hide it. It does remove alot of space from the edit box, especially if you have a small screen to begin with. Bgwhite (talk) 07:03, 29 May 2013 (UTC)
Added in revision
Add minimize/Maximize functionality to Normal Find and replace box
Could we please add the ability to minimize and maximize the find and replace box that appears under Normal settings? Currently the window has to be manually resized. Kumioko (talk) 14:56, 30 May 2013 (UTC)
Added in revision
Add minimize/Maximize functionality to Advanced Find and replace box
Could we please add the ability to minimize and maximize the find and replace box that appears under Advanced settings? Currently the window has to be manually resized. Kumioko (talk) 14:56, 30 May 2013 (UTC)
Can we add the capability to compare differnces in the code when updating code in modules? It would be very beneficial if we could compare the changes before making them to ensure the change is correct. Kumioko (talk) 14:58, 30 May 2013 (UTC)
My workaround is to populate the Skip Text Contains field with {{(Af[Dd]|Prod blp/dated|db-|speedy deletion-significance|Proposed deletion|Article for deletion). GoingBatty (talk) 03:59, 15 June 2013 (UTC)
The new cite templates complain when encountering |isbn=aaaaaaaaax, 978bbbbbbbbby since only one ISBN is allowed in the field (and there is no need for the equivalent 10-digit value). If aaaaaaaaa == bbbbbbbbb, "aaaaaaaaax, " should be removed (i.e. |isbn=aaaaaaaaax, 978bbbbbbbbby → |isbn=978bbbbbbbbby. —[AlanM1(talk)]—22:16, 15 September 2013 (UTC)
Added in revision
Comment: Not sure what the policy says on this, but for all books published before 2007, we should use ISBN-10. In general, we should always use the number printed on the book. But I agree in general that there's no need to have both ISBN-10 and ISBN-13 for one book. --bender235 (talk) 16:40, 23 September 2013 (UTC)
Beg to disagree. I recently saw a book which had only the 10-digit ISBN printed in the book, but which did have a 13-digit assigned. (No idea how common this might be.) The 13-digit ISBN might be useful in locating the book in various contexts, but most certainly should not cause the 10-digit ISBN to be removed.
To fix the doubled-up ISBN problem one of them should be moved (not removed) outside of the template. Perhaps even stuffing the second instance into an {{ISBN}} template. Or perhaps the templates could be modified to handle two ISBNs? ~ J. Johnson (JJ) (talk) 20:36, 6 October 2013 (UTC)
It is very, very common to have books published with only a 10-digit number to have a 13-digit number assigned to it. Two reasons. One is that they are essentially the same numbers. To convert a 10-digit number to 13, slap 978 on the front and recalculate the checksum digit (last digit). Second is that the 10-digit's equivalent 13-digit was assigned to each other when the transition happened. If you goto WorldCat, Google or Amazon, you can find pre-2000s books with a 13 digit number.
Example: As Ender's Game, the film, is about to be released, type "9780312932084" anywhere you want. That is the 13-digit equivalent of the 1985 first edition.
WP:ISBN states, "Please use the 13-digit one if available"
In most cases, there should be no reason to list more than one ISBN number. One ISBN shouldn't be moved. Template shouldn't contain more than one. If the 10-digit and 13-digit numbers are equivalent, then use the 13 as they both point to the same spot. Should AWB being automatically doing that? I don't know. But it could give a warning. Bgwhite (talk) 06:39, 7 October 2013 (UTC)
I am a little concerned that an editor with a printed book in hand that lists only the ISBN-10 might not recognize the equivalent ISBN-13 as referring to the same book. Perhaps that is something be mentioned on the Special:BookSources page?
I would be less concerned about AWB removing a 10-digit ISBN if it checks that the 13-digit version is indeed correspondent. If they are not, then that should should be flagged as an error, and no changes made. ~ J. Johnson (JJ) (talk) 20:51, 8 October 2013 (UTC)
Right. That's what I attempted to say (perhaps cryptically). AFAIK, there is an equivalence between 10- and 13-digit numbers as described by Bgwhite (and in the logic I gave) above. In the cases I looked at, where both numbers were present and the root 9 digits did NOT match, one of the two numbers was wrong – usually because it was for a different edition than the one being cited. —[AlanM1(talk)]—11:08, 3 March 2014 (UTC)
13-digit ISBNs described as such are a relatively recent phenomenon: since circa 2006 (in the lead-up to 2007, as noted above). But at least twenty years earlier (in Europe at least), these numbers were already being allocated, and being shown on the books in the form of the EAN-13 barcode. I've got several books bearing such barcodes from as long ago as 1984, which are definitely first printings, not post-2007 reprints. All that happened in 2007 is that the publishers stopped putting the ISBN-10 on the books. If it begins 978 or 979, and has 13 digits, there is no difference between an EAN-13 barcode and an ISBN-13, apart from the spacing and hyphenation (which is of no consequence). We only need one ISBN for each cite template: if there are two, one should be removed - but for a pre-2007 book, it doesn't really matter which you remove and which you keep. --Redrose64 (talk) 19:54, 31 August 2014 (UTC)
Per WP:LISTGAP in Manual of Style/Accessibility. Quote from Signpost Technology report, "Avoid separating list items by blank lines, especially in bulleted lists, as it makes lists harder to read with screen readers. For example, this list of tips to make articles accessible contains three items, and would be read by a screen reader like this: 'list of 3 items, <text of the list> ... list end'. If there were a blank line between each list item (i.e. if I had pressed the enter key twice between each point), it would be read out like this: 'list of 1 items, <text>, list end; list of 1 items ...' and so on." Bgwhite (talk) 20:27, 19 September 2013 (UTC)
Added in revision
It's true that having blank lines between bullet points is wrong and should be fixed in many cases. I was also thinking about this but came to the conclusion that it's not so easy to automate. People are using this as a feature in both numbered …
Series a, item a
Series a, item b
Series b, item a
Series b, item b
… as well as unnumbered lists. If the user wanted the two lists to be separated AWB shouldn't merge them. Suggestion: Check for loneley list items between other list items:
A general fix I'd like to see implemented: replacing interlanguage pipe links (like [[:de:Bernd Raffelhüschen|Bernd Raffelhüschen]]) with {{ILL}} (like {{Ill|de|Bernd Raffelhüschen}}), producing a temporary interwiki link. bender235 (talk) 16:33, 23 September 2013 (UTC)
In normal editing mode, AWB will notify us when a page contains non-Unicode characters and can't be edited, so we can check it manually. (This was done at my request a couple years ago.) However, in pre-parse mode the page will still be skipped. We should be able to manually review such articles regardless of editing mode. Perhaps if it can't be edited by AWB, it wouldn't be removed from the list, but wouldn't be colored green either; then when we go back in regular edit mode, we'd get the notification.
An example is Tuscarora language. I'm reviewing language articles for redundant or otherwise useless links to sites like Ethnologue, and with 8,000 articles I'm pre-parsing them. All that clutter is being left in whenever the article has non-Unicode characters, which isn't uncommon for language articles. — kwami (talk) 14:42, 1 October 2013 (UTC)
Added in revision
@Kwamikagami: - One workaround would be to enable the logging feature, pre-parse the list, and then sort the log by reason to find the articles with non-Unicode characters and fix them manually. GoingBatty (talk) 16:47, 1 October 2013 (UTC)
Thanks for pinging me.
I doubt many people would think of that, or would even be aware that they're missing anything. Would it be possible to have it as a feature? — kwami (talk) 16:51, 1 October 2013 (UTC)
@Kwamikagami:, @GoingBatty: we could add this information somewhere in the Manual with the help of GoingBatty. I think the current behaviour is satisfying. Kwamikagami, you can edit the pages that remain in the list and add all the pages skipped for having Unicode control characters too. -- Magioladitis (talk) 13:12, 18 November 2013 (UTC)
Thanks. Would it be possible to have a bot change the ones in the PUA range to &#x...; format, and tag them with {{PUA}}? And tag other things with some other marker, so we can find them? — kwami (talk) 13:16, 18 November 2013 (UTC)
Update: We've taken care of nearly all the PUA on WP-en, but new instances do crop up from pasting refs and the like. It would still be nice IMO to have them left white rather than skipped in pre-parse mode, but is much less of a problem for now. — kwami (talk) 02:08, 19 January 2014 (UTC)
Remove day of the week from citation date parameters
When creating new articles with multiple sections it is often convenient to start with the section headers (in the manner of an outline), then fill them individually. It would be irksome to have them removed before one gets to filling them! Therefore automatic removal should allow a grace period for new sections, and respect any tags indicating the article or section is under construction (see following proposal). ~ J. Johnson (JJ) (talk) 19:47, 17 October 2013 (UTC)
It wouldnt change the default behavior, but in cases where the user is knowingly needing to remove either the contents of or the entire section across multiple articles (External links, gallery sections are examples) as part of a cleanup process the adding the templates is annoying. Werieth (talk) 20:25, 17 October 2013 (UTC)
...except where the article is about a person, and the DEFAULTSORT is there in first name surname format so no one adds an incorrect DEFAULTSORT surname, first name (e.g. Bao Ninh). In the meantime, you can update your F&R rules to find {{DEFAULTSORT:%%pagename%%}}\n and replace it with nothing. GoingBatty (talk) 05:09, 20 February 2014 (UTC)
An option to not move a stubtags. Just as the option to sort interwiki's. This is good due to that Swedish Wikipedia does not have any rules/guidlines on where to put them. I do not want any edit-war...(haha) (t) Josve05a (c) 22:53, 25 November 2013 (UTC)
AutoWikiBrowser is incredibly useful to me when making large scale edits both on and offline, since I have a few offline test wikis I use for testing and development. However, it would be nice to be able to use it completely offline, or at least independently of Wikipedia servers, especially at times when AutoWikiBrowser cannot reliably connect to them and its use would be highly useful on a non Wikimedia wiki. Arcane21 (talk) 18:42, 28 November 2013 (UTC)
I find this approach to be a bit back to front. Writing lots of regexes is fine, but has the analysis been done to identify what are the most common errors. Otherwise we can spend ages writing clever regexes to find that only one page matches. Rjwilmsi14:40, 17 December 2013 (UTC)
What I would like to do/have done is get a databsae dump of all pages in the category, or get a full database dump and cross reference to those pages in the category. Then parse database dump to pull out all citation templates, all the date fields out of those. Put this in a flat file, load to MySQL, then do analysis to see how many dates are non-standard and how many have particular errors (e.g. how many start with a zero). Can then target fixes specifically at what's found. AWB genfixes likely already fix hundreds or thousands of these pages, but need to pull the hundreds/thousands out of the > 100k pages in the category. I take this approach with journal cites. Rjwilmsi17:04, 17 December 2013 (UTC)
The replace expression seems to have very limited possibilities/features. I would like for example the possibility to change $1 from upper-case to lower-case and vice versa. It would also be useful with for example \n, \u or \x that only can be used in the find expression for the moment. Svensson1 (talk) 19:54, 14 December 2013 (UTC)
Added in revision
There is a workaround using some of the MediaWiki builtin functions. Something like {{subst:lc:$1}} should convert $1 to lowercase, for example. But this won't work well if you have ticked "Add replacements to edit summary", since "subst" doesn't work in edit summaries. -- John of Reading (talk) 20:11, 14 December 2013 (UTC)
I just want the possibility to do a replace like this: "$1\n$2". But it is not extremly important for me and you say that it already works (at least outside the Regex tester), so maybe nothing needs to be done concerning \n. The two other I mentioned, \u and \x, is not so important for me for the moment. Svensson1 (talk) 14:24, 22 December 2013 (UTC)
@John of Reading: Substituting also doesn't work within <ref>...</ref> tags, so doing something lik changing months from "november" to "November" and "january" to "January" each have to be spelled out in different regex replacements. GoingBatty (talk) 14:19, 16 December 2013 (UTC)
I suspect there was confusion about \n because it does not work for replacements in the Regex Tester – it displays the text "\n" rather than the desired effect of a new line. It would be great if there was a way to get it to work there. MANdARAX•XAЯAbИAM20:43, 15 December 2013 (UTC)
@Mandarax: - If I remember correctly, you need to check one of the boxes in the Regex tester to get it to work, but don't need to check the corresponding box in the Find and replace screen to get it to work there. I'll look in to that further later. GoingBatty (talk) 14:16, 16 December 2013 (UTC)
As far as I know, there's nothing a user can do to make this work correctly in the Tester. When I first encountered this problem long, long ago, I assumed that the Tester was a tool which accurately mirrored what happens with the actual Find and replace, but I eventually discovered that that is not the case for \n. MANdARAX•XAЯAbИAM04:30, 18 December 2013 (UTC)
I second this, as it is a badly needed feature for disambiguators. For example, Saint Thomas is a disambiguation page, but many of the incoming links will be to redirects like St. Thomas, St Thomas, St. Thomas College, and St. Thomas' College. When using the disambig feature of AWB, these redirect links are ignored in articles, when they should be treated just like other disambiguation links. bd2412T18:12, 5 January 2014 (UTC)
Even if it was technically possible, their approaches to making changes are very different. For example, WPCleaner allows you to fix multiple ambiguous links in an article before clicking save. On the other hand, AWB will automatically make many changes for you, whereas WPCleaner will suggest some but make you manually do each one. GoingBatty (talk) 19:50, 5 January 2014 (UTC)
AWB's disambig feature suggests possible alternative fixes for a disambiguation page, but it can only be loaded with links to one disambiguation page at a time (and, as noted above, ignores redirects to that disambiguation page). It would be great if AWB could recognize those redirects and treat them the same as the disambiguation title to which they redirect, and if AWB would let you load links to multiple disambiguation pages at a time. For example, it is not uncommon for articles to be written that include multiple disambiguation links to nationality/language terms (e.g. "she speaks English and French"; or "the region was colonized by the English and the French"). Should any of this be technically challenging to implement? bd2412T20:00, 5 January 2014 (UTC)
Let's see an example. I have a disambiguation page (let's name it "thing") with disambiguation links "thing1", "thing2", "thing3". Then I take a "what's links to" to "thing" pages list and get disambig links from "thing". Works great on start. But then I found a redirect page "thingy" that points straight to "thing" (#REDIRECT [[thing]]). So I take a "what's links to" to "thingy" pages list and... ups, pages doesn't contain word "thing", they contain "thingy", so disamig doesn't work and nothing changed. Ok, I put "thingy" to disambig link and get disambig links. Ups again, there is only one page - "thing". For now I need process pages twice: first to replace "thingy" with "thing", secong to disambig "thing" to correct page. I think it can be simplified. KorneySan (talk) 17:07, 9 January 2014 (UTC)
The example should be, if the disambiguation page is "thing", with redirects "thingy" and "thing (song)" and "thing (film)", that AWB should treat the disambiguation title and all of its redirects the same, so that it provides a solution irrespective of whether the link is the page title or a redirect to it. Run once, solve anywhere. This is particularly important for pages like Destiny (disambiguation), for which links to the actual title do not need to be fixed at all but for which links to redirects Destiny (comics), Destiny (album), Destiny (film), and Destiny (song) need to be fixed. bd2412T20:03, 11 January 2014 (UTC)
It will be useful if AWB would replace dmy dates or mdy dates with mdy or dmy when {{use mdy dates}} or {{use dmy dates}} is used on page respectively. Wikiuser13 (talk) 16:14, 9 January 2014 (UTC)
@GoingBatty: I tired it after you suggested. Couldn't it be built in? And a problem is that it convert all dates to dmy or mdy without seeing {{use dmy dates}} or {{use mdy dates}}. If you are editing a page with {{use mdy dates}} and using module to convert dates to dmy, it will convert all mdy dates to dmy.--Wikiuser13 (talk) 08:06, 10 January 2014 (UTC)
@Wikiuser13: You're right, so if you're using the dmy module, you should probably skip pages that don't already have {{use dmy dates}}. As documented in Template:Use dmy dates#Usage, some people use this template to indicate that ALL dates should be in dmy format, while others think that dates in references can be in a different format than the body of the article. Since Ohconfucius' modules/scripts change ALL dates, if this were added to AWB's general fixes, I imagine that people in the latter camp would be upset at the bot owners who are running general fixes that changed their formatting en masse. GoingBatty (talk) 13:30, 10 January 2014 (UTC)
I'm not sure if this is a "major" problem but I've noticed when using AWB to delete, protect or move articles on Wikia wiki's the actions do not update any of the counts at the bottom of the AWB update counters. I assume its because these are log actions and not "edits" but wanted to mention it. It might be useful to add a counter for log actions or something if its decided these shouldn't be included in the edit counts. Reguyla (talk) 18:25, 6 January 2014 (UTC)
{{cite web|url=http://web.archive.org/web/YYYYMMDDTTTTTT/http://example.com/|title=Example title}} to {{cite web|url=http://example.com/|archiveurl=http://web.archive.org/web/YYYYMMDDTTTTTT/http://example.com/|archivedate=YYYY-MM-DD|title=Example title}} (Y=Year e.g. 2010, M=Month e.g. 12 (December), D=Day e.g. 10, T=Time e.g. 112345 (11:23:45).
(Example: {{cite web |url=http://web.archive.org/web/20070715115352/http://www.paxilcr.com/safety_information/important_safety_information.html |title=Important Safety Information about Paxil CR |publisher=[[GlaxoSmithKline]]}}</ref> to {{cite web |url=http://www.paxilcr.com/safety_information/important_safety_information.html |archiveurl=http://web.archive.org/web/20070715115352/http://www.paxilcr.com/safety_information/important_safety_information.html |archivedate=2007-07-15 |title=Important Safety Information about Paxil CR |publisher=[[GlaxoSmithKline]]}}</ref> ) (t) Josve05a (c)17:54, 18 January 2014 (UTC)
InternetArchive have now started using wayback.archive.org instead of web.archive.org, so that shold be included as well.(t) Josve05a (c)01:32, 19 January 2014 (UTC)
In the find statement, I suggest looking for both "http" and "https". In the replace statement, I suggest adding |archiveurl=//web.archive.org... to create a protocol-relative link. GoingBatty (talk) 03:41, 19 January 2014 (UTC)
archive.org certainly hasn't changed over to wayback.archive.org completely. Links to their site from archive pages still go to web.archive.org. In addition, I made a data request to them for the URL to an archive page. The response indicated an archive address in the web.archive.org domain. In a brief look, I found no mention of a change to wayback.archive.org on their site. When I went to wayback.archive.org I was redirected to archive.org. It appears that archives are available at wayback.archive.org URLs, but is is not instead ofweb.archive.org.
web.archive.org still appears to be their primary choice.
Josve05a, I also agree with GoingBatty and Makyen in using protocol relative links, per VPP. The decision there was:
"Use HTTPS links for HTTPS only sites, protocol relative links for sites that support both HTTP and HTTPS, and HTTP links for sites that don't support HTTPS at all"
That effectively means that this proposal would be changed to:
Make AWB change the following:
{{cite web|url=http://web.archive.org/web/YYYYMMDDTTTTTT/http://example.com/|title=Example title}} or
{{cite web|url=https://web.archive.org/web/YYYYMMDDTTTTTT/http://example.com/|title=Example title}} to
{{cite web|url=http://example.com/|archiveurl=//web.archive.org/web/YYYYMMDDTTTTTT/http://example.com/|archivedate=YYYY-MM-DD|title=Example title}} (Y=Year e.g. 2010, M=Month e.g. 12 (December), D=Day e.g. 10, T=Time e.g. 112345 (11:23:45).
Since AWB can't possibly know if the http://example.com site supports protocol relative links, the http//: for the |url= parameter should remain unchanged for now; however, if this is to be added, there's no reason not to add the same fix for citations that already include both the |url= and |archiveurl= parameters; that would add the following to the find statement of the proposal:
{{cite web|url=http://example.com/|archiveurl=http://web.archive.org/web/YYYYMMDDTTTTTT/http://example.com/|title=Example title}} or
{{cite web|url=http://example.com/|archiveurl=https://web.archive.org/web/YYYYMMDDTTTTTT/http://example.com/|title=Example title}} or
This could be the start of the implementation plan the decision reached at the Village Pump lacked. If successful, it could be duplicated for other websites known to support protocol relative links. Additional steps would be to create the same tasks for the other {{cite ...| formats and modify the citation tools to exclude http//: or https//: in the |archiveurl= parameter to avoid the problem to start with.—D'Ranged 1talk12:32, 10 May 2014 (UTC)
When the "Make List" section has template-namespace pages in it, there should be an option to open the /doc page instead of the template page itself, for recategorization and other useful editing purposes. APerson (talk!) 21:48, 25 January 2014 (UTC)
Hi! I like AWB very much, but it'd be even better if it supported wildcards. Like, I'd love to be able to search for [[*|turnip]] so that I could find anything that looks like a link for "turnip" and probably wants turnip but is directing readers somewhere else. If wildcards currently exist, they aren't documented as far as I saw. Red Slash03:32, 26 January 2014 (UTC)
Added in revision
If you have built a list of articles and want to search for [[*|turnip]] within them, then the existing "Find & Replace" will find them for you. A regular expression such as \[\[[^\|\]\[]+\|\s*turnip\s*\]\] will find two open square brackets, then arbitrary text that doesn't include any pipes or square brackets, then a pipe, then the word "turnip", and then two close square brackets.
If you asking about using this kind of search to create the list of articles, then this is already available, just about. You can search with arbitrary regular expressions if you download a copy of Wikipedia and then use AWB's database scanner. But that's a big download. I'd be happy to do searches like this for you, but my current download is two months old. If you watchlist User:John of Reading/Latest download you'll get a notification when I have the February database. -- John of Reading (talk) 08:41, 26 January 2014 (UTC)
This page has character(s) in the Unicode Private Use Area so unfortunately can't be edited with AWB. The page will now be skipped
Pages that are "This page has character(s) in the Unicode Private Use Area so unfortunately can't be edited with AWB. The page will now be skipped" should be added to the false positives list by default so editors can go back and fix those pages manually. Technical 13 (talk) 19:18, 1 February 2014 (UTC)
Added in revision
Technical 13, this doesn't answer your question, but.... I often delete the Unicode Private Use (PUA) if I don't see a reason for it. kwami has started deleting the PUAs or adding {{PUA}}. Template docs show which articles use the template and how kwami finds articles with a PUA. Any article with a PUA would have it added in the past month. Bgwhite (talk) 21:50, 1 February 2014 (UTC)
Interesting. Well, alternatively or concurrently with my request above, if AWB could tag those pages with a new section at the bottom of the page with that template, they would be easier to find, no? Unfortunately, I'm not an admin (not sure I want it although I've played with the idea in my head to be more useful in technical areas), so I can't delete and currently, AWB offers me no way to go back and find which pages those were without running the queue over again from scratch, which takes forever when you have ##,### articles to go through... Technical 13 (talk) 22:01, 1 February 2014 (UTC)
I don't have logging enabled. Last time I had it enabled, I couldn't find anything in it anyways because it was too cluttered and there isn't a good filtering option. Technical 13 (talk) 22:45, 1 February 2014 (UTC)
AWB can not edit these pages because unintentional things happen when saved by AWB. For instance, characters may change and text may be altered in a way we do not control. -- Magioladitis (talk) 22:52, 1 February 2014 (UTC)
Mag, I'm not suggesting it should edit the page, I'm suggesting it should append a category to the bottom in a new nameless section (so it doesn't touch the existing code at all) using &action=edit§ion=new and for these pages to be logged in my false positives text file so I can go back and fix them after. Technical 13 (talk) 23:17, 1 February 2014 (UTC)
I think it should be explored to see if editing/submitting the page in safemode would enable AWB to edit these pages (perhaps just to ID and/or remove the PUA characters). Safemode is automatically entered if the requesting browser is identified as being on the blacklist. The blacklist is a list of browsers known to not properly handle unicode characters. When in safemode all "the non-ascii characters in the initial content of the textarea are converted into the form &#xxxx; and the value safemode=1 is hidden in the form".
In a brief look, I did not see a direct way to request for the data to be delivered with the conversions to &#xxxx; already done. However, if AWB is unable to perform such conversions itself and submit the data using safemode, AWB could spoof the browser ID to have the data sent pre-converted. However, doing this would be a much larger/different RFE than just logging the event. It is also possible that this method of handling PUA characters has already been investigated and found to not be viable. Makyen (talk) 19:46, 3 February 2014 (UTC)
Make Arabic Wikipedia list of typos usable on other Arabic Wikimedia projects
I would like to use Arabic Wikipedia list of typos on Arabic Wikinews. When enabling the "Regex typo fixing" option, it gives me the following message: "No list of typos was found. Would you like to use the list of typos from the English Wikipedia? Only choose 'Yes' if this is an English wiki." I would like to make the Arabic Wikipedia's list of typos usable on other Arabic Wikimedia projects and also on other wikis that use Arabic as a content language. One might suggest that I create a local copy of the list of typos on Arabic Wikinews, however maintaining 7 local copies for the 7 Arabic Wikimedia projects is hard. It is easier to maintain just one list centrally. Ideally, AWB should be aware of both the local list of typos on Arabic Wikinews and the global one on Arabic Wikipedia and the local list takes precedence over the global one. Meno25 (talk) 19:09, 7 February 2014 (UTC)
I would like to recommend a change to the Append/Prepend function under the more function. Currently, this function seems to fall at the end of all other edits. Unfortunately for me I often find that I have to do another pass to move what I have just added. With that in mind I would like to suggest either moving the append/prepend function to be first processed rather than last or even better, allow the editor to specify the order as is possible for the Normal settings options. Reguyla (talk) 19:07, 14 February 2014 (UTC)
Thank you for the quick response. Because I am editing at Wikia the majority of the general fixes don't apply and most of my logic is built into either modules or the Find and Replace functions. It does not appear that checking the sort meta data after has any effect on either of those things. I have figured out how to prepend things by coding it into a module but I have not yet been able to do so with the append stuff. Do you happen to know if there is an example of how this can be done? Reguyla (talk) 03:43, 16 February 2014 (UTC)
Thanks, your right that will work for somethings, but it still seems like its a lot harder than it needs to be. I'm fairly technical so I will figure it out but not everyone can or should be required too IMO. Thanks for the help though. Reguyla (talk) 04:35, 16 February 2014 (UTC)
Add two options to AWB to enable new and experimental typos, this would involve creating a new section atAWB/T for the experimental ones. The new typos would include everything under 'New additions' and would be enabled by default. The change would be the addition of an section on that page for experimental typos and an option in AWB, disabled by default, to use them. This experimental section could include typo fixes that are a work in progress, or possibly ones that will never 'graduate' due to the high false positives. It would mean that these type of typos fixes could be more collaboratively worked on and used by anyone who knows what they're getting in for.
What about adding an alert (beep sound) when a task (list) is done? Sometimes you are leaving AWB on a background (watching a movie :) and I think would be good to have a simple sound when a list is done, to load another one or to close a program. Blackfish (talk) 20:01, 11 March 2014 (UTC)
I have found myself yearning for a one-button settings load ability. I've been working my way through Category:Pages with citations using unsupported parameters with a script running on one computer that gives me unsupported parameter names that I then add to a script running on another computer. I make those additions using Notepad++ because of it's syntax coloring and bracket matching abilities and because it's a real editor. But, that does mean that I have to make the change, ctrl+s, go to AWB then File > Recent settings > select the file then Start.
Perhaps, to save space, the start button can be overloaded with this functionality. AWB sees that the currently loaded settings file is older than the current file so it changes the button's color/legend (perhaps to Reload settings - a tool tip might show the file name). At a fresh start up, the Reload settings button would reload the last used settings file.
Sometimes disambiguation links are put inside the "disambiguation needed" template like the following: {{disambiguation needed|Jones Beach|date=June 2011}}; this yields the outcome, Jones Beach[disambiguation needed]. AWB does not seem to pick up links that are inside the template like this, so these still must be fixed manually. It would be great if AWB could recognize and address disambiguation links inside such a template, and eliminate the template entirely when the link is fixed. bd2412T14:35, 8 May 2014 (UTC)
For the 'sort metadata after' feature in Append Text function, I found that it cannot sort interlanguage link in wikia after those appended text. Please add the support to it. C933103 (talk) 06:46, 21 June 2014 (UTC)
A certain wiki that i want to use the AWB on does not have the wiki open in default port number, and typing http://IP address:port number/ into the AWB, it would ignore the port number and result in failure.C933103 (talk) 06:56, 21 June 2014 (UTC)
This is most commonly seen at the beginning of articles where it gives the person or place's name in its native language.
The reason to use {{lang}} is given at Template:Lang#Rationale. However, the most important reason is for accessibility – screen readers need language info to speak text in the correct language – and to satisfy Wikipedia accessibility guidelines. Bgwhite (talk) 21:48, 7 July 2014 (UTC)
It would be nice if we could optionally request AWB to warn us before saving if our changes introduce red links. I doubt many would want to be told about existing red links, but that could be an additional option. (I specifically want it for File: links especially, because I've been running with changing all text, which I've just turned off again -- I'd like those two check boxes to be able to be broken down into about 6 detailed boxes so we can fix things more automatically, but not break things :-) ) Mark Hurd (talk) 08:01, 27 August 2014 (UTC)
Added in revision
It would definitely be a good idea - and this should also include transclusions of red-linked templates. Of course, the extra work involved with detecting these (including geting data from the wiki) would make it undesireable for anyone who doesn't actually want it. עוד מישהוOd Mishehu13:02, 12 October 2014 (UTC)
Yes, definitely optional, but I'd hope it can be limited to noting where text has been changed within [[<here>|...]] and, given your suggestion, {{<here>|...}} and only rechecking those links. However, it may be rather simpler to just render the page and look for new red links. Mark Hurd (talk) 09:33, 27 November 2014 (UTC)
Recently over on Wikisource I had the issue of 're-aligning' a number of pages where the underlying File: had been moved and the Index: and Pages: needed to be re-aligned. (see my Recent contirbutions on that project for the manual effort.)
Would it be possible for AWB to implement a feature which would allow for a list of pages to be Moved based on a specified pattern?
(I will note here that this would be something like this might need to be admin only if implemented for high traffic projects though.)
I will also note that AWB currently doesn't recognise that I can move pages on Wikisource, albiet I'm not an admin equivalent there ( I assume this is by design to prevent page move vandalism.)
Alternatively would an AWB developer/contibrutor be willing to open discussions with the Wikisource Community about a customised AWB version which would help assist with large scale cleanup tasks at Wikisource.
Another problem that arises at Wikisource is when scan sources (typically djvu) are patched for missing pages by the insertion or deletion of additional sheets in the original file. This means that any existing transcription has to be moved to the new layout which is time consuming when performed manually.
Would it be possible for there to be a 'page block shift' feature in a future AWB version which can do the following?
For a specified Index:filename (typically something.djvu) which has a set of subpages Page:filename/x to Page:filename/y move those pages so that they are Page:filename/x to Page:filename/y+n where n is an integer offset supplied by the user of the function in AWB, x and y are the start and end points of the portion to shift.
(NB. Because this specifically relates to the Index: and Page: nammespaces, it would only be applicable to Wikisource).
Added in revision
Generate a list of user pages/user talk pages based on editors who have commented on a particular page (or better yet, particular section of a page)
To be more precise: get the list of links; change to talk pages (some users have links to both their user page and their talk page, while others only have one); and remove duplicates. עוד מישהוOd Mishehu11:48, 7 October 2014 (UTC)
An "Are You Sure" dialog for when the user accidentally hits Stop or the Escape key. I've sometimes lost an edit because of a stray click. Maybe it could be made so that the dialog only happens when there is the potential for lost work. David Brooks (talk) 02:00, 6 October 2014 (UTC)
Added in revision
This is a good idea for when an editor has made additional edits on top of the automated ones. Otherwise, 'Stop' isn't really a problem as the article can be submitted again. This idea could be applied to 'Skip' as well. Stevie is the man!Talk • Work12:58, 6 October 2014 (UTC)
Yes, my thought. I wouldn't want to interrupt the workflow in cases where the user wants an immediate stop. Is there a simple test the code can make for the editor having made manual edits? David Brooks (talk) 17:08, 6 October 2014 (UTC)
The simplest one would be: create a boolean (bool) flag. When loading a page, set it to false; as soon as the user makes any manual change, set it to true. עוד מישהוOd Mishehu07:25, 12 October 2014 (UTC)
No. If I have that set, then it will remind me even if I had just saved the settings; and if I hadn't, I need to closew that window, save the settings, and then give the exit command again. עוד מישהוOd Mishehu11:46, 7 October 2014 (UTC)
An AutoWikiBrowser server client. This would be a GUI management console (with a command line interface built in if needed) that allows for AutoWikiBrowser servers. The server client would function as a way to hook AWB to work specifically with a MediaWiki installation (allowing customization of AWB for users to for with the settings of the wiki in question). It cab also host tasks and settings for AWB which allow anyone who has access to that server cab use specific tasks for their bots. Also it can have a panel to manage bots usernanes , passwords z account details etc. I am proposing this because large wikis like Wikipedia could benefit from having a server in ehich all the AWB users could more easily share bot tasks as well as manage it better).66.102.7.86 (talk) 20:03, 11 October 2014 (UTC)
It would be nice if AWB had a feature for changing all user-entered string values with a find-and-replace feature. For example, when creating *CountyTN-geo-stub tags and populating them, I need to do the following:
Replace {{Tennessee-geo-stub}} with {{*CountyTN-geo-stub}}; this means that if multiple stub tags are present on the page, and {{Tennessee-geo-stub}} is one of them, then the new tag will replace it.
Append the page with {{*CountyTN-geo-stub}}; this means that the page wilol end up with the tag, even if {{Tennessee-geo-stub}} wasn't present on the page.
Of course, I have the county name in here twice (and some times, when dealing with similar issues, it could be a lot more), I would like to be able to change it once. עוד מישהוOd Mishehu14:07, 14 October 2014 (UTC)
This works correctly when that stub tag is the last one; however, if there are other stub tags afterwords (such as Harrison Bay State Park), ther stub tags end up in the wrong order - I replace one stgub tag with an other, but it's in a different place.
There is no prescribed order for stub templates. The only rule about their positioning (see MOS:LAYOUT or the doc page of any stub template) is that they be at the very end of the page, except that any interlanguage links that may still be present should be after the stubs. --Redrose64 (talk) 23:11, 14 October 2014 (UTC)
There may be no MOS reason that the stub tags already prewsent are in that order, but I want to make it easy to see what I'm doing; if the removed tag and the added tag are in the same place, that's musch easier. And some times I also add the template in question to the summary (in the form of {{[[template:Foo-stub|Foo-stub]]}}), which means that it's there 2 more times. עוד מישהוOd Mishehu10:31, 15 October 2014 (UTC)
And even for this task, the external tool I have been using to create the county geo-stub lists is no longer working, so I have an other use of the county name in the settings - in the list generator, where I use "Geography of * County, Tennessee". עוד מישהוOd Mishehu13:33, 19 October 2014 (UTC)
I've been doing this; I want a more user-friendly solution. This would also make sure that I son't accidently mess up any of the mainingful strings (words like "true" and "false", the xml tag names, etc.) עוד מישהוOd Mishehu19:03, 17 March 2015 (UTC)
Option to increase number of articles loaded from Random pages
I'd like to request a programming update the next time a new version is released. My apologies if this is something that's already been dealt with, but I thought I should raise it just in case.
What I'm finding is that when I do disambiguation runs, I'm running into an increasing number of links where the pipetext has been set off from the dablink by means of {{!}} instead of the conventional | . But the result of that ends up being that AWB doesn't recognize it as a link to be disambiguated — so instead of being able to fix it within AWB, I end up having to manually edit that link.
I'd like to be able to configure two regular expressions via the "Options" menu somewhere, one to identify a suspect edit and one to identify a suspect edit summary. The "Save" button would check the article text and the edit summary against these regular expressions, and perhaps put up Yes/No warnings: "Are you sure you want to save this edit?" and "Are you sure you want to save this edit summary?".
Background: My typo-fixing method replaces text with the word "INVESTIGATE", to tell me that I need to edit the text manually or to tweak one of my typo-fixing rules. Usually this works well, but there is a constant danger that I'll accidentally save an edit or edit summary containing the bogus word "INVESTIGATE". Sometimes I realise my mistake instantly, but a few have slipped through. The latest database dump shows that my error at John Henry Mears lasted two months. -- John of Reading (talk) 16:23, 10 December 2014 (UTC)
Some times, I want to stop in the middle of an AWB task (usually to save and come back later, occasionally I have a specific reason to skip to a different part of the list); unfortunartely, that means that I have to wait for the program to parse the next page (which I don't want to do immediately). I would like a button which would save the current page, but not load and parse the next one. עוד מישהוOd Mishehu 16:02, 15 December 2014 (UTC) עוד מישהוOd Mishehu16:02, 15 December 2014 (UTC)
It would be nice to be able to build a list of CNRs for specified namespaces. For example, if I wanted to see how many article pages were redirected to Draft: or User: — {{U|Technical 13}} (e • t • c)17:35, 17 December 2014 (UTC)
Added in revision
Beep sound when done skipping and ready to review an edit with AWB in focus
When going through a sparse list, it'd be very useful to have a "beep when ready to edit" option whenever AWB presents a page for review. There seems to be a timeout feature which blanks the viewing area of the last edit as the list is processed with many skips. That's kind of useful, when you see it. Even just a system beep when an edit is ready for review would be much more useful.
Most of the missing DEFAULTSORT on the French WP are now added using AWB with %%key%%, which removes accents: for instance, "é" is changed to "e", though doing so has been useless for more than a year, because $wgCategoryCollation is now uca-fr (almost identical to uca-default), so the sorting inside categories no longer puts "é" after all "e" words. This gives a wrong example to users not aware of this technicality, who are thus encouraged to believe that, when adding by hand a DEFAULTSORT, they still need to make the small extra effort of removing accents. With or without accents makes no difference on category pages when the sorting algorithm is uca-..., but users of WPs in languages that use accents (about all languages except English) would appreciate a %%key-uca%% similar to %%key%% but not removing accents, or as another solution, %%key%% not removing accents when the sorting algorithm is uca-.... Oliv0 (talk) 07:38, 7 January 2015 (UTC)
"Internal link" is the my most frequently used button in the Edit box. A keyboard shortcut would be very convenient, and to have it/them show up on mouse-hover over their respective buttons, as the others do. Thanks! ~Tom.Reding (talk ⋅contribs ⋅dgaf)21:10, 22 January 2015 (UTC)
Added in revision
Make user-defined rules able to handle nested templates
Current behavior: on a page with {{Reflist {{citation ...}} {{citation ...}} }}, rules are applied to the entire Reflist indiscriminately, and not on a citation-template to citation-template basis.
Desired behavior: Operate on the inner-most-nested template, out. Not infinitely deep, but at least 2 levels would solve a lot of problems, as triple-nested template are more rare. ~Tom.Reding (talk ⋅contribs ⋅dgaf)15:14, 3 February 2015 (UTC)
Added in revision
Workaround: at the beginning of your rules/script, replace {{Reflist|refs=\s (and variants) with {{Reflist|refs=REMOVE-CLOSING-BRACES}}\s and create a subsequent rule to remove "REMOVE-CLOSING-BRACES}}" at the end of your script/rules. This speeds things up dramatically, but it would still be good to have a proper solution. ~Tom.Reding (talk ⋅contribs ⋅dgaf)17:31, 23 March 2015 (UTC)
Remember "AWB Regex Tester" window size for session, just like "Replace Special"
When recursing a rule several times, it's necessary to copy the results from the "Result:" textbox to the "Text to search:" textbox. This feature would make that faster than performing manually (Ctrl+C works in "Result:"). ~Tom.Reding (talk ⋅contribs ⋅dgaf)13:50, 5 February 2015 (UTC)
The pipe trick doesn't work in some places (such as within <ref> and <gallery> tags), which leaves code such as [[Subject (disambiguator)|]] visible to the readers. I have a BRFA filed to change these to [[Subject (disambiguator)|Subject]] using AWB's Find/Replace functionality so the intended wikilinks will be displayed properly (e.g. this edit). However, if this could be added to AWB's general fixes, then these could be fixed by many more people while they're making other edits. Thanks! GoingBatty (talk) 02:40, 8 February 2015 (UTC)
Added in revision
Add a "skip if page </> x bytes" option, where x is user-defined
There seems to be a limit at "Skip if Contains: .{935,}". This 17 kB page is skipped only if ".{935,}" and lower, but not ".{936,}" and higher. ".{936,100000}" doesn't skip either. Is this an AWB problem or a regex engine one?
I think the "skip" regex is being run against each paragraph separately, not the whole article. The longest paragraphs in that article have 935-ish characters. -- John of Reading (talk) 20:32, 17 March 2015 (UTC)
See Template:Citation#Authors and Editors below that for author-link and editor-link functionality in citations. It would be nice to be able to find all pages that refer to a given author and generate links for them. Jerodlycett (talk) 13:29, 4 March 2015 (UTC)
For some of us who do crosswiki work (xwiki spam, and commons work), and much of it, there would be value in having AWB be able to login for all the wikis rather than having to login per wiki, even with a saved account. So some sort of persistent login. If we can get the login, then the ability to add interwiki links in the form s:en:User:Blahblah would add some good portability to these tasks. Sort of things can be adding {{delete}} template, or doing image replacements that fail with CommonsDelinker. — billinghurstsDrewth11:11, 8 March 2015 (UTC)
While editing a page, sometimes you can get disconnected (especially if on WiFi), which erases all changes made in the Edit box. I suggest saving the Edit box text if you get disconnected and restore it on reconnect with an updated diff. ~Tom.Reding (talk ⋅contribs ⋅dgaf)16:54, 15 March 2015 (UTC)
Or at least pausing to allow the user to copy the edit box, or a "Copy edit box text to clipboard?" button on the "You have been disconnected" popup. ~Tom.Reding (talk ⋅contribs ⋅dgaf)23:40, 17 March 2015 (UTC)
Added in revision
@Tom.Reding: Are you aware of the "Replace text with last edit" option at the bottom of the right-click menu in the text box? I think that's supposed to handle this case; after the re-start, you should be able to select this and have a second attempt at saving your changes. -- John of Reading (talk) 05:22, 18 March 2015 (UTC)
In addition to applying a rule X number of times, please consider adding a feature that will stop when no changes are made, with X being the cap. If X = 0 (or maybe -1), and "Repeat until no changes made" is checked, then continue without a cap. Or grey-out the "Apply No. of times" box when "Repeat ..." is checked. Whichever's easier for y'all. ~Tom.Reding (talk ⋅contribs ⋅dgaf)15:15, 16 March 2015 (UTC)
Added in revision
Make a "Skip list" tab that skips pages on the main list that are also on the "Skip list".
Make a tab somewhere, maybe next to the "Skip" tab, which has similar list creation capabilities (or at least allows you to paste in a list of your own) that are skipped if on the main list. One useful case is when someone wants to omit pages that they have recently edited from a larger list. ~Tom.Reding (talk ⋅contribs ⋅dgaf)15:22, 16 March 2015 (UTC)
Moving ==See also== to its proper location seems to happen only when it is below ==References==, and when ==Notes== is either absent or below the original ==See also== location. The presence of ==External links==, ==Further reading==, and/or ==Notes== only alerts the user and doesn't automatically move ==See also== to the top, even though such a fix seems to be a small extension of the current fix-behavior. I didn't test every permutation, so I'm just recalling from the many manual ==See also== placement fixes I've done. ~Tom.Reding (talk ⋅contribs ⋅dgaf)23:35, 17 March 2015 (UTC)
Currently when disambiguating, if you are unsure what the link should be disambiguated to AWB lets you tag it with {{dn}} so that it's more obvious and hopefully another editor will see it and fix it. This behaviour should be changed so that AWB adds the 'full' {{Disambiguation needed}} template along withe the date parameter. This will reduce the need for future cleanup by bots etc. Jamesmcmahon0 (talk) 12:31, 18 March 2015 (UTC)
There is a problem that if you reparse the page, AWB finds the Disambig link again and brings up the disambig dialogue box. Could the AWB logic be changed so that the disambiguation is done before general fixes etc. are applied? Jamesmcmahon0 (talk) 15:31, 18 March 2015 (UTC)
Add keyboard shortcut Alt+f to toggle between "If" and "Find" tabs in the "Replace Special" window
The only problem with many regex rules is that, even with good rule descriptors, finding all rules which touch an input string can take a long time. It'd be great to have a text dialog similar to the existing AWB Regex Tester window, except that it runs the text through all rules sequentially, then highlights in light blue (say) all rules in the rule list which operated on it. ~Tom.Reding (talk ⋅contribs ⋅dgaf)23:02, 22 March 2015 (UTC)
To make picking the most appropriate (long) edit summary much easier.
Perhaps make the right-most edge of the dropdown dynamically extend to the right-most edge of AWB. That way, users can tailor its width to their needs (if they wikilink summaries, for example). AWB has a minimum width, and the dropdown is still short of that. ~Tom.Reding (talk ⋅contribs ⋅dgaf)13:47, 26 March 2015 (UTC)
Added in revision
Smarter Find & Replace skip with respect to minor replacements
Currently there is a feature to "skip if only minor replacement made". However, this works in a narrow manner, in that if there's only minor replacements, but other changes, such as "general fixes" or typos, these other changes are not done. I want this: "skip if only minor replacement made and no other changes are made". This will allow me to do more comprehensive cleanups. Stevie is the man!Talk • Work11:42, 9 April 2015 (UTC)
Added in revision
Make the Regex Tester's Find editbox vertical-scrollable
New
Done there if I can't fine one. ~ Tom.Reding (talk ⋅contribs ⋅dgaf) 20:21, 17 March 2015 (UTC)
Feature added in next release