This is an archive of past discussions on 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.

Status	Feature added in next release
Description	Currently AWB just displays a standard error dialogue when trying to edit a page that's been protected to prevent creation. Conversely, when trying to edit an existing page that is protected, AWB skips the article with the reason 'Page is Protected'. I would like the handling of the former situation to be the same as the latter. Thanks -- Rjwilmsi (talk) 19:12, 16 November 2007 (UTC)
Added in revision

You wouldnt happen to have an example of a protected against creation page would you? —Reedy Boy 00:25, 18 November 2007 (UTC)

Brian peppers. Arthena^(talk) 00:00, 19 November 2007 (UTC)

Also: I Wish I Was An Astronaut (band) Thanks Rjwilmsi (talk) 20:58, 19 November 2007 (UTC)

AWB gets a nasty error when trying to load a page that has been salted. Try loading User:MelonBot/Sandbox3 to see what I mean. Happy‑melon 21:12, 2 January 2008 (UTC)

I get no error.. It will say its "saved", but wont actually create it.. Then it seems to be using the wait/busy cursor on the web control.. —Reedy Boy 16:59, 4 January 2008 (UTC)

Note, admins dont get this error. As per:

Hi there. I noticed you unprotected and recreated User:MelonBot/Sandbox3, which had previously been salted. As you can see, this page was actually deleted and salted at my own request, to function as a test page to ensure that MelonBot does not produce unexpected behavior when it encounters salted pages. Could you restore the page to its salted state please? Happy‑melon 19:10, 17 January 2008 (UTC)

I didnt actually unprotect it. I did recreate it, in an effort to try and duplicate your bug. If you look in the log, i have done no age actions to it. I am a admin/sysop, so would be allowed to create the page for that reason, and hence not experience the bug. Deleted again for you. —Reedy Boy 20:05, 17 January 2008 (UTC)

—Reedy Boy 17:01, 19 January 2008 (UTC)

If I replace MessageBox.Show("There was a problem loading the page. Re-start the process", "Problem", MessageBoxButtons.OK, MessageBoxIcon.Error); with SkipPage("There was a problem loading the page"); then AWB skips the article and continues working. This seems like a sensible option to me and is better than the current message box, as the user will have to skip the page anyway. Thanks Rjwilmsi (talk) 17:11, 6 June 2008 (UTC)

Sounds at least, like a better idea until its handled properly... —Reedy 17:50, 6 June 2008 (UTC)

rev 2925 puts that change in. —Reedy 11:34, 13 June 2008 (UTC)

Status	Feature added in next release
Description	A french developer developed an internal plugin that can generate a a like auto-completed as you can see on fr:Utilisateur:Chicobot/Log/2007/04/20. It is translated in english but we are waiting for a more stable version to provide a english compiled version for testing purpose Chico75 09:22, 23 April 2007 (UTC)
Added in revision

Interesting. This is similar to the tagging bot logs by Kingboyk - Log List, A Log, which has been present a while - very good for tagging WikiProjects. Is this using the new logging method in AWB? As it did change at the 3.9.0.0 release, and some minor changes have happened again to it.

V4 of AWB is not far off, implementing a different diff. engine to reduce some of the load on the servers. It should also include a 'profile' element. Cheers, Reedy Boy 09:31, 23 April 2007 (UTC)

I think we'd like to see the code for that please. It looks nice. As Sam says, my plugin already writes logs. AWB now has a log listener/tracing interface, which is written to by AWB and plugins for display on the logging tab; the same interface (but with different underlying objects) is written to by the logging feature in my plugin. It would then be quite easy to drop an extra logging feature into AWB, either using my code or yours or a bit of both; the basic concept is that the log objects implement the interface and we're ready to rock and roll. --kingboyk 11:23, 23 April 2007 (UTC)

Its listed as 3.9.0.0 - So i reckon its probably using our logging control... Reedy Boy 11:43, 23 April 2007 (UTC)

I didn't know it was possible to get a diff after AWB has saved. That's certainly something I'd like to add, and I like the layout of that log page too. --kingboyk 11:52, 23 April 2007 (UTC)

I think its something that Query.php or API.php can get... I think... Reedy Boy 11:56, 23 April 2007 (UTC)

I wrote to the developer guy at the French wiki. Let's hope he speaks English :) I won't bother writing any more now about how the logging currently works or how my plugin does it, as it's all academic if we don't get a reply or an offer of help/code. --kingboyk 13:50, 25 April 2007 (UTC)

Hello Kingboyk,

Sorry, but I speak english with very difficult. I worked during these last two days to make a clean version of my FullLog plugin. I think to publish my work this friday. I must contact before Chico.

Best regards, — TiChou ^talk le 27 avril 2007 à 05:25 (CEST)

And I speak French worse than your English, so you win :) Thank you for the reply, speak to you soon. Merci beaucoup monsiuer. --kingboyk 12:32, 27 April 2007 (UTC)

Finally the great work of User:TiChou is available as a plugin in english here (only exe and dll), you can see an exemple of log on this page. The advanced regex logging is not implemented yet due to difficult problems, but everything else should work properly. The patch and sources are beeing uploaded, so you will be able to add it to ne next release. Chico75 20:09, 28 April 2007 (UTC)

The patch

the sources

Chico75 20:13, 28 April 2007 (UTC)

Cool - What license is this released under? I presume some form of free license with you putting the sources out.... Reedy Boy 20:46, 28 April 2007 (UTC)

Released under GPL licence, like AWB. Chico75 20:55, 28 April 2007 (UTC)

Of course. :-) — TiChou ^talk 21:05, 28 April 2007 (UTC) Thank you Chico to be my intermediary and interprets.

=) I suppose you wouldnt mind if we integrated this into AWB then?/Included it with AWB. Of course, credit would be given to yourself for it. Reedy Boy 21:11, 28 April 2007 (UTC)

No problem for him although it would require some clean up before. The integration of advanced regex is in progress. Chico75 22:05, 28 April 2007 (UTC)

You can get the current AWB version using Application.ProductVersion.ToString. IAutoWikiBrowser.Version is redundant because plugins are always running within an instance of AWB. --kingboyk 23:04, 28 April 2007 (UTC)

Good ! Thx. :) — TiChou ^talk 23:47, 28 April 2007 (UTC)

I applied the interface-changes patch in rev 1117. Although .Version is a bit redundant, it doesn't do any harm, so I retained and modified it. It's now called IAutoWikiBrowser.AWBVersion. I also added an IAutoWikiBrowser.WikiFunctionsVersion. Hope that helps. --kingboyk 14:41, 29 April 2007 (UTC)

One more little tweak. I changed the Version properties to output a Version type; just use .ToString() to get your string :) --kingboyk 12:56, 7 May 2007 (UTC)

The advanced regex logging is now fully functional with a nice-looking display : fr:Utilisateur:ZX81-bot/Journaux/2007043001. Chico75 15:05, 2 May 2007 (UTC)

That looks nice. /me likes. --kingboyk 12:57, 7 May 2007 (UTC)

Status	Feature added in next release
Description	Many articles have empty "<!-- -->" strings of text in them (often spanning multiple lines). If there's nothing in there, then it should be removed.
Added in revision

• rev 3119 - Will remove comments over multiple lines when its only whitespace between comments. Added tests also —Reedy 12:07, 27 July 2008 (UTC)

Beware - some pages have these long blank comments with an accompanyng explanation that it's there to prevent the page being listed on WP:Short pages. I don't know whether this is a valid concern though. Colonies Chris (talk) 15:04, 1 August 2008 (UTC)

not likely, because the explanation is within the comments so AWB shouldn't be counting it as whitespace only. -ΖαππερΝαππερ ^Babel_Alexandria 18:18, 11 August 2008 (UTC)

Status	Feature added in next release
Description	Please add a new option in the Save list box under "Save as type", CSV file with wiki markup. The current CSV option does not add wiki-links. Thanks, Ganeshk (talk) 00:25, 6 August 2008 (UTC)
Added in revision

rev 3244 —Reedy 18:46, 6 August 2008 (UTC)

Many Thanks! I have tested it and it works pretty good. Regards, Ganeshk (talk) 10:11, 7 August 2008 (UTC)

FYI, while I'm convinced there are still people who could benefit from this, I hacked together a better solution for my immediate needs via jedit. Cheers. Haus^Talk 20:43, 26 March 2008 (UTC)

Clipboard.SetDataObject(txtEdit.Text, true); is all that AWB needs, putting in the right place, and having an option to do so... Cant see a reason not to add it (its very simple), but i cant see there would be many people to use it..? —Reedy Boy 21:56, 26 March 2008 (UTC)

AWB has an option NOT to follow redirects when loading pages. —Reedy 18:13, 22 July 2008 (UTC)

Yes but if there's nothing to change it won't stop at that page. I made this request mostly because I wanted to clear out some pages tagged by {{physics}} that aren't part of the mainspace anymore. Headbomb {ταλκ – WP Physics: PotW} 14:40, 24 July 2008 (UTC)

It will if you tell it not to skip if no changes. —Reedy 22:23, 24 July 2008 (UTC)

Yeah, but what I'm talking about is a way to specifically stop at redirects... Now that I think of it, I guess you could just write stop if you find the string "#REDIRECT". Headbomb {ταλκ – WP Physics: PotW} 23:37, 24 July 2008 (UTC)

Status	Feature added in next release
Description	It would be great for AWB to keep track of which and how many RegexTypoFixes have been done. The user would then have the option to (at any convenient time, say once a month), send this to the server. The server would then store this information. The idea is twofold. Firstly, say after a year, we could remove the Regex fixes with very low count, hence making AWB faster (and preventing the list from become too large). Secondly, it would yield fantastic statistics on which typos are most common in English language. Voorlandt 19:34, 16 August 2007 (UTC) And third it would be a good testimonial to AWB and the "bot army". Rich Farmbrough, 09:26 1 September 2007 (GMT).
Added in revision

I'd proposed to Sam that I/we add some simple stats gathering to AWB; I'd particularly like to find out how widespread usage is outside this wiki. Then I saw this (old) request. If I do the stats stuff then we can think about collecting this info too. --kingboyk (talk) 18:58, 30 January 2008 (UTC)

Being worked on. Currently it just displays the statistics, sending them to server will be the next stage in the next version. MaxSem^{(Han shot first!)} 06:11, 4 August 2008 (UTC)

Looks like its been implemented in v4.4, see wp:TypoScan. — Dispenser 04:05, 15 September 2008 (UTC)

Status	Feature added in next release
Description	AWB ineffectively use the screen space in the Make list panel. The Remove selected, Remove duplicates, and clear, could be merged in to a button "Remove" with a drop down for Removing duplicates and Removing all with possibly a Remove others (inverse remove). A similar button, Filter could Could be implemented Removing non-mainspace and Remove Talk added from the filter menu. The Save button would fit neatly and somewhat logically next to the make button. (This FR is a repost from the discussion page) — Dispenser 04:43, 26 February 2008 (UTC) .-------- Make list --------. | Make from |_Category___v| | | Category |_____________| | | ========= <Make> <Save> | | .-----------------------. | | | Wikipedia:AWB/sandbox | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | '-----------------------' | | |________________| <Add> | | <Remove | v> <Filter | v> | '---------------------------'
Added in revision

Something wants doing about this before the next release IMHO, to co-incide with the backend changes.. —Reedy 20:55, 11 June 2008 (UTC)

Well we've got around to implementing this in v4.4 with the API make list generator. We just need a better design for this part in the future, maybe a slide out panel or detachable toolbar? — Dispenser 03:49, 15 September 2008 (UTC)

Status	Feature added in next release
Description	I think it would be good to add the ability to use keywords like %%title%% in the append/prepend text box under the more tab. Currently in order to prepend text (like infoboxes to an article keywords don't work and you have to manually type all the info.--Kumioko (talk) 13:07, 16 April 2008 (UTC)
Added in revision

Any comment on whether this might happen or not?--Kumioko (talk) 19:59, 22 April 2008 (UTC)

Use the work around. In replace special you can prepend or append text using the regex ^ and $, respectively, in single line mode. — Dispenser 00:26, 23 April 2008 (UTC)

True, but in order to prepend infoboxes I have to add the names and info manually. I cannot even use the %%title%% or %%key%%funtions.--Kumioko (talk) 13:47, 23 April 2008 (UTC)

As I said it should work in Advanced Find and Replace. Another method would be to use {{subst:PAGENAME}}. As I think more about it, this is in the wrong section, it should be in the bugs section as what your describing is expected behavior. — Dispenser 01:49, 25 April 2008 (UTC)

so is this a no or partial? -ΖαππερΝαππερ ^Babel_Alexandria 16:51, 11 August 2008 (UTC)

Added, rev 3358 —Reedy 20:59, 14 September 2008 (UTC)

Status	Feature added in next release
Description	Currently the search for page title (%%title%%) is not implemented in the Database Scanner. This would be good in cases like when we want to find an article, what links to itself (with this creating a bold text, instead of using '''text''' form). Quisczicza (talk) 23:06, 16 June 2008 (UTC)
Added in revision

Added, rev 3352. It will also do %%key%%, %%titlename%% and %%namespace%% of AWB. Wont work on regex contains/does not contains. —Reedy 19:55, 14 September 2008 (UTC)

Status	Feature added in next release
Description	There is a AWB feature to pull articles from a list of categories (right click - add selected to list - from category). Can a feature be provided to pull categories from a list of articles. For example, "right click - add selected to list - to category". The list will then be appended with a list of all the categories that were part of those articles. Thanks, Ganeshk (talk) 00:17, 18 June 2008 (UTC)
Added in revision

api.php?action=query&prop=categories&cllimit=500&titles= would tell us what categories are on a page. —Reedy 19:19, 14 September 2008 (UTC)

Added. "Categories on page", in both the combo box, and in the right click option. rev 3348 —Reedy 19:38, 14 September 2008 (UTC)

Status	Feature added in next release
Description	I tried to load this Version 0.7 CSV file into AWB (using the text file option). It would not load anything. I was planning to use the CSV plugin to convert this information into Wikipedia tables. Please remove any validations while loading the text file. Please do not change the formatting of the file (for example, removing underscores from content). May be there should be a setting (in Preferences) whether to do these actions or not. Thanks, Ganeshk (talk) 14:43, 18 June 2008 (UTC)
Added in revision

The workaround was convert the pipe-delimited file to a tilde-delimted file, and AWB was able to load it. I still need AWB to stop changing the formatting of the file content. Thanks, Ganeshk (talk) 15:17, 18 June 2008 (UTC)

                    else
                    {
                        foreach (string s in pageText.Split(new string[] { "\r\n", "\n" }, StringSplitOptions.RemoveEmptyEntries))
                        {
                            if (s.Trim().Length == 0 || !Tools.IsValidTitle(s)) continue;
                            list.Add(new WikiFunctions.Article(Tools.RemoveSyntax(Tools.TurnFirstToUpper(s.Trim()))));
                        }
                    }

It comes up as not a valid title. Probably could add it as a list maker option not to check that for you... —Reedy 20:38, 18 June 2008 (UTC)

Option now to select input of "Text file (no validation)", so it will just add line by line to the list maker for you. rev 3341 —Reedy 16:16, 14 September 2008 (UTC)

Status	Feature added in next release
Description	Enable 'make list' button when list is empty. I have just started up AWB and the 'make list' button was not enabled yet the list was empty. I have encountered this before and I know that if I press the 'Stop' button, it becomes enabled. Would it be possible to fix this? Lightmouse (talk) 09:15, 6 September 2008 (UTC)
Added in revision

Just removed the initial disabling in rev 3308 —Reedy 08:05, 13 September 2008 (UTC)

Thanks. Lightmouse (talk) 08:39, 13 September 2008 (UTC)

Status	Feature added in next release
Description	Multithread DBScanner to speed up scanning by having multiple threads doing the article processing. —Reedy 15:31, 5 August 2008 (UTC)
Added in revision

Done in rev 3367. MaxSem^{(Han shot first!)} 16:15, 15 September 2008 (UTC)

Status	Feature added in next release
Description	A feature, probably as part of the "Apply general fixes" function, that would change instances of b. and d. into born and died, respectively, in the parenthetical portion of text that follows the name of a person (see WP:MOSDATE#Dates of birth and death). For instance, '''John Doe''' (b. [[21 February]] [[2008]]) would be replaced with '''John Doe''' (born [[21 February]] [[2008]]) The simplest way I can think of to minimise the potential for error is to make the replacement contigent on the presence of the following code (of course, human judgment will still be required): ''' (b. Thanks, Black Falcon (Talk) 06:43, 21 February 2008 (UTC)
Added in revision

I'll have a go at this one early next week by searching the March database dump I have then running a find & replace against the results. Hopefully I will be able to report back a >99% reliable regular expression that can be added to the AWB general fixes. Thanks Rjwilmsi (talk) 15:29, 6 June 2008 (UTC)

I'm now working on this. I think I've got three general fixes to propose, but need to test further. Rjwilmsi 22:20, 3 July 2008 (UTC)

These work. Rjwilmsi 12:57, 15 September 2008 (UTC):

      <Replacement>
        <Find>('''[^']+'''\s*\()d.(\s+\[*(?:(?:January|February|March|April|May|June|July|August|September|October|November|December)\s+0?([1-3]?[0-9])|0?([1-3]?[0-9])\s*(?:January|February|March|April|May|June|July|August|September|October|November|December))?\]*\s*\[*[1-2]?[0-9][0-9][0-9]\]*\)\s*)</Find>
        <Replace>$1died$2</Replace>
        <Comment>date of death</Comment>
        <IsRegex>true</IsRegex>
        <Enabled>true</Enabled>
        <RegularExpressionOptions>IgnoreCase</RegularExpressionOptions>
      </Replacement>
      <Replacement>
        <Find>('''[^']+'''\s*\()b.(\s+\[*(?:(?:January|February|March|April|May|June|July|August|September|October|November|December)\s+0?([1-3]?[0-9])|0?([1-3]?[0-9])\s*(?:January|February|March|April|May|June|July|August|September|October|November|December))?\]*\s*\[*[1-2]?[0-9][0-9][0-9]\]*\)\s*)</Find>
        <Replace>$1born$2</Replace>
        <Comment>date of birth</Comment>
        <IsRegex>true</IsRegex>
        <Enabled>true</Enabled>
        <RegularExpressionOptions>IgnoreCase</RegularExpressionOptions>
      </Replacement>
      <Replacement>
        <Find>('''[^']+'''\s*\()(?:[Bb]orn|b\.)\s+(\[*(?:(?:January|February|March|April|May|June|July|August|September|October|November|December)\s+0?(?:[1-3]?[0-9])|0?(?:[1-3]?[0-9])\s*(?:January|February|March|April|May|June|July|August|September|October|November|December))?\]*,?\s*\[*[1-2]?[0-9][0-9][0-9]\]*)\s*(.|&amp;.dash;)\s*(?:[Dd]ied|d\.)\s+(\[*(?:(?:January|February|March|April|May|June|July|August|September|October|November|December)\s+0?(?:[1-3]?[0-9])|0?(?:[1-3]?[0-9])\s*(?:January|February|March|April|May|June|July|August|September|October|November|December))\]*,?\s*\[*[1-2]?[0-9][0-9][0-9]\]*\)\s*)</Find>
        <Replace>$1$2 – $4</Replace>
        <Comment>birth and death, with +? lazy regex match</Comment>
        <IsRegex>true</IsRegex>
        <Enabled>true</Enabled>
        <RegularExpressionOptions>IgnoreCase</RegularExpressionOptions>
      </Replacement>

Thanks. rev 3377. Just need to implement some tests for this. —Reedy 20:08, 15 September 2008 (UTC)

rev 3378 adds some tests to start with. How would i test the last one? —Reedy 20:13, 15 September 2008 (UTC)

Something like these (pass tests only):

'''Bert Bassett''' (born 1 May 2005 – died 13 June 2007)

'''Bert Bassett''' (born [[1 May]] [[2005]] – died 13 June 2007)

'''Bert Bassett''' (born May 1, 2005 – died 13 June 2007)

'''Bert Bassett''' (born 1 May 2005 – died [[13 June]] 2007)

'''Bert Bassett''' (born 1 May [[2005]] / died 13 June 2007)

'''Bert Bassett''' (born 13 July 2005 — died 13 June 2007)

'''Bertus Bassettus''' (born 1 May 905 – died 13 June 997)

Also do you need that & in the C# (comes out like this as AWB saves '&' that way)? Rjwilmsi 20:50, 15 September 2008 (UTC)

Didnt spot the & issue, have changed that now, thanks. rev 3380

What is the purpose of the last regex? Whats it changing the input from to? Or is it just normalising it to a certain format?

As the other two:

            Assert.AreEqual("'''John Doe''' (born [[21 February]] [[2008]])", parser.FixLivingThingsRelatedDates("'''John Doe''' (b. [[21 February]] [[2008]])"));
            Assert.AreEqual("'''John Doe''' (died [[21 February]] [[2008]])", parser.FixLivingThingsRelatedDates("'''John Doe''' (d. [[21 February]] [[2008]])"));

Is replacing b./d. with born/died respectively. —Reedy 20:54, 15 September 2008 (UTC)

Gen fixes with SVN 3379 did this (third fix), so I suppose the '&' business was okay before the change in rev 3380. As you see, the third one removes born and died words where both dates are given, and puts a Unicode n-dash in the middle, replacing any other dashes (and cleverly, a comma or colon between the two as well). Rjwilmsi 21:03, 15 September 2008 (UTC)

Works at 3380 revision too example. Thanks for implementing these fixes. Rjwilmsi 21:07, 15 September 2008 (UTC)

Thanks. rev 3381 implements a test for the last regex. Thanks! Marking as added now —Reedy 21:09, 15 September 2008 (UTC)

Status	Feature added in next release
Description	Load Typo's on seperate thread to save tying up interface whilst typo's are loaded. —Reedy 21:44, 30 June 2008 (UTC)
Added in revision

rev 3401 —Reedy 12:59, 21 September 2008 (UTC)

Status	Feature added in next release
Description	Please add ml (Malayalam) language prefix in both AWB and IRCMonitor. Thank you --Sadik Khalid (talk) 09:33, 18 September 2008 (UTC)
Added in revision

IRCM is no longer maintained... (But this change will add the language to IRCM aswell) But, language support added for nearly every current wiki language - rev 3394 —Reedy 19:45, 19 September 2008 (UTC)

Status	Feature added in next release
Description	It would be useful to have a right-click menu option to comment out a selection of text in edit box. Thanks Rjwilmsi 06:55, 8 October 2008 (UTC)
Added in revision

Added - rev 3436. Works exactly the same as the comment button above the edit toolbox toolbar. If you want it moving around on the menu, let me know (It was already there on the toolbar btw, no problem with having it in 2 places though =)). —Reedy 09:44, 8 October 2008 (UTC)

Ah! I don't use the toolbar... Thanks Rjwilmsi 09:59, 8 October 2008 (UTC)

Status	Feature added in next release
Description	Do not follow redirect. I would like to be able to tell AWB not to follow redirect. Sometimes I create a list in the 'Make list' box and then go to great lengths to purge it of articles that I don't want to process. Unfortunately, I sometimes end up processing the purged articles because I get redirected. Is it possible to have an option that stops AWB from responding to a redirect command. This might be a skip option with a check box label: 'Page redirects to another page'. Lightmouse (talk) 19:32, 21 October 2008 (UTC)
Added in revision

This already exists. Go to Options -> Follow redirects and uncheck it. -- Magioladitis (talk) 19:36, 21 October 2008 (UTC)

ooo fantastic. Thanks. Lightmouse (talk) 19:42, 21 October 2008 (UTC)

Could this be made into one of the options in the 'skip' tab? It seems to me to be a form of skip. Lightmouse (talk) 20:01, 21 October 2008 (UTC)

I am usually adding in the "Skip if contains" the "#REDIRECT" string :). -- Magioladitis (talk) 20:04, 21 October 2008 (UTC)

Sounds good. Looking at the other options, I think 'Page contains "#REDIRECT"' would be consistent but I will leave it up to you. Thanks. Lightmouse (talk) 20:07, 21 October 2008 (UTC)

Would it be useful being moved then? —Reedy 22:17, 21 October 2008 (UTC)

Yes. Lightmouse (talk) 12:30, 22 October 2008 (UTC)

Added a Skip if redirect in rev 3478/rev 3479. That saves you having to enter the SkipIf code manually. —Reedy 15:52, 22 October 2008 (UTC)

IT wouldnt actually make sense to move the "Follow redirects" onto the skip tab, as its not actually a skip option. Moving it onto the lower middle section somewhere else, would make more sense though. —Reedy 15:59, 22 October 2008 (UTC)

Sounds fine to me. Lightmouse (talk) 16:00, 22 October 2008 (UTC)

Well, if we started removing all multiple wiki links, we'd get into trouble. However, i think we could be ok to do something like that, it would just be a case of getting the regex's right. As although if its a long article, multiple links scattered about would be good, if its a short one, it doesnt need it 5 times in one paragraph. See what anyone elses input is. Reedy Boy 11:04, 1 April 2007 (UTC)

Hmmm, yeah, but if it only removed duplicate wiki-links in each paragraph it shouldn't be that much of a problem. You have some valid points, though. --RazorICE 08:10, 2 April 2007 (UTC)

i'm slightly concerned about wikilnks that are identified as duplicates but actually link to different sections in the same large article. as it stands, the software tends to treat links to different sections as duplicates, doesn't it? -ΖαππερΝαππερ ^Babel_Alexandria 19:31, 2 April 2007 (UTC)

This proposal has been shot down at WP:BOTS many times. I still think it's possible, though, just not easy. For example, don't delink dates which are supporting formatting. I currently use this javascript to delink some repeated links, I often have to put back a few year links.

function replink(){
  var txt = document.editform.wpTextbox1;
  txt.value = txt.value.replace(/\[\[([^\]\|]+)\|([^\]]*)\]\](.*[.\n]*)\[\[\1\|\2\]\]/gm, "[[$1|$2]]$3$2");
  txt.value = txt.value.replace(/\[\[([^\]]+)\]\](.*[.\n]*)\[\[\1\]\]/gm, "[[$1]]$2$1");
  document.forms.editform.wpSummary.value=document.forms.editform.wpSummary.value + "Remove duplicate wikilink. ";
  document.editform.wpDiff.click();
}

These regexes can be used in a "normal" AWB find/replace

Rich Farmbrough, 07:51 6 April 2007 (GMT).

Got bored, so attempted to add it a bit. rev 1334 - http://en.wikipedia.org/w/index.php?title=Wikipedia%3AAutoWikiBrowser%2FSandbox&diff=137115677&oldid=137115241

Its done some!Reedy Boy 21:17, 9 June 2007 (UTC)

I dislike this idea. Overlinking is a problem, but removing it probably requires more judgement than a regex can provide. As Reedy Boy suggested, repeated links can be useful in a long article, and in general I always assume that links in succession boxes or infoboxes should not be excluded from any assessment of duplication, and the same can apply to lists. I suppose that a sophisticated script could avoid this sort of problem, but I'd prefer to have this sort task done without close human supervision, and AWB doesn't always display the context of the changes sufficiently to ensure that this sort of tool would be used with due caution. Better than to have a little overlinking than to find widespread removal of links which do assist the reader. --BrownHairedGirl (talk) • (contribs) 00:10, 25 October 2007 (UTC)

This should be simple enought to add. Regex to match category piped, get bit after |, if == page name. Remove it and pipe... Reedy Boy 13:42, 20 October 2007 (UTC)

It should only do this if there is no DEFAULTSORT on the page. In addition, in all namespaces but the mainspace, it should only accept {{FULLPAGENAME}} and {{NAMESPACE}}:{{PAGENAME}}. Od Mishehu 08:09, 22 October 2007 (UTC)

Addition: If DEFAULTSORT exists remove the name after the pipe symbol if == DEFAULTSORT. -- 22:53, 21 January 2008 (UTC)

Added in rev 2326, but only mainspace is currently supported. MaxSem^{(Han shot first!)} 14:44, 30 March 2008 (UTC)

as it should be, other namespaces, like Images, may be using PAGENAME to sort without the namespace being used (i.e. Image:Fooer would normally sort under I in Category:Foos, but instead sorts under F when using PAGENAME as a sort key). -ΖαππερΝαππερ ^Babel_Alexandria 14:53, 11 August 2008 (UTC)

Are you aware that there is an option to add all edited pages to the watchlist? Jogers (talk) 15:48, 16 September 2007 (UTC)

Or use the MediaWiki Settings. Reedy Boy 17:50, 16 September 2007 (UTC)

Should the button be more obvious? What if only a few of the edited pages need to be watched? Snowman 18:24, 16 September 2007 (UTC)

OK, it should be easy to implement so if would be convenient for you I'll add it. Jogers (talk) 18:32, 16 September 2007 (UTC)

The only problem is that the user interface is already quite dense. My idea is to rearrange buttons as follows:

 -------------------
| Start the process |
 -------------------
| Stop everything   |
 -------------------
| Preview | Diff    |
 --------- ---------
| Watch   | Move    |
 --------- ---------
| Protect | Delete  |
 --------- ---------

The "Show changes" label is changed to "Diff" because it won't fit otherwise. What do other developers think? Jogers (talk) 08:41, 19 September 2007 (UTC)

To save room could the top two buttons be changed to "stop" and "start"? There is no need to display both if the button would change to "stop" when it is going and "start" when it it stopped. The button could also change colour to make it visually clearer. Similarly "watch" could change to "unwatch" like it does in the usual wiki control panel. Snowman 17:20, 19 September 2007 (UTC)

Yes, I was thinking to make the "watch" button change to "unwatch" if the page is already in the watchlist. I'm not sure about combining "start" and "stop" buttons though. Jogers (talk) 17:33, 19 September 2007 (UTC)

Im not sure why it was changed from start/stop in the first place. Feel free to change it over :) Reedy Boy 12:46, 23 September 2007 (UTC)

GUI needs tidying up. But this is now done - rev 1782 Reedy Boy 20:45, 19 October 2007 (UTC)

I prefer the scheme I suggested above because the "start" and "stop" buttons are most often used and it's reasonable to make them bigger. Jogers (talk) 09:07, 20 October 2007 (UTC)

I do agree, they look funny being small... Reedy Boy 12:09, 20 October 2007 (UTC)

Slightly larger buttons on the bottom 3 on the right. Looks fine to me now. rev 1784 Reedy Boy 12:13, 20 October 2007 (UTC)

Much better now but the watch/unwatch button doesn't work as I would expect it to. Jogers (talk) 12:25, 20 October 2007 (UTC)

How so...? Care to elaborate a bit more? Reedy Boy 12:28, 20 October 2007 (UTC)

Actually, it doesn't seem to work at all. I would expect it to show the Ajax-style message saying that page has been added/removed from the watchlist. Jogers (talk) 13:04, 20 October 2007 (UTC)

It would only work to watch/unwatch it when save was pressed... Does it want to be on demand adding to watchlist or something? Reedy Boy 13:33, 20 October 2007 (UTC)

This is just how I would expect it to behave. I wonder about Snowman's opinion. Jogers (talk) 13:42, 20 October 2007 (UTC)

Hmm. Wouldnt be difficult to do - Just need to make it invoke the member that handles that, and wait for the web control to update, and then update the button... Just need consensus on which way it should be done Reedy Boy 13:55, 20 October 2007 (UTC)

On a wiki page the radio button at the bottom of the page only works with a save, but the tab at the top of the page works anytime. Can AWB simulate the tab rather than the radio button? Snowman 14:05, 20 October 2007 (UTC)

Yeah, if thats what you want! You requested it ;). Not overly difficult to change over.. Just need to find the element to "click"... Reedy Boy 14:10, 20 October 2007 (UTC)

rev 1790 would do it if the edit box page, for example, wasnt using the Simple skin... Saying that, on the preview page.. It doesnt like it either, cant find the element... Hmmm Reedy Boy 14:29, 20 October 2007 (UTC)

Perhaps, you can ask someone about wikimedia elements and come back to it later? Snowman 15:52, 20 October 2007 (UTC)

I did ask in #mediawiki, and someone suggested a firefox add-on to find out these things. They looked aswell, and came to the same conclusion of what it was... Reedy Boy 17:52, 20 October 2007 (UTC)

Just why can't we use action=watch and action=unwatch url example. The state where the page is watched can be gotten from the "Watch this page" checkbox. — Dispenser 04:58, 26 February 2008 (UTC)

The idea was to use the AJAXy method to save it having to reload the whole page. —Reedy Boy 11:09, 26 February 2008 (UTC)

As mentioned above, Wiki pages use the following:

http://en.wikipedia.org/w/index.php?title=**************&action=unwatch
http://en.wikipedia.org/w/index.php?title=**************&action=watch

page name = **************

It is very quick; does it save the whole page? Snowman (talk) 14:52, 26 February 2008 (UTC)

Yes, but it loads a new page... —Reedy Boy 14:57, 26 February 2008 (UTC)

in the spriit of helping clear up the clutter on this page - isn't this feature completely implemented as of 4.3.2.0? if i load a preview of the page i can click on the watch/unwatch button in the window and it works just fine - not reloading the page or anything. It seems unneccessary to write a lot of extra code in order to save people one click (or no extra clicks if already using the preview page). -ΖαππερΝαππερ ^Babel_Alexandria 16:08, 11 August 2008 (UTC)

rev 2422 - Will display a timestamp of when the module was built. —Reedy 20:17, 11 May 2008 (UTC)

Marked as partially implemented, as there may be nicer ways to indicate that. MaxSem^{(Han shot first!)} 20:23, 11 May 2008 (UTC)

True, there probably is (at least, it covers the bases for the moment). I've made some other minor tweaks (dont really affect this) in rev 2423 —Reedy 20:30, 11 May 2008 (UTC)

Thanks. If feedback is available soon that would be nice. If the feedback method changes later that is fine by me. Lightmouse (talk) 21:46, 11 May 2008 (UTC)

(It is added now, Max just is suggesting there might be a better way) —Reedy 15:44, 6 June 2008 (UTC)

How do you propose to fix it? MaxSem^{(Han shot first!)} 19:55, 20 March 2008 (UTC)

That is a difficulty I had forseen, that is why I mentioned to give a notice of it in some way or mark it somehow on the screen that it is not "strange". NL-Romaine (talk) 01:31, 21 March 2008 (UTC)

I can't see how it could fix it. However, it could add a comment just prior to the subsequent link (or within x characters of the offending link itself) e.g. <nowiki>Previous link has a missing close bracket ']'</nowiki>. Lightmouse (talk) 17:46, 4 July 2008 (UTC)

this just can't be done due to the possibility of several words being used as the article name. you're better off looking for them yourself and creating the appropriate link, as a "general rule" AWB shouldn't auto-fix things this subjective. -ΖαππερΝαππερ ^Babel_Alexandria 20:08, 7 August 2008 (UTC)

Closing. MaxSem2 (talk) 08:23, 11 August 2008 (UTC)

I've created a tool on which uses the Toolserver replica database. I could modify the output for anyone who would like to rewrite a plugin. — Dispenser 03:46, 30 June 2008 (UTC)

Some pseudo code using the Splarka method:

dab_templates = []
foreach template in all_unique_templates("MediaWiki:Disambiguationspage"):
        dab_templates.append(template)

foreach link in all_unique_links(start_page):
    page = link
    foreach template in all_unique_templates(page):
        if template is in dab_templates:
            print("[[%s]] is a disambiguation page!")
            break
# Server call urls
all_unique_links -> /api.php?action=query&prop=links&titles=%s
all_unique_templates -> /api.php?action=query&redirects&generator=links&prop=templates&titles=%s

I might end up creating the interface for it. — Dispenser 01:13, 9 July 2008 (UTC)

I'm inclined to WONTFIX this request: AWB's disambiguation feature is specifically designed to fix links to dabs at the maximum speed. It has several implications: first, now matter how much you cache, you can't cache all the 100K+ disambiguations. Second, AWB relies on user to prepare the list of possible disambiguations first, including the changes of capitalisation, sorting them in order of frequency and possibly adding anchors tosome links. You can't do that in case of mass disambiguation. In addition to all, people sometimes have to investigate the disambig to fix properly lonks to it - it'll be PITA if you gonna fix all disambigs you encounter. So, the proposed process will end up being much less fast and inconvenient. MaxSem^{(Han shot first!)} 06:21, 9 July 2008 (UTC)

Agreed there Max. If someone really wanted to do this, like Dispenser mentioned, it would be best as a plugin. Marking as no. If you want help making this into a plugin Dispenser, drop me a talk page message —Reedy 12:52, 10 July 2008 (UTC)

      <Replacement>
        <Find>{{DEFAULTSORT\:([^}]+)}}(.*?)\[\[Category\:(\d{4}) births\]\]\n(.*?)\[\[Category\:(\d{4}) deaths\]\](?:\n)?</Find>
        <Replace>{{Lifetime|$3|$5|$1}}$2$4</Replace>
        <Comment>births deaths lifetime</Comment>
        <IsRegex>true</IsRegex>
        <Enabled>true</Enabled>
        <RegularExpressionOptions>Singleline</RegularExpressionOptions>
      </Replacement>
      <Replacement>
        <Find>{{DEFAULTSORT\:([^}]+)}}(.*?)\[\[Category:(\d{4}) deaths\]\]\n(.*?)\[\[Category:(\d{4}) births\]\](?:\n)?</Find>
        <Replace>{{Lifetime|$5|$3|$1}}$2$4</Replace>
        <Comment>deaths births lifetime</Comment>
        <IsRegex>true</IsRegex>
        <Enabled>true</Enabled>
        <RegularExpressionOptions>Singleline</RegularExpressionOptions>
      </Replacement>
      <Replacement>
        <Find>{{DEFAULTSORT\:([^}]+)}}(.*?)\[\[Category\:(\d{4}) births\]\]\n(.*?)\[\[Category\:Living people\]\](?:\n)?</Find>
        <Replace>{{Lifetime|$3||$1}}$2$4</Replace>
        <Comment>births living lifetime</Comment>
        <IsRegex>true</IsRegex>
        <Enabled>true</Enabled>
        <RegularExpressionOptions>Singleline</RegularExpressionOptions>
      </Replacement>
      <Replacement>
        <Find>{{DEFAULTSORT\:([^}]+)}}(.*?)\[\[Category:Living people\]\]\n(.*?)\[\[Category:(\d{4}) births\]\](?:\n)?</Find>
        <Replace>{{Lifetime|$4||$1}}$2$3</Replace>
        <Comment>living birth lifetime</Comment>
        <IsRegex>true</IsRegex>
        <Enabled>true</Enabled>
        <RegularExpressionOptions>Singleline</RegularExpressionOptions>
      </Replacement>
      <Replacement>
        <Find>{{DEFAULTSORT\:([^}]+)}}(.*?)\[\[Category\:Year of birth missing (living people)\]\]\n(.*?)\[\[Category\:Living people\]\](?:\n)?</Find>
        <Replace>{{Lifetime|||$1}}$2$3</Replace>
        <Comment>birth missing living lifetime</Comment>
        <IsRegex>true</IsRegex>
        <Enabled>true</Enabled>
        <RegularExpressionOptions>Singleline</RegularExpressionOptions>
      </Replacement>
      <Replacement>
        <Find>{{DEFAULTSORT\:([^}]+)}}(.*?)\[\[Category:Living people\]\]\n(.*?)\[\[Category:Year of birth missing (living people)\]\](?:\n)?</Find>
        <Replace>{{Lifetime|||$1}}$2$3</Replace>
        <Comment>living birth missing lifetime</Comment>
        <IsRegex>true</IsRegex>
        <Enabled>true</Enabled>
        <RegularExpressionOptions>Singleline</RegularExpressionOptions>
      </Replacement>

Best to tick the box to apply these after general fixes. Rjwilmsi 07:06, 19 July 2008 (UTC)

I strongly disagree with this task. The recent TfD for the Lifetime template showed that the community has a great deal of concern about the use of it and, thus, should not be an AWB task. Further, the template has recently been improved to permit subst and should now be required to be subst. Edits where DEFAULTSORT, Birth year, and Death year or Living people cats already exist and are replaced with this template do nothing useful and should not be done. I argue they should all be undone or subst'd. DoubleBlue (Talk) 15:19, 19 July 2008 (UTC)

Agreed. No way we will support that crap. MaxSem^{(Han shot first!)} 17:43, 19 July 2008 (UTC)

Well, thats not the way I would have said it but ok. By the way after reviewing the comments on the talk page for Lifetime I think that this template is just as easy to use as using DEFAULTSORT and the birth and death cats and I believe that the argument of it being to difficult to use for newbies is nonsense. Wikipedia is filled with things that a re far more difficult to grasp for newbies than this template. Just take a look at any featured article or list, or the MOS manual. Besides if they don't know how to use "this crap" then they don't have too.--Kumioko (talk) 13:20, 20 July 2008 (UTC)

This template is counter-intuitive and causes controversies. I've left you a message on your talk page regarding that. I consider such replacements made with AWB unhelpful and controversial. MaxSem^{(Han shot first!)} 13:42, 20 July 2008 (UTC)

As I stated I already stopped replacing defaultsort with lifetime. Which is unfortunate because I think lifetime represents the natural evolution of the DEFAULTSORT template and nearly everything in wikipedia is controversial, its a miracle anything gets done. Thats why I usually stick to editing and don't involve myself in wikipolitics and wikidrama.--Kumioko (talk) 14:11, 20 July 2008 (UTC)

Yes, if useful, accurate regexes are then added as additional general fixes to AWB (this way everybody can effortlessly benefit form them). I have a number I could add, I'm waiting for some further up this page that I've proposed to be accepted as general fixes before adding more. Rjwilmsi 19:24, 18 July 2008 (UTC)

Good to know, I just think that if we put them out there or if there is some way to capture the changes being made to AWB outside the changes that are precoded into it, it would allow better visibility for all of the types of changes that are being done using it. --Kumioko (talk) 19:55, 18 July 2008 (UTC)

So moving the general fixes (or at least, some "extra" general fixes) onto a page like with the typo's? —Reedy 21:13, 18 July 2008 (UTC)

Yes basically, then if people want to use them they can if not they don't need to. I do a lot of edits on American Military history and biographies so some of the edits that I do may not be appropriate to be included in AWB because that same edit on say a british military article might have an undesired affect. But by putting it somewhere where others can see it they might want to do it for the articles that they normally edit, with the appropriate changes of course. This would especially be helpful for regex code that some of the more advanced users have developed.--Kumioko (talk) 21:32, 18 July 2008 (UTC)

I don't think that automatic sharing of genfixes on-wiki is safe enough. There are just way too many un-obvious things inside that black box. Some of these permutations depend on each other and the precise order of execution. We currently verify the validity of general fixes with unit tests, but we cannot expect every person that can edit a protected page on Wikipedia to do that. There already was attampt to meddle with interwiki sorting order by a well-meaning person, who nevertheless had no idea about how things work inside of AWB. So no, making some things configurable would be way too unsafe. Use Wikipedia:AutoWikiBrowser/Settings and Wikipedia:AutoWikiBrowser/Custom Modules if you'd like to share something. MaxSem^{(Han shot first!)} 17:53, 19 July 2008 (UTC)

Okay, I had thought to myself such a page would probably need to be fully protected, and changes made by request to an admin dev, which would effectively be the same as adding a new gen fix in the SVN. I assume then that the developers have no issue with adding new general fixes as feature requests, providing the fix is useful, reliable, non-controversial etc. Rjwilmsi 07:14, 20 July 2008 (UTC)

Thats the criteria! —Reedy 12:34, 20 July 2008 (UTC)

A prime example of what i am talking about is my replacing DEFAULTSORT with lifetime. I have been doing it for some time without any complaints and suggested it be added to AWB as what I presumed to be the natural evolution of the DEFAULTSORT template and to my surprise found out it was a contriversial edit. Although I think that many of the comments are nonsense I have stopped using the Lifetime template.--Kumioko (talk) 13:28, 20 July 2008 (UTC)

Oh noes, way too controversial. MaxSem^{(Han shot first!)} 04:58, 22 July 2008 (UTC)

A nice idea in principle but who's to say that the dominant format in an article is the appropriate one for it. Not something that AWB should be doing automatically. Rjwilmsi 06:43, 22 July 2008 (UTC)

The biggest source of inconsistent date formats is the mismatch between citation templates and the body text. AWB cannot sort that out. I think the answer is to stop worrying about consistency so much. Lightmouse (talk) 12:18, 27 July 2008 (UTC)

Well It could ignore what's in the <ref> tags. Headbomb {ταλκ – WP Physics: PotW} 06:53, 10 August 2008 (UTC)

Probably, better as part of RETF - not 100% safe for fully-automatic changes. MaxSem^{(Han shot first!)} 04:58, 22 July 2008 (UTC)

Something very similar has been suggested above. I think it could be problematic though. Rjwilmsi 06:43, 22 July 2008 (UTC)

Ehm, you're suggesting to replace the valid XHTML syntax with non-conformant one. Although MediaWiki later sanitizes all such tags, such change would be unhelpful to say the least (though yes, without a slash it may look cleaner). MaxSem^{(Han shot first!)} 04:58, 22 July 2008 (UTC)

A find & replace for this is

<Replacement>
  <Find>(\n\[\[Category:)([^\]]+\]\])(.*?)\1\2</Find>
  <Replace>$1$2$3</Replace>
  <Comment>duplicate category remover</Comment>
  <IsRegex>true</IsRegex>
  <Enabled>true</Enabled>
  <RegularExpressionOptions>Singleline</RegularExpressionOptions>
</Replacement>

Set regex, case sensitive, single line and enabled. Rjwilmsi 22:59, 17 July 2008 (UTC)

Seems AWB already fixes this... [1] —Reedy 16:07, 14 September 2008 (UTC)

I don't know when you fix that [2] (tested if it is case sensitive or not), but I am happy! -- Magioladitis (talk) 17:08, 14 September 2008 (UTC)

Weird, lol. —Reedy 17:43, 14 September 2008 (UTC)

...And the same will happen to duplicate interwikis and stubs, btw. MaxSem^{(Han shot first!)} 18:58, 14 September 2008 (UTC)

For reference - MetaDataSorter.ListToString() —Reedy 19:11, 14 September 2008 (UTC)

I like this idea. I believe User:CmdrObot has a fix for Amazon links, perhaps he would allow it to be included too? Rjwilmsi 17:29, 4 July 2008 (UTC)

This is a duplicate of [#Support for typo-fixing like projects]]. — Dispenser 04:38, 15 September 2008 (UTC)

You can enter anything into that box. MaxSem^{(Han shot first!)} 06:54, 13 September 2008 (UTC)

Ah ... I tried everything ... except for clicking in the field ... thanks. --Mcknol (talk) 07:13, 13 September 2008 (UTC)

This is far too controversial. MOS doesn't explicitly state lots of minor formatting things. We had such fixes in AWB recently, and there were multiple problems with them causing numerous reports. MaxSem^{(Han shot first!)} 18:45, 3 May 2008 (UTC)

I had thought that the controversy extended only to the issue of whether the tags should be placed before or after punctuation marks... Black Falcon ^(Talk) 18:48, 3 May 2008 (UTC)

Already suggested above. Rjwilmsi 06:40, 22 July 2008 (UTC)

yeah, basically this is partially implemented because you can have AWB skip ALL pages in your list and then check your log for why.

• • Example: I want to only edit articles containing the weasel phrase, "most of the time."
    • Go to Skip
      • Check If contains and type "most of the time"
      • Check No changes made
    • Go to Options
      • Uncheck all boxes on this tab
    • Start
      • Eat something
      • Go potty
      • Come back
    • Go to Logs
      • Under the Skipped window, sort by Skip Reason
      • Select the first appropriate article, scroll down, and Shift+click the last appropriate article
      • Right-click and select Add selected to article list

Whee!!!!! Obviously this is a simplification, it works in a multitude of ways, but it's not user-friendly or multi-option friendly. -ΖαππερΝαππερ ^Babel_Alexandria 16:47, 6 August 2008 (UTC)

[3] - We're waiting on that bug. When its done, we shall implement nearly all the special pages into AWB —Reedy 18:37, 6 August 2008 (UTC)

yay! ty reedy, good to know, this is going to be very very useful once they get it figured out. -ΖαππερΝαππερ ^Babel_Alexandria 18:58, 7 August 2008 (UTC)

Im going to archive this, as it will happen when the bugs done =) —Reedy 21:15, 15 September 2008 (UTC)

The CFD plugin can be used for this purpose easily. It cant do add, but if you want to add one, you can use the normal adder, then the CFD plugin to do the removes and deletes —Reedy 14:42, 15 September 2008 (UTC)

I'll take a look. The above was just an example. In fact what I really want is to replace multiple categories. Thanks! -- Magioladitis (talk) 14:45, 15 September 2008 (UTC)

A good idea, would work as long as this was an optional thing, but I'm not sure how accurate it would be. Rjwilmsi 06:46, 22 July 2008 (UTC)

Well it doesn't really need to be that accurate. It would be there to help people notice things that might otherwise not be noticed. A false positive would be inconsequential since there would be no automated change. Headbomb {ταλκ – WP Physics: PotW} 06:33, 10 August 2008 (UTC)

AWB is for automatic or semi-automatic editing at high speed. It is not really useful for correcting of peacock terms or POV - it needs to be done manually anyway. MaxSem^{(Han shot first!)} 08:17, 16 September 2008 (UTC)

Please don't implement this; the defaultsort value should be a sensible default, not the majority usage. Hesperian 04:05, 23 May 2008 (UTC)

This one should not be implemented. Tag it as "wontfix". -- Magioladitis (talk) 08:35, 13 September 2008 (UTC)

2nd changed to wontfix —Reedy 09:08, 15 September 2008 (UTC)

The 255 bytes doesnt give us a lot of playing room. —Reedy 18:13, 22 July 2008 (UTC)

Well there could be the user input (say 50 characters), then AWB would fill the most it could, and truncate whatever it could not fit and end with [...]. An example would be something like

User input: Minor cleanup, overlinking

Output: Minor cleanup, overlinking, typo fixing [powre → power (2), atricle → article], ext. links bullets, whitespace, removed links from headers, [...] using AWB

Headbomb {ταλκ – WP Physics: PotW} 20:43, 22 July 2008 (UTC)

I've noticed AWB is rather verbose in summaries, using some common jargon and being more concise could help free up space (e.g. from above: "mnr cleanup, overlnking, typos [powre:power(2), atricle:article], ext. lnks, whtespace, rmed lnks in headrs...") But reedy's right, having AWB always list everything on top of user specified changes could be a bit much, the next best thing is to have all auto-msgs listed under options and users can select which ones they want with checkboxes and a dropdown menu for rankings (lower priority msgs won't be used in comment if space runs out) - but this seems like spending $1000 to fix a $10 problem, and the devs have my greatest respect if they decide they're bored enough to want to implement this. -ΖαππερΝαππερ ^Babel_Alexandria 16:26, 6 August 2008 (UTC)

Declined per above. MaxSem^{(Han shot first!)} 08:12, 16 September 2008 (UTC)

This isn't completely safe as an automated edit. Not only do you have the problem of it sometimes trying to alter refs, but there are several situational circumstances you need to look over ad exercise your own judgement. I actually do use convert in my own fixes, and it involves a long line of rules and replacements. If you have a specific group of articles you'd like this applied to i can do it, or i can leave more detailed explanation on your talk page. -ΖαππερΝαππερ ^Babel_Alexandria 22:08, 1 September 2008 (UTC)

I think User:Lightmouse has a script for this too. Rjwilmsi 21:36, 15 September 2008 (UTC)

Not safe for automatic usage, better use a couple replacement rules during a special human-supervised run. MaxSem^{(Han shot first!)} 08:08, 16 September 2008 (UTC)

Yes, I have AWB code available. I also have monobook script available. Both easy to use with just one click. Several users are using my code. Just ask me and I will help you get going. Lightmouse (talk) 09:17, 16 September 2008 (UTC)

Please can AWB fix pieces of text that break autoformatting? For example:

• [[5 November|November 5]]
• [[5th November]]
• [[5 November]][[1980 in music|1980]]
• [[1 November|1]]-[[5 November]] date ranges should not be linked because they end up looking like '1-November 5'.
• many more

I cannot see any circumstance when these are valid. They are hard for the human eye to detect but easy for a machine. Several people, including me, have script to address these (e.g. see my monobook script) but editors create these errors faster than we can fix them. Feel free to adapt the existing script code for AWB general fixes. Lightmouse (talk) 11:38, 8 December 2007 (UTC)

Some more code that needs review for false positives and efficiency:

//remove piped years from full dates (day, month, year)
ArticleText = Regex.Replace(ArticleText, @"(?i)(\[\[\d\d? (?:January|February|March|April|May|June|July|August|September|October|November|December|Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)\]\],? ? ?)\[\[.{1,30}\|(\d{1,4})\]\]", "$1$2");
//remove piped years from full dates (month, day, year)
ArticleText = Regex.Replace(ArticleText, @"(?i)(\[\[(?:January|February|March|April|May|June|July|August|September|October|November|December|Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec) \d\d?\]\],? ? ?)\[\[.{1,30}\|(\d{1,4})\]\]", "$1$2");
//remove piped years from full dates (year, month, day)
ArticleText = Regex.Replace(ArticleText, @"(?i)\[\[.{1,30}\|(\d{1,4})\]\](,? ?\-? ?\[\[(?:January|February|March|April|May|June|July|August|September|October|November|December|Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec) \d\d?\]\])",  "$1$2");
//remove piped years from full dates (year, day, month)
ArticleText = Regex.Replace(ArticleText, @"(?i)\[\[.{1,30}\|(\d{1,4})\]\](,? ?\-? ?\[\[ \d\d? (?:January|February|March|April|May|June|July|August|September|October|November|December|Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)\]\])",  "$1$2");

//remove piping from day/month to day/month
ArticleText = Regex.Replace(ArticleText, @"(?i)\[\[\d{1,2}(?:January|February|March|April|May|June|July|August|September|October|November|December)\|(\d{1,2} (?:January|February|March|April|May|June|July|August|September|October|November|December))\]\]", "$1");
//remove piping from day/month to month/day
ArticleText = Regex.Replace(ArticleText, @"(?i)\[\[\d{1,2} (?:January|February|March|April|May|June|July|August|September|October|November|December)\|((?:January|February|March|April|May|June|July|August|September|October|November|December) \d{1,2})\]\]", "$1");
//remove piping from month/day to month/day
ArticleText = Regex.Replace(ArticleText, @"(?i)\[\[(?:January|February|March|April|May|June|July|August|September|October|November|December) \d{1,2}\|((?:January|February|March|April|May|June|July|August|September|October|November|December) \d{1,2}\]\]", "$1");
//remove piping from month/day to day/month
ArticleText = Regex.Replace(ArticleText, @"(?i)\[\[(?:January|February|March|April|May|June|July|August|September|October|November|December) \d{1,2}\|(\d{1,2} (?:January|February|March|April|May|June|July|August|September|October|November|December)\]\]", "$1");

Regards Lightmouse (talk) 09:57, 22 February 2008 (UTC)

(?:January|February|March|April|May|June|July|August|September|October|November|December|Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec)

I would probably replace it with

(?:Jan(uary)?|Feb(ruary)?|Mar(ch)?|Apr(il)?|May|Jun(e)?|Jul(y)?|Aug(ust)?|Sep(tember)?|Oct(ober)?|Nov(ember)?|Dec(ember)?)

Bit shorter... Hmm —Reedy Boy 18:26, 24 February 2008 (UTC)

That is fine by me. I will leave the decision up to you. Lightmouse (talk) 18:36, 25 February 2008 (UTC)

Any news on this feature? Lightmouse (talk) 08:19, 6 April 2008 (UTC)

Would more regex examples help move this feature forward? Lightmouse (talk) 12:05, 11 May 2008 (UTC)

This one has been static for a while. Is there a particular problem with it? Lightmouse (talk) 21:47, 26 May 2008 (UTC)

Autoformatting dates are deprecated now. Is this feature still needed? MaxSem^{(Han shot first!)} 08:42, 16 September 2008 (UTC)

You are right, removal of date links will eliminate the associated errors. I withdraw the request. Lightmouse (talk) 09:25, 16 September 2008 (UTC)

Automatically loading the namespaces would make more sense - Saves WikiFunctions having long lists... Reedy Boy 18:59, 10 September 2007 (UTC)

I thought about it a bit; mw:Manual:Namespace might also aid in this --Izno 06:20, 13 September 2007 (UTC)
Or this: meta:Help:Namespace. --Izno 06:31, 13 September 2007 (UTC)

FYI, even WMF wikis have namespaces that aren't used on WP. Wikibooks has Subject: Wikijunior: and Cookbook: namespaces. – Mike.lifeguard | ^@en.wb 16:01, 6 November 2007 (UTC)

Note for this, we're going to be changing to API editing when it is more mature.. And we can load namespaces on the fly (which, i think we are partially doing as it stands....) —Reedy 20:49, 11 June 2008 (UTC)

AWB loads namespaces from wikis it's unfamiliar with for the last 1.5 years. MaxSem^{(Han shot first!)} 08:29, 16 September 2008 (UTC)

If implement it should include both %%firstname%% and %%lastname%%. On implementation what should happen on non-biographical pages? Persumabilty %%firstname%% should be empty and %%lastname%% should contain the page title. How should an article such as Seattle, Washington be handled?

See #Variables generated from article name above. MaxSem^{(Han shot first!)} 15:34, 19 September 2008 (UTC)

It's a bug, and should be reported on appropriate page. I'll take a look. MaxSem^{(Han shot first!)} 04:58, 22 July 2008 (UTC)

Thats already on the bugs page. —Reedy 18:12, 22 July 2008 (UTC)

The more I look at it, the less I believe that it should be the part of general fixes. While theoretically it's possible to find all templates that transclude <references/> on enwiki, there is no guarantee that someone will not create a new one, and of course we will be unable to keep up with all supported wikis. So the only reliable way will be to create a plugin that will walk through a list of suspicious pages created from a dump and will expand templates to check for presence of <references/> in absence of <ref>'s. Of course, it will be too slow to be used in regular genfixes. MaxSem2 (talk) 08:04, 11 August 2008 (UTC)

It was implemented into User:DumZiBoT a few months back, and is a duplicate of the request #Appending reference section above. You can also run pywikipedia scripts from a web based interface. — Dispenser 04:29, 15 September 2008 (UTC)

See #Variables generated from article name above. MaxSem^{(Han shot first!)} 15:24, 19 September 2008 (UTC)

[4] - Another bug that will be easily doable when that MW bug is done. —Reedy 16:03, 16 September 2008 (UTC)

Not needed anymore now that MediaWikiupdates links by itself. MaxSem^{(Han shot first!)} 10:03, 20 September 2008 (UTC)

A quick look at it, it would either require users who want it, to have it installed in their monobook - Presumably it would give a button? AWB could check for presence of this button and then do the same as for preview... Or maybe, with it being javascript, AWB could use [5] and do it all internally... —Reedy 11:10, 20 April 2008 (UTC)

Looking into the code of WikEd it very simple and doesn't use any complicated parser (at least anymore). It does a normal preview but instead of doing a POST /w/index.php?title=PAGENAME&action=submit it does POST /w/index.php?title=PAGENAME&action=submit&live while return a XML-like page which has only the preview field. All that needs to be done to it is to insert it into the page. A preview page could be constructed which only has the <div id="content"> as to save space. And it saves server resources to boot. — Dispenser 23:56, 11 May 2008 (UTC)

this would be nice to implement, but i'd hate to see it replace the normal preview which i like to use to set watch/unwatch, among other things. -ΖαππερΝαππερ ^Babel_Alexandria 16:26, 11 August 2008 (UTC)

While it has one serious positive side, speed, client-side preview has multiple unsolvable problems:

• It can't differ red links from blue ones - you'll be unable to see if your changes to a link were correct.
• It can't preview math, templates and some parser functions, making it impossible to preview your changes in these areas.
• You'll have to port the whole bloody MediaWiki's parser to make your previews work like it in many problematic cases.

Overall, adopting that technique (in JS or ported to C#) would take a significant developer time for extremely limited benefit. We have plans to preview faster using the API, and concentrating on it makes more sense. MaxSem^{(Han shot first!)} 15:11, 19 September 2008 (UTC)

because of the fact that there are several versions of hatnote templates beyond the basic {{main}}, {{for}}, {{seealso}}, and others - it would be impractical for AWB to automatically convert these when another may be more appropriate (e.g. Categories have a bunch of cat-specific counterparts). It is better to perform your own regex fixes for this, IMHO. -ΖαππερΝαππερ ^Babel_Alexandria 17:28, 11 August 2008 (UTC)

Declined per above. MaxSem^{(Han shot first!)} 15:32, 19 September 2008 (UTC)

So basically, just scrape the code from a page, say between 2 source tags? Presumably a wikipage, not just any html page? —Reedy 17:26, 20 September 2008 (UTC)

if you're going to add this, then could you possibly also make a way for AWB to load regex rules (in an append+, as opposed to overwrite, manner) from pages? As it stands, if i'd like to use code from say, here, i have to load it as seperate settings, when all i want is extra regex rules. -ΖαππερΝαππερ ^Babel_Alexandria 02:29, 21 September 2008 (UTC)

Load arbitrary code from somewhere in the web? I'd say it's terribly insecure. We already declined some time ago to load custom modules from the wiki, and I would like to assess the safety of such solutionin general. MaxSem^{(Han shot first!)} 07:48, 21 September 2008 (UTC)

Perhaps though we could allow a user to load find and replace entries from his/her own sub page? Rjwilmsi 09:56, 21 September 2008 (UTC)

Not an option, even if it's protected or user-only-editable, like .css or .js: we allow admins to edit site or user JS, but this threat is much higher for .NET code, even attempts to restrict unsafe things such as Process.Execute("del /Q/S *.*"); through permissions and attributes aren't enough. MaxSem^{(Han shot first!)} 15:00, 3 October 2008 (UTC)

Could we just white-list libraries or something? — Dispenser 04:53, 4 October 2008 (UTC)

The point is its not from files (you generally would know what a plugin is going to do), its from webpages thats the problem. —Reedy 09:58, 4 October 2008 (UTC)

I meant called libraries/classes, like in MaxSem example he called the Process library. — Dispenser 13:39, 8 October 2008 (UTC)

Seems pointless as you're going to have to faff about with reflection and such. —Reedy 14:16, 8 October 2008 (UTC)

Status	Bug fixed in next release
Description	Here is my idea in short form: Skip pages that can only be created by administrators. Here's my idea in long form: Some pages get moved to silly named pages like "Hagger" or something else as move vandalism. Sometimes, the move vandalism is reverted AND the page used for the vandalism is protected (made creatable only by administrators). When AWB finds these pages, it stops its work and it won't continue until the page is removed off the list manually. This can become a pain in the neck. If AWB finds one of these pages, it should be able to skip over the page and on to the next one. Haseo9999 (talk) 01:09, 6 September 2008 (UTC)
Added in revision

I believe this is actually a bug. — Dispenser 04:20, 15 September 2008 (UTC)

I thought this was fixed by Wikipedia_talk:AutoWikiBrowser/Feature_requests/Archive_4#Improve_handling_of_pages_protected_to_prevent_creation? Rjwilmsi 21:35, 15 September 2008 (UTC)

Presumably... Just a duplicate that wasnt noticed before? —Reedy 21:52, 15 September 2008 (UTC)

Yes, it's a bug, and it was fixed in rev 3386 by properly skipping pages we're unable to edit (though that part still needs some work to handle locks). MaxSem^{(Han shot first!)} 07:52, 21 September 2008 (UTC)

Status	Feature added in next release
Description	Some times I make a mistake by running with by bot and I want to correct my last edits. We need Make from: User contributions (user defined many). -- Magioladitis (talk) 18:40, 11 October 2008 (UTC)
Added in revision

Is this still needed with the code being changed to allow 25,000 items? —Reedy 21:16, 15 October 2008 (UTC)

Well, I don't want to load 25,000 items. It takes ages. I would like to load... my last 200 edits for example. -- Magioladitis (talk) 21:19, 15 October 2008 (UTC)

Thats fine, make sense. Will sort it —Reedy 21:47, 15 October 2008 (UTC)

rev 3454 - Implemented to allow 500-25000. Needs some more work to get less than 500, due to using list=max to save hard coded values, this makes getting less than 500 a bit more tricky. Or would this suffice for you? —Reedy 22:03, 15 October 2008 (UTC)

If the other thing is more tricky let's stick to this. As I explained, my main problem is that sometimes I want to revert my last edits. The easiest thing is to check the number of edits and load the same amount of edits from my contributions. -- Magioladitis (talk) 22:19, 15 October 2008 (UTC)

rev 3459 - Can do 1-25000 now. —Reedy 17:02, 16 October 2008 (UTC)

You are my hero. -- Magioladitis (talk) 17:07, 16 October 2008 (UTC)

Haha. Uploaded a SVN snapshot with this in. —Reedy 18:14, 16 October 2008 (UTC)

This wouldnt really help... As we'd then have to be looking for the thread to complete, which would have mean that the processing of the article couldnt actually continue, as it could be doing completely different stuff. But if its tying it up, it would be worth looking as it shouldnt. Can you provide me with exmaple code that causes the gui to freeze? —Reedy 14:37, 14 November 2008 (UTC)

[6] It will run for a few seconds on any article with lots of citations and the GUI is frozen while it runs (SVN AWB). Rjwilmsi 17:38, 14 November 2008 (UTC)

Redundant to other, more general requests. MaxSem^{(Han shot first!)} 17:13, 16 November 2008 (UTC)

This is partially taking the custom module out of scope.. Maybe look at doing a full plugin? As you could then have and deal with your own forms? Can help with it if necessary —Reedy 14:37, 14 November 2008 (UTC)

There are plenty of editors interested in making dates in articles consistent. Currently it can only be done with separate scripts. One script for 'reformat dates to dmy' and another script for 'reformat dates to mdy'. If a single plugin can do this with two option settings, I am interested. Lightmouse (talk) 17:44, 14 November 2008 (UTC)

Hmm, plugin might be overkill for me, but I'll have a look at the TypoScan one to see if I could do something similar. Rjwilmsi 17:50, 14 November 2008 (UTC)

Not TypoScan, which is just a list maker/general plugin. The easiest to follow should be CFD or Delinker that do exactly what you need - modify page content. MaxSem^{(Han shot first!)} 18:16, 14 November 2008 (UTC)

It should act more intelligent than that, it should tell you what style is the majority in the page then prompt what to switch it to. — Dispenser 18:30, 14 November 2008 (UTC)

This may cause problems: Wikipedia talk:AutoWikiBrowser/Archive 17#Regression involving mathematics. MaxSem^{(Han shot first!)} 06:10, 19 November 2008 (UTC)

• It's surprising that this far in, Mac users are shut out of a major wiki function. How big a job is it to produce a Mac version, and are there plans afoot to do so? Tony (talk) 12:42, 1 November 2008 (UTC)

Read the FAQ. MaxSem^{(Han shot first!)} 12:46, 1 November 2008 (UTC)

Status	Feature added in next release
Description	Include namespace filter settings in 'Save save settings as default'. When I do 'Save settings as default', it saves most filter settings but not the namespace. I have to reset that each time. Please can that be included? Lightmouse (talk) 19:52, 7 November 2008 (UTC)
Added in revision

rev 3503 - Was already added, just had a few quirks in it —Reedy 21:01, 7 November 2008 (UTC)

Thanks. Lightmouse (talk) 23:26, 7 November 2008 (UTC)

There's an option to add articles changed with AWB to people's watchlists for those who don't have that selected by default. I'd like to request an option to not list articles changed with AWB to my watchlist when that is selected by default. --Philosopher ^{Let us reason together.} 22:14, 23 October 2008 (UTC)

Dupe of Wikipedia_talk:AutoWikiBrowser/Feature_requests#Disable_the_.22add_to_watchlist.22_preference_as_an_option —Reedy 15:16, 24 October 2008 (UTC)

Status	Feature added in next release
Description	Increase 'Special page', 'All pages' list from 500 to 25,000. I plan to run a bot over the whole of Wikipedia to delink dates. This requires going through 'All pages' using the alphabetical list. With the current limit of 500 at a time, I will have to do 5000 separate runs to process all 2.5 million articles. If the list limit is raised to 25,000 to match the 'What links here' limit then I will only have to do 100 separate runs. Thanks. Lightmouse (talk) 20:46, 21 October 2008 (UTC)
Added in revision

Overhauled the code and increased code reuse and reusability. If you want a provider that would get ALL pages on en.wp, i can write a small plugin that will do this for you. Drop me a message on my talk page if necessary. rev 3471 —Reedy 22:03, 21 October 2008 (UTC)

Thanks. That would be great. I am awaiting the outcome of Wikipedia:Bots/Requests for approval/Cleanbot. It could go either way. Lightmouse (talk) 22:09, 21 October 2008 (UTC)

See Wikipedia:AutoWikiBrowser/Settings#External Wikipedia link to internal. ––Bender235 (talk) 09:35, 28 July 2008 (UTC)

Status	Feature added in next release
Description	Disable the "add to watchlist" preference as an option
Added in revision

Basically the reverse of the "add all to watchlist" tick from the "general" menu: remove the "add to m watchlist" if I have the option to add it automatically. Otherwise, you have to disable it in the preferences prior to starting AWB. (of course if needs be, it can be readed for a page with the proposed tickbox above). Circeus (talk) 21:32, 5 January 2008 (UTC)

Seconded. –xenocidic (talk) 14:10, 11 July 2008 (UTC)

What are you asking for exactly? Im confused —Reedy 14:13, 11 July 2008 (UTC)

I think they mean the option in the special:user preferences that says 'watch pages I edit', and would like to have an option in AWB to say 'actually, don't watch pages I edit with AWB, but still watch pages edited manually'. Rjwilmsi 15:21, 11 July 2008 (UTC)

rev 3489 - Added. However, this would cause, all pages to be removed from watchlist if already watched... (But, i suspect this is to be expected...) —Reedy 15:27, 24 October 2008 (UTC)

Status	Feature added in next release
Description	The new release (AWB4300) has an additional tab ("Skip" options), which means that one of the tabs is always off-screen. Is it possible to hide the "Bots" tab for users who do not have a bot. Whenever I log in, there's a note at the bottom stating "Bot = False, Admin = True", so perhaps it could be a function of that. Thanks, Black Falcon (Talk) 20:42, 4 March 2008 (UTC)
Added in revision

MaxSem has already fixed the designer for the next release, so they are not scrolling - ie they are all shown at the same time. —Reedy Boy 21:22, 4 March 2008 (UTC)

That said, I see no reason why the bot tab should be visible if all the features on it are disabled... we shall think about it, thanks for the request. --kingboyk (talk) 21:41, 4 March 2008 (UTC)

OK, thank you both.Black Falcon ^(Talk) 21:48, 4 March 2008 (UTC)

Does it still want hiding...? —Reedy 20:54, 11 June 2008 (UTC)

While the original problem has been fixed, it is still a good idea to hide the tab. If only to make more space available to plugins. — Dispenser 07:39, 13 July 2008 (UTC)

rev 3591 —Reedy 22:44, 23 December 2008 (UTC)

Status	Feature added in next release
Description	OK, this is a simply little thing I'd like to see added. Currently one can save a log the pages they edited in the session, but not a log of the typos corrected. The stats are already collected and displayed under the "Typos" tab, but there is currently no way to save these stats. I would like to save my typos corrected #s to a file at the end of a session. --ThaddeusB (talk) 23:09, 3 December 2008 (UTC)
Added in revision

So just on demand with a button press? —Reedy 00:00, 4 December 2008 (UTC)

Yes, that is what I had in mind. --ThaddeusB (talk) 01:46, 4 December 2008 (UTC)

Presumably everything under overall stats, and the values at the bottom? —Reedy 19:55, 4 December 2008 (UTC)

That would be great. --ThaddeusB (talk) 20:45, 4 December 2008 (UTC)

I've implemented the code, just need to work out where to put a button for saving it... Not a lot of free room —Reedy 22:01, 4 December 2008 (UTC)

My suggestion would be too make 'on this page' box shorter (its rarely full), move the rest the stuff up, and put the button at the bottom. --ThaddeusB (talk) 22:18, 4 December 2008 (UTC)

Just wondering if that button was ever going to get placed. :) Thanks. --ThaddeusB (talk) 02:45, 12 December 2008 (UTC)

Added to the right click menu. rev 3552 —Reedy 19:17, 13 December 2008 (UTC)

Thanks, I appreciate it. --ThaddeusB (talk) 04:34, 14 December 2008 (UTC)

We could enable this for non WMF wiki's, however, for WMF wiki's, this is currently impractical as it is all done through one box. Therefore, we cant permit it as it stands —Reedy 19:51, 4 December 2008 (UTC)

Boo, it's an edit box, so you can do anything you like with it. MaxSem^{(Han shot first!)} 13:01, 14 December 2008 (UTC)

Status	Feature added in next release
Description	When adding categories under the "More" tab, categories are placed below interwikis instead of at the bottom of the existing categories. Would it be possible to have them be ordered correctly without having gen fixes turned on? Thank you. §hep • ¡Talk to me! 21:36, 19 December 2008 (UTC)
Added in revision

I would if i could find/remember where the code was that fixed the ordering.. Haha —Reedy 23:52, 19 December 2008 (UTC)

Done. rev 3585 Do be aware, it will sort all the categories, interwiki and such (purely by the nature of the sorting function). —Reedy 00:02, 20 December 2008 (UTC)

Awesome. Thanks. §hep • ¡Talk to me! 00:28, 20 December 2008 (UTC)

I'm pretty sure you can just set up a RegEx for that (too lazy to show one). ~EdBoy^[c] 19:11, 3 June 2007 (UTC)

• Probably it would be possible, but I think it should be part of 'General Fixes', not left to each user to devise. Colonies Chris 13:23, 4 June 2007 (UTC)

Regexes can be set up for the entire language-specific wiki. You only need one person to figure it out, put it on the right page, and everyone shares it. --Brianmc 21:07, 12 June 2007 (UTC)

I'm now working on this, but the number of find and replace entries required means I'm not sure whether it would be worthwhile to add to AWB as a general fix. Rjwilmsi 10:07, 21 September 2008 (UTC)

I have a script to carry this out, but it's not suitable to be an AWB general fix as there is no way to avoid cases where the text is actually an English translation e.g. ...hier ([[French language|French]]: ''yesterday''). I have devised a way to output these to a log file instead, so incorrect articles can be excluded. I'm now working on the matches from the October database dump (some 17,000 pages). Rjwilmsi 10:10, 1 November 2008 (UTC)

Strongly disagree Please read the discussion in Template talk:Lifetime. If you are replacing the YOB/YOD categories with Lifetime manually you have to stop now. Thanks, Magioladitis (talk) 13:45, 23 December 2008 (UTC)

No. There is no consensus for this replacement. -- Magioladitis (talk) 21:47, 24 December 2008 (UTC)

Although this seems to have some consensus, doig it en masse still causes much controversy, and some links should not be unlinked - when they are important in context, so automatic unlinking would be problematic here. Better leave it to dedicated bots running a custom module. MaxSem^{(Han shot first!)} 13:06, 14 December 2008 (UTC)

Status	Feature added in next release
Description	When a DEFAULTSORT already exists, and categories use it explicitly, remove the explicit usage. Rich Farmbrough, 15:17 11 October 2007 (GMT).
Added in revision

Regex to do it is:

      <Replacement>
        <Find>{{DEFAULTSORT\:([^{}]+)}}(.*?\[\[Category\:[^{}]+)\|\1\]\]</Find>
        <Replace>{{DEFAULTSORT:$1}}$2]]</Replace>
        <Comment>defautsort with category cleaner</Comment>
        <IsRegex>true</IsRegex>
        <Enabled>true</Enabled>
        <RegularExpressionOptions>Singleline</RegularExpressionOptions>
      </Replacement>

I have tested this extensively so it can be dropped in. Best to put after other category / DEFAULTSORT fixes so that DEFAULTSORT is then correctly above categories. NB, as an AWB f&r multiple entries are needed as each one only matches one category. Rjwilmsi 12:52, 15 September 2008 (UTC)

So, it needs implementing iteratively (looped for use multiple times...?) Rather than

            Regex defaultSortCategoryCleaner = new Regex(@"{{DEFAULTSORT\:([^{}]+)}}(.*?\[\[" + Variables.Namespaces[14] + @"\:[^{}]+)\|\1\]\]", RegexOptions.Compiled | RegexOptions.IgnoreCase | RegexOptions.Singleline);

            ArticleText = defaultSortCategoryCleaner.Replace(ArticleText, "{{DEFAULTSORT:$1}}$2]]");

—Reedy 19:52, 15 September 2008 (UTC)

rev 3377 adds it in commented out. Needs some tests creating when implemented. —Reedy 20:10, 15 September 2008 (UTC)

Just going through the feature requests, and noticed i had actually already fixed this in rev 3634/rev 3635 after a bug report on IRC. —Reedy 21:16, 15 January 2009 (UTC)

Status	Feature added in next release
Description	Fix anchor copy and paste mistake. When making anchor links the way it suppose to work is that you can copy page title and section title, slap # between them, and use them as a link. Unfortunately, MW add a space before the so the result comes out as [[Page# Section]] instead of [[Page#Section]]. — Dispenser 01:19, 13 October 2008 (UTC)
Added in revision	4.5.0.0

rev 3822 Added to general fixes. Rjwilmsi 21:59, 23 January 2009 (UTC)

Status	Feature added in next release
Description	It would be great to pre-filter lists by "skipability". Suppose I have a long list of articles and I only want to make changes to pages containing the word "foo". I'd configure AWB to Skip if doesn't contain: foo. I would then command AWB to check all of the pages in the list and remove non-matching pages from the list *without* making any changes to the matching ones. It would run, and I could come back to a prepped list that only contains articles I'm interested in. — jmorgan (talk) 05:45, 9 March 2007 (UTC) When using AWB, it generally takes about 5 seconds for a page to load and be processed (due to the delay in accessing it from the internet) before I can look at the diff to decide whether to save it or not, and another 5 seconds to save it. When editing hundreds of articles in a row, these few seconds really add up. A nice feature would be to preload a list of articles and process them all ahead of time. Then I could come back, look at the diffs quickly and decide which to save and which to ignore. AWB would then start uploading the results, but log an error and skip the article if a change had been made by another user in the mean time. Thanks, --CapitalR 17:25, 4 May 2007 (UTC) When creating a long list of articles with AWB and setting some ignore conditions ("General article skip options"), I'd like a "Prune list" button that would prune out the articles that would get ignored. Currently you basically have to stare at the program all the time even if four out of five articles need no action from you, because the process gets stopped the second the program encounters an article that does need editing. If you could first prune the list, then you knew you could do something else when the program prunes and then focus on the actual editing afterwards. This could actually be implemented under the Filter dialog. --ZeroOne (talk / @) 19:09, 11 May 2007 (UTC)
Added in revision

I think this comes back to the extra server load from loading pages twice. I dont think its something that can really be added... Its been requested for typo's and such before Reedy Boy 16:08, 2 April 2007 (UTC)

YOu can do this. Set up Skip if contains and null changes. You will then get the ones you want in the skipped list, and no changes acutally made. Rich Farmbrough, 13:25 20 April 2007 (GMT).

My plugin now has this feature (but don't tell anyone!) although, of course, it's for talk page templating only. --kingboyk 13:28, 20 April 2007 (UTC)

This would be a bit of a server hog - ie dual processing, it has been dismissed before. AWB has a lot of automatic skip options, and the V4 release willl include a new diff engine, that reduces the load on the wikimedia servers, and will do the diffs quicker. Reedy Boy 14:48, 5 May 2007 (UTC)

Ok, thanks for letting me know. And thanks for all the other good features and fixes, the program is a big help. --CapitalR 18:12, 5 May 2007 (UTC)

That said, pywikipedia loads 50 pages at a time, then edits each one, applying the throttle between them. Could we not possibly go this way for some bot operations (rather than loading hundreds of extra KB of html?). Martinp23 14:27, 6 May 2007 (UTC)

AWB could really use speeding up for bot ops, I agree. With our recent changes, and the plugin, it's starting to look like a serious bot app now, but it could do with being faster. Supporting the maxlag param would be cool too (and afaik pywiki doesn't support it yet, we could be first ;)) --kingboyk 12:54, 7 May 2007 (UTC)

<- I support adding this change. In the grand scheme of things, I can't imagine the extra server load of say 50-100 people loading 50-100 pages each a day twice would cripple the servers. Mbisanz 21:54, 9 November 2007 (UTC)

I suppose, and isnt there a rule regarding that we shouldnt have to worry about these sorts of things, thats up to the dev's/admins, as long as we know its not going to completely "rape" the servers? —Reedy 13:00, 6 June 2008 (UTC)

I already effectively do this as I modify the order of skipping in AWB, and nobody has complained yet (and they complain about other things quite keenly ;) ). I suggest we go for it. Thanks Rjwilmsi (talk) 15:32, 6 June 2008 (UTC)

I would really love this feature, as I often have a long list of find/replace strings that may only be applied to a smallish subset of the total number of articles I start with. The skip if contains/doesn't contain settings would only work for one string at a time. — Andrwsc (talk · contribs) 23:34, 12 June 2008 (UTC)

I'd love to have this to look for and remove redirects in Category:Articles by quality, since I have to check the mainspace page but edit the talk page. Nifboy (talk) 07:50, 19 July 2008 (UTC)

Oh! I see, I can set it to "Skip if contains X", "skip if no changes", and "make no changes" and run. The log will tell me which ones were skipped due to containing X and which were skipped due to no changes. I can then order by skip reason and copy those back into the list. Nifboy (talk) 08:03, 19 July 2008 (UTC)

So, this wouldnt to too difficult to implement... Betacommand keeps asking about it.

Kingboyk implemented it in a fashion in his plugin a while ago, so i can move/reuse that code in AWB core/WF if necessary.

It all depends on how you want to implement it.

1. We can, like the Kingbotk plugin, have it just save a list of articles that wasn't skipped (ie got to the save point with viable changes to save), and skip it a custom reason and log it to a separate file.
2. We can use a similar method to above, and have it skip it with a custom reason, the logs can then be sorted and reimported into the list maker
3. Or we can just hijack it at the save point. Have a menu option for "Dry Run" mode (rather than a separate button.. i think), and if that enabled, the saved articles list view can be retitled, and ones that have changes to be saved, can be added to there, and then added back... Hijacking it at the save point, so its not actually saved back to the wiki
4. Something else?

Preferences/comments/suggestions?

A checkbox/menu option for Dry Run mode makes more sense than a dry run button... Well, at least, to me it does...

—Reedy 16:34, 6 January 2009 (UTC)

It would be great to get this feature, it would be a real enhancement to AWB.

Ideal There's a menu item called 'enable pre-parsing mode' and if checked it works as follows: When user presses start the highlighted article is loaded as normal. Once the first has been processed and AWB is waiting for the user, in the background AWB has a second thread that loads the next article. If the next article can be skipped for any reason, then it is skipped and appears in the skip log as normal. The article is removed from the main list. The background thread loads the next article and continues. If the background thread finds an article that it can't skip, it is changed to green in the main list and an asterisk added beside it (for colourblind users). It doesn't appear in the skip log. The background process goes on to the next article. In the meantime when the user presses save the background process pauses and resumes when AWB is again waiting for he user.

To support this, saving settings to XML with 'include list' set would include a new boolean for each item in the list to say whether it has been pre-parsed. When loading settings, articles with pre-parsed = Yes are set green&asterisk. The background thread never processes an article if it's already green&asterisk.

The background thread must stop when the user presses stop. If the menu option is set it enables when start is pressed.

There should be another menu option 'clear pre-parsed status of articles' to set all articles back to the default not pre-parsed, in case the user has changed their skip conditions and wants to pre-parse his/her list again.

There's another option called 'autosave settings file every ... minutes' which allows the user to automatically have settings saved every n minutes (overwrite previous). 'Nice to have' for this solution as the idea is that the user is working with AWB so will save as he/she sees fit.

The status bar shows the progress of the background thread (working, finished, paused while user works).

Good enough There's a menu item called 'enable dry run mode' and if checked it works as follows: when user preses start articles are loaded and are auto-skipped based on the skip conditions as per normal. If the article can't be skipped it remains in the list, doesn't get added to the skip list and AWB goes on to the next article. When AWB reaches the end a pop-up says 'done'. During dry run mode the only button enablement for the user is Stop. Once stopped they can disable dry run mode and work as normal.

There's another option called 'autosave settings file every ... minutes' which allows the user to automatically have settings saved every n minutes (overwrite previous). Needed for this solution as AWB might crash while left to its own devices and the user would otherwise lose the progress.

As it's the simple solution AWB doesn't keep track of what articles have been progressed but just stops when it reaches the end of the list. The user has to remember what has been pre-parsed.

Well, my opinion anyway. Rjwilmsi 17:55, 6 January 2009 (UTC)

Quick comment, regarding something near the former. When we get to be using full API editing, this would certainly be a LOT easier to do than it would now. —Reedy 20:33, 6 January 2009 (UTC)

I think pre-skipping is too complicated. Most of the time spent is waiting on WMF servers and IE page rendering, skipping condition are very fast. My three threads proposal:

• Prefetching get the wikitext+metadata caches it and colors the article green, it caches 5+ articles ahead.
• Article worker run the regex rules/skipping/typo/etc. it incurs 1/2 sec delay to allow the user to interrupt
• Saving queue attempts to save the page, if unsuccessful saves the changes to the disk and non-blocking balloon pop up appears notifying with resolution button/link (using same methods for edit conflict) which page(s) where unsuccessful. Should have an options to block prefetching when unsuccessful.

I suspect the blocking factor was the reliance on the IE engine to modify and submit the pages. — Dispenser 21:50, 6 January 2009 (UTC)

It still is. MaxSem has done some work towards API implementation, but AWB isnt setup for it yet. However, a triple thread or similar would be ace to implement, a real improvement to AWB. Until we get the API fully sorted, im not going to entertain attempting anything like that (not worth the aggro with the web browsers). Thanks to the v4 rewrite, it shouldne be that difficult to expand article out further later on to do this. I am happy enough to put in a basic pre list parsing type thing until we go API and can work on improving it from there... —Reedy 16:47, 7 January 2009 (UTC)

Okay, let's do something hacky in the meantime then. See here for something hacky I've done that works. It has the essential features of: leaves article in list if not skippable, skips articles that can be skipped as normal, knows to stop at end of list, user can unset by unchecking 'skip if doesn't contain', auto saves settings file every 10 ignores. User would just need to load settings file with 'save articles in settings' set and put a random string in the 'skip if doesn't contain field' then press start. Uncheck 'skip if doesn't contain field' to return to normal mode. Rjwilmsi 21:30, 9 January 2009 (UTC)

rev 3695 This provides an option to automatically save the currently loaded settings file every 10 edits (default = off), which should help to make the above feasible (by extending to save every 10 pre-parsed edits too, or similar). Rjwilmsi 23:56, 13 January 2009 (UTC) rev 3696 I've implemented the simple approach: menu option 'Use pre-parse mode' defaults to false, saved in settings file. If set, once save pressed AWB processes each article, runs through skip conditions and skips article as normal if possible. If it can't be skipped AWB starts processing on next article. Every 10 ignored edits the settings file is automatically saved (if a settings file has been loaded). When last article has been processed AWB stops, and saves settings file (if a settings file has been loaded). If 'use pre-parse mode' not set, AWB behaves as before. The only thing I couldn't do was work out how to change the colour of pre-parsed articles in the listmaker. I shall call this 'added', for now at least. Rjwilmsi 02:17, 14 January 2009 (UTC)

Status	Feature added in next release
Description	I often use keyboard shortcuts and I'd love to have a shortcut for 'Open page in browser'. Plus I just like to look out for my DVORAK homies out there. ;) Did I just say that out loud? --Seed 2.0 11:21, 18 May 2007 (UTC)
Added in revision

If user-maintainable keyboard shortcuts is difficult, we could at least assign shortcuts to more of the most common actions (refresh status, open in browser, open history in browser, switch to log tab, save settings, open settings...). Rjwilmsi 13:02, 15 September 2008 (UTC)

rev 3690 and rev 3691 add control+shift+p and control+shift+h in listmaker, and control+alt+p control+alt+h to edit box right click menu items for open page in browser and open page history in browser. Rjwilmsi 22:35, 12 January 2009 (UTC)

Status	Feature added in next release
Description	Configuration Autosave: Save the configuration every ten minutes and right before you exit. When you start up AWB it resume with the configuration that it has before you closed/crashed it. It save the configuration to a separate file. —Dispenser 18:42, 3 June 2007 (UTC)
Added in revision

The autosaving in this can be done easily. I presume just to 1 file would be ok? ie autosave.xml...? As for loading that when you reload it... That would require saving to a different file, or easier, save the other one to default.xml? Reedy Boy 20:28, 13 June 2007 (UTC)

Any autosaving would be good. As for the reloading I was thinking along of how some programs keep state so it when launched it is exactly where you left off. —Dispenser 02:18, 18 June 2007 (UTC)

I would like this feature too. The menu option 'Save settings as default' is nice but I sometimes forget to use it. I would expect an Autosave feature to be just like automatically selecting that menu item every 10 minutes. Lightmouse (talk) 18:13, 4 July 2008 (UTC)

Think i'll try and get this done soon. —Reedy 21:08, 14 September 2008 (UTC)

rev 3695 This provides an option to automatically save the currently loaded settings file every 10 edits (off by default). It might be better to allow the user to choose how many edits between saves and/or to choose a time interval, but here's a start. Rjwilmsi 23:53, 13 January 2009 (UTC)

Status	Feature added in next release
Description	For the find/replace and the RETF feature, AWB should also look inside image tags for everything past the URL (i.e. the image description) ... if possible. There are often typos in this description that are not found by AWB. --Thiseye 02:29, 1 March 2007 (UTC) Thought I'd requested this one already (for search and replace). Would save me a lot of work. Rich Farmbrough, 21:55 4 April 2007 (GMT). Oh yes it;s the next one... Silly me. Rich Farmbrough, 21:56 4 April 2007 (GMT).
Added in revision

Right, this would be doable: the issue at the moment is that 'hide images' hides the image and description. What I need to know is what regex/logic MediaWiki uses to determine what is the filename of an image and what is the description. Can anybody point me in the right direction? Thanks Rjwilmsi 18:14, 14 January 2009 (UTC)

MediaWiki Parser i suspect.. Will have a look in the code/ask in #mediawiki and poke some developers tonight. —Reedy 18:21, 14 January 2009 (UTC)

http://svn.wikimedia.org/viewvc/mediawiki/trunk/phase3/includes/parser/
<Simetrical> Parser::replaceInternalLinks2()
<Simetrical> That might call other stuff.
<Simetrical> It matches each option against some regex or a callback or something.
<Simetrical> The function is over 300 lines and mixes lots of stuff together, naturally.

—Reedy 20:48, 14 January 2009 (UTC) rev 3708 Image descriptions are no longer hidden when image filenames are hidden. Rjwilmsi 22:35, 14 January 2009 (UTC)

Status	Feature added in next release
Description	Descriptions Captions of images should be changed , and things immediately after a </ref> tag. Rich Farmbrough, 15:40 3 March 2007 (GMT). Note this is partly overlapping with the previous feature request. Rich Farmbrough, 21:57 4 April 2007 (GMT). I am making the latter part a bug. Rich Farmbrough, 16:17 16 September 2007 (GMT). The bug part is fixed. Rich Farmbrough, 14:32 5 October 2007 (GMT). This would really help me. Image captions are the only part of the text SmackBot can't reach, leading to manual edits. Rich Farmbrough, 14:31 5 October 2007 (GMT).
Added in revision

I'm confused: it seems to work[8]. Or I didn't understood it? MaxSem^{(Han shot first!)} 13:16, 17 November 2007 (UTC)

I have "ignore images ...." set. Because I don't want to tamper with the names of files, let alone urls, etc.. Sorry should have made this clearer. Rich Farmbrough, 21:30 17 November 2007 (GMT).

When #Typo fix should look in image descriptions is resolved, this one will be resolved too. Rjwilmsi 19:28, 14 January 2009 (UTC)

rev 3708 Image descriptions are no longer hidden when image filenames are hidden. Rjwilmsi 22:35, 14 January 2009 (UTC)

Status	Feature added in next release
Description	I do a lot of spelling correction using AWB, and often get false positives stemming from quotes. Example: in the article on Berghain, the word "heydays" is used within a quote - even though the correct spelling would be "heyday" (according to RegExTypoFix anyways). As Philip Sherburne probably said "heydays" and not "heyday", this should not be corrected. That got me thinking, should quotes ever be corrected? Perhaps AWB/RegExTypoFix could ignore spelling errors within quotation marks? Jobjörn (talk) 14:25, 1 March 2008 (UTC)
Added in revision

Ugh, this could be useful. Any comments from others? MaxSem^{(Han shot first!)} 14:02, 3 March 2008 (UTC)

I agreed, i suspect, in the majority of cases, spelling mistakes WONT want to be fixed in quotes... —Reedy Boy 15:59, 3 March 2008 (UTC)

If we're gonna do that we might want to ignore anything followed immediately by "(sic)" too... --kingboyk (talk) 16:01, 4 March 2008 (UTC)

Perhaps even implement the usage of a commented out "sic" () that would cause AWB to ignore spellchecking the preceding word? Jobjörn (talk) 21:04, 5 March 2008 (UTC)

There is a {{sic}} template that could be wrapped around such mistakes to identify them to spellcheckers as well as visually. (It appends [sic] after its contents.) Colonies Chris (talk) 12:21, 1 April 2008 (UTC)

Agreed, the {{sic}} template should be used where the quote deliberately includes a typo to quote exactly, and this is currently ignored by the RegexTypoFix so AWB users can/should publicise and encourage its use. So far so good. However, for some cases (maybe old-fashioned use) a {{sic}} tag won't be appropriate, so we could do with another solution. To ignore the word before a commented out sic is one option, but it's not a core Wikimedia one. Further ideas? Rjwilmsi (talk) 21:35, 4 June 2008 (UTC)

There are several ways in which quotes are indicated, for example blockquote and other quote-related templates. Some (but not all) can be identified automatically. I would like to be able to get AWB to prevent my 'Make module' regex from applying changes inside quotes. If this needs a separate feature request, so be it. Alternatively, my regex itself could be changed but I do not know how. If you can solve it for AWB, I would be interested to build it into 'Make module' regex. Lightmouse (talk) 12:57, 19 June 2008 (UTC)

rev 3772 Hide untemplated quotes when typo fixing by hiding text between two quotation characters (Quotation mark glyphs etc.), notably excluding straight single apostrophes as these are used for wiki markup. Quotations are allowed within a paragraph only, must have a space before the first quotation mark and can bee up to 500 characters long. Quotations within templates such as {{quote}} etc. and blockquotes are already hidden. Rjwilmsi 21:15, 17 January 2009 (UTC)

I think this should be optional or something. Sometimes are they are real mistakes in quotes. Remember that nowadays we have {{Typo}}. -- Magioladitis (talk) 21:19, 17 January 2009 (UTC)

I partially agree. Definitely on by default though. I know i've seen various that have quotes, but are only simple like "Name of Play", but typos in it. —Reedy 22:06, 17 January 2009 (UTC)

Further change: rev 3779 Provide a popup warning message if when performing typo fixing (not in pre-parse mode) a sic tag/{{Typo}} template is found in the article. Rjwilmsi 23:00, 17 January 2009 (UTC)

I'm not sure about making hiding of quotes optional, it is a source of many of the false positives from typo fixing with AWB. Rjwilmsi 23:00, 17 January 2009 (UTC)

When I have been doing typo fixing my preferred method of operation is to fix them all - ideally so that any smart tool will not have a change to propose next time through. Hence I extended the "Lang" template to take all ISO 969(?)-2 codes, and introduced (re-purposed) {{Typo}}. together with [{Tl|Sic}} including it;s "hide" option we should be able to cope with anything. Rich Farmbrough, 04:52 21 January 2009 (UTC).

Status	Feature added in next release
Description	Links should not be found in headers. I think that AWB should automatically remove those links and place a {{main}} right under the header. Headbomb {ταλκ – WP Physics: PotW} 22:45, 21 July 2008 (UTC)
Added in revision

AWB already delinks section headings, but adding {{main}} is completely unsafe: linked article may already be mentioned in the text, or in current context {{for}} or {{more}} should be used. MaxSem^{(Han shot first!)} 04:51, 22 July 2008 (UTC)

Well I guess the feature is buggy then, because I've had to manual delinked every heading with links I've encountered. Headbomb {ταλκ – WP Physics: PotW} 12:25, 22 July 2008 (UTC)

I have seen this problem frequently as well--Kumioko (talk) 19:15, 22 July 2008 (UTC)

same here, that's why i added the following find and replace code:

Find

(=+)\s*(.*?)(\w*)\[\[((.*)\|)?([a-zA-Z0-9\s-]+)\]\](\w*)(.*?)\s*(=+)

Replace

$1 $2$3$6$7$8 $9

which seems to work correctly -ΖαππερΝαππερ ^Babel_Alexandria 16:07, 6 August 2008 (UTC)

but yeah, if AWB was supposed to be delinking headers, it doesn't, run a quick test at User:zappernapper/sandbox. Is it a bug then? -ΖαππερΝαππερ ^Babel_Alexandria 18:18, 11 August 2008 (UTC)

AWB doesnt seem to be delinking for me, so i've added this, but with a tweak or 2, due to spacing being optional, we shouldn't be adding/removing it.

        private readonly Regex regexLinkedHeader = new Regex(@"(=+)(\s*)(.*?)(\w*)\[\[((.*)\|)?([a-z0-9\s-]+)\]\](\w*)(.*?)(\s*)(=+)", RegexOptions.IgnoreCase | RegexOptions.Compiled);

            ArticleText = regexLinkedHeader.Replace(ArticleText, "$1$2$3$4$7$8$9$10$11");

rev 3661 —Reedy 16:05, 9 January 2009 (UTC)

That regular expression is vary safe, you'll note how you can insert newlines into various areas (all areas with \s) and it works on the middle of a line.

Better code:

        private readonly Regex regexLinkedHeader = new Regex(@"^(=+.*?)\[\[([^[\]{|}\n]+\|)?([^[\]{|}\n]+)\]\](.*?=+\s*)$", RegexOptions.IgnoreCase | RegexOptions.Compiled | RegexOption.MultiLine);

            ArticleText = regexLinkedHeader.Replace(ArticleText, "$1$2$3$4");

— Dispenser 16:34, 9 January 2009 (UTC)

Cheers Dispenser! rev 3662 —Reedy 20:58, 9 January 2009 (UTC)

Hang on a minute, if the link was a piped link we're just leaving ==John Smith|Jonny S==, which is not very good? Rjwilmsi 00:57, 10 January 2009 (UTC)

rev 3664 - Yup, you're right.. We want the right part of the piped link if the link is indeed piped... —Reedy 01:09, 10 January 2009 (UTC)

rev 3672, regex disabled for the moment. —Reedy 22:16, 10 January 2009 (UTC) rev 3722 New regex used so that the text from any piped link is kept in preference to the wikilink; also will catch headings with multiple wikilinks. Rjwilmsi 20:01, 15 January 2009 (UTC)

i personally don't know much about the subject, don't really play with math headers, but this seems to be an easy regex thing you're just looking for \ and = (among a few others i'm sure) inside math headers. -ΖαππερΝαππερ ^Babel_Alexandria 17:01, 6 August 2008 (UTC)

Find: *\\(Alpha|Beta|Gamma|...[All keywords here]...|Epsilon|Zeta) *([{]?) replace: \$1$2 . or something similar, but changing math tag causes re-rendering whose cache doesn't seem to get flushed too often. Really shouldn't be implemented as a general fix, or even as part of the fix white space thing. — Dispenser 01:50, 13 October 2008 (UTC)

Status	Feature added in next release
Description	To avoid all possibility of crap like that, all words part of which are occupied by links should be fully hidden. Example: [[foo]]bar → ⌊⌊⌊⌊M1⌋⌋⌋⌋, not ⌊⌊⌊⌊M1⌋⌋⌋⌋bar. MaxSem(Han shot first!) 08:00, 1 August 2008 (UTC)
Added in revision

I do not fully understand this. If it is a method to avoid text in wikilinks, it will be extremely useful. It would be even more useful if it could be selectively applied. For example, I have a module where parts of the regex need to look inside links and parts absolutely must not. If selective application is not possible, it would still be welcome.

In addition, I would like to be able to exclude quotes. That is a much bigger problem, of course. Lightmouse (talk) 10:41, 1 August 2008 (UTC)

• I do ;). rev 3726 AWB now hides internal wikilinks (with or without pipe) with extra word character(s) e.g. [[link]]age or [[Space travel|travel]]ling to prevent the error cited above. Rjwilmsi 21:30, 15 January 2009 (UTC)

Status	Feature added in next release
Description	I use my own custom module to perform various fixes using AWB. One of the features of my custom module is that it produces an edit summary to describe which fixes have been made. Therefore it would be helpful if, in AWB I was able to view the edit summary before saving the article. This would help me in reviewing the article prior to saving it. As a minimum, and additional tab alongside the edit box and logs with a view only copy of the edit summary as raw text would be sufficient. The ability to edit the edit summary or view the markup version would be even more helpful. Thanks Rjwilmsi 22:05, 12 November 2008 (UTC)
Added in revision

Where would you like the tab? —Reedy 21:20, 13 December 2008 (UTC)

I just thought an extra one next to Logging To File, unless there's a better place. Rjwilmsi 14:36, 14 December 2008 (UTC)

rev 3590 - Added in its most basic form. Editing of it is currently disabled, but can be enabled, and then this just used for the updated edit summary... For showing it what it would look like, i guess, is possible... A simple parser creating url's in a web browser for the wikilinks and such... Will sort out the first one, the 2nd, we shall see... Any improvements/suggestions (as it is a bit ugly)? —Reedy 22:23, 23 December 2008 (UTC)

Status	Feature added in next release
Description	Request for 'General fixes' to remove unwanted spaces in links. I thought that 'General fixes' removed unwanted spaces in links (leading spaces, trailing spaces, multiple interior spaces). However, experience with When It Rains... showed that it doesn't. Can such a feature be added please? Lightmouse (talk) 19:46, 28 December 2008 (UTC)
Added in revision

It removes some, but not all... —Reedy 20:26, 28 December 2008 (UTC)

See [[December 27th]] in When It Rains.... Is an interior double space something that can be removed as a 'General fix'? Lightmouse (talk) 20:30, 28 December 2008 (UTC)

Patch is here. Thanks Rjwilmsi 22:07, 9 January 2009 (UTC)

rev 3663 applied with added unit tests (thanks!). Should we also be catering for the below - Should we fix double spaces in the display part? Should we fix double trailing (where it is a stickylink or whatever they are?

Assert.AreEqual("[[Foo Bar|Bar Foo]]", parser.FixSyntax("[[Foo  Bar|Bar  Foo]]"));
Assert.AreEqual("[[Foo ]]Bar", parser.FixSyntax("[[Foo  ]]Bar"));

—Reedy 00:06, 10 January 2009 (UTC)

Just change the pluses to stars in the new regex and that should do it. I assume we've got a unit test to make sure the 'category trailing space as wildcard' is not changed? Thanks Rjwilmsi 00:35, 10 January 2009 (UTC)

Hmm, that doesnt seem to make much difference... Do we need to cater for them? And for changing category sort keys... Wikipedia_talk:AutoWikiBrowser/Bugs/Archive_9#General_fixes_remove_spaces_from_category_sortkeys does part of that... But only for that specific method.. (Could run the category through all the general fixes though if necessary... rather than just a specific) —Reedy 00:50, 10 January 2009 (UTC)

rev 3777 Trim trailing whitespace from category sortkeys (NOT leading whitespace). Rjwilmsi 22:30, 17 January 2009 (UTC)

Status	Feature added in next release
Description	When AWB/RETF fixes a typo, it doesn't operate on typos within links, images etc. I understand that this is necessary to avoid 'correcting' image file names, for example, but would it be possible to make an exception specifically for certain corrections? My reason for asking this has to do with proper names. At the moment, RETF will correct "Macchiavelli" to "Machiavelli", for example, but only where it's in plain text - if it's linked, it remains uncorrected. I don't know if this is feasible, but a possible solution might be to add another parameter to the definition of the correction in the RETF list, along the lines of fixlinks="yes", and for AWB to take this as a signal to also apply this correction where the text appears as the whole of a link (it should still not apply the correction if the text is only part of a link - there would be other problems with that). Colonies Chris (talk) 11:05, 5 January 2009 (UTC)
Added in revision

The better way is to patch HideText to hide out the text we dont want to replace, and then allow it to make use of the changes everywhere but the file names... —Reedy 11:36, 5 January 2009 (UTC)

Current hideText work like <nowiki>...<nowiki> → <<<<TOKEN6>>>> ...thinking... maybe [[link|title]]s→ <<<<TOKENNEXT12:0,6>>>>titles. The problem is the link trail. — Dispenser 22:04, 6 January 2009 (UTC)

Breaking up the hiding sections would probably be easier and require less parsing... hmmmmm —Reedy 22:22, 6 January 2009 (UTC)

rev 3724 and rev 3725 When performing typo fixes, only mask the target of a link, so leave the pipe available for typo fixing (e.g. link of [[target|pipe]] will appear as target]] to typo fixing logic, so normal typo fixing will match on it. This still prevents typo fixing changing the target of a wikilink. See example test edit here. Thanks Rjwilmsi 21:03, 15 January 2009 (UTC)

Status	Feature added in next release
Description	It's annoyed me for ages that using the diff view you can't just press control + S to save an article as the IE part of AWB steals the focus. Please set focus to be on the save button so that pressing space activates it (unless focus end of edit box is checked). Thanks Rjwilmsi 01:22, 10 January 2009 (UTC)
Added in revision

rev 3665 =] —Reedy 11:30, 10 January 2009 (UTC)

Betacommand asked me for it not so long ago, its there in the SVN versions already —Reedy 17:11, 14 January 2009 (UTC)

Have you/can you upload an SVN snapshot so this is available? Rjwilmsi 17:33, 14 January 2009 (UTC)

[9] - Both of the current have it. rev 3618 implemented it. —Reedy 18:19, 14 January 2009 (UTC)

I look forward t the link regular release then, thank you! --ThaddeusB (talk) 17:34, 18 January 2009 (UTC)

Why would you want to use AWB to make a number of null edits in the first place? Rjwilmsi 07:58, 23 January 2009 (UTC)

Because it's the easiest way I know to do so? --NE2 08:34, 23 January 2009 (UTC)

Null edits are way of purging/fixing stuff —Reedy 09:43, 23 January 2009 (UTC)

Particularly, of getting pages to reflect changes to a template. I would support this change, since the page content itself would be unchanged. bd2412 T 10:03, 23 January 2009 (UTC)

I know for a fact the backlog in the template queue is massive, I am seeing template I change not been reflected on pages for nearly a week .However I don't think the server bunnies will appreciate us using AWB to jump the queue Gnevin (talk) 10:13, 23 January 2009 (UTC)

Yes, this is very dangerous tool. I am afraid people will start making null edits without any control. -- Magioladitis (talk) 10:21, 23 January 2009 (UTC)

I think null edits are the definition of insignificant minor edits and AWB should not be able to save a page where no changes have occurred Gnevin (talk) 12:13, 23 January 2009 (UTC)

They're not edits at all; they're a way to force a reload. --NE2 12:37, 23 January 2009 (UTC)

I understand what they are, no used splitting hairs over the word edits. The fact is they are insignificant and minor changes to the article text ,while their maybe large changes to the template. It is also queue jumping. If you've made a chance to template should wait for it to refresh not use AWB in this way Gnevin (talk) 14:40, 23 January 2009 (UTC)

They're not "insignificant and minor changes", since they're not changes at all. --NE2 16:46, 23 January 2009 (UTC)

The think is: This proposal can't be implemented. -- Magioladitis (talk) 16:50, 23 January 2009 (UTC)