Wikipedia talk:AutoWikiBrowser/Archive 13

Sadly Martin seems to have left Wikipedia (see User talk:Bluemoose). It looks to be a genuine departure rather than a wikibreak. He'll be greatly missed and I hope he changes his mind. Of more immediate importance though is the future of AWB - it's too great a tool to lose and Martin's hard work mustn't be lost. Has he handed over control of the repository to somebody else? Will you be continuing to develop AWB MaxSem? --kingboyk 14:51, 2 December 2006 (UTC)

It is an unfortunate loss, and a permenant one it would seem from his Talk page. I agree totally with kingboyk, the hard work, and god knows how many hours martin would have put into AWB, cant just be lost.

Reedy Boy

14:55, 2 December 2006 (UTC)

Yes, before leaving Martin granted me admin rights on Sourceforge, and I will continue the development. MaxSem 21:54, 2 December 2006 (UTC)

OK, so "admin rights" will allow you to do such things as assign developer rights? i.e. you have full control? --kingboyk 23:38, 2 December 2006 (UTC)

If he is a project Admin, then yes he can. here shows maxsem as a project admin

Reedy Boy

01:05, 3 December 2006 (UTC)

OK. I'm reluctant to say anything is "excellent" in the circumstances but that at least is good news. Cheers. --kingboyk 02:35, 3 December 2006 (UTC)

I was very sorry to hear of Martin's departure, and hope he changes his mind. I'm glad that AWB will continue though. It's an amazing tool that I have spent many many hours with. But I agree, the Moose will be missed. --Elonka 00:02, 4 December 2006 (UTC)

Martin's departure is really sad. As being a previous hacker on AWB (I did the MWB fork of AWB, from which Martin backported that advanced find and replace dialog), I would be willing to serve as back-up admin on the AWB sourceforge project. My sourceforge login is "ligulem" (in case you want to add me, MaxSem). But I won't and can't replace Martin. --

Ligulem

17:03, 4 December 2006 (UTC)

I've granted you access to SVN. Help from other developers is always welcome. MaxSem 13:09, 6 December 2006 (UTC)

I've uploaded a couple of scripts. They seem to have had this strnge character in them, e.g.:

False" multi="False� single...

(Updated versions coming.) They are OK in MSVC editor, but not in wordpad. Any ideas?

Hi, unfortunately, AWB does not support platforms other than Windows, and porting would require significant rewrite, so it's not likely to happen. MaxSem 06:41, 9 December 2006 (UTC)

Wine (software). Lcarsdata 09:35, 9 December 2006 (UTC)

Did anyone try this in practice? MaxSem 14:32, 9 December 2006 (UTC)

A better idea would probably be VMWare. --kingboyk 20:24, 9 December 2006 (UTC)

I'll give'em a try! Thanks for the suggestions! Cheers,

31415

18:20, 10 December 2006 (UTC)

I'm going to release v. 3.0.6.1 soon, so please enable it on

I created

WP:AWB/CP, but I realized: can AWB follow the transclusion to find out which versions are enabled? —Mets501 (talk

) 02:10, 10 December 2006 (UTC)

No, it can't. MaxSem 06:57, 10 December 2006 (UTC)

MaxSem - It may be worth noting for yourself,

Reedy Boy

16:18, 10 December 2006 (UTC)

The AWB don't clear the memory after save or ignore the page editions; it keep all the text in cache (RAM memory) and for big lists of large pages things not good happen... Giro720 23:17, 10 December 2006 (UTC)

I've noticed it does that for a while, then it seems to clear it and drop down in file size... I dont know if thats some form of garbage collection after x minutes)

Reedy Boy

09:53, 11 December 2006 (UTC)

===[[Luxembourg]]===

Is there some reason that this sort of thing is bad? With 'apply general fixes' it always seems to want to tear these out, and they seem like they work just fine to me. -- Saaber 22:14, 6 December 2006 (UTC)

It is a personal view, but I think that blue headings (or blue and black headings) do not look good. If there are just a few, I look for a the word in the text to link instead. There might be a wiki policy on this, but I am not sure. Snowman 22:26, 6 December 2006 (UTC)

See

Avoid links within headings. 88.110.5.144

14:30, 7 December 2006 (UTC)

This kind of thing is best avoided, but I'd really rather AWB didn't enforce these slightly iffy guidelines. Could it stick to enforcing only totally uncontroversial rules? Thanks. Stevage 12:34, 8 December 2006 (UTC)

What do you mean by "uncontroversial"? It's Manual of Style that should be followed by every editor. MaxSem 13:32, 8 December 2006 (UTC)

The rule says "Avoid". It doesn't say "Do not ever under any circumstances even consider doing". Let human editors decide if and when to remove the link. There is also the problem that the linked text *should* be linked - just not in a heading. As follows:

===[[Luxembourg]]===
The 1997 conference, held in Luxembourg, was...

If this is going to be rewritten, it should be done as follows:

===Luxembourg===
The 1997 conference, held in [[Luxembourg]], was...

Simply naively removing the links in the heading is too simplistic and is generally harmful to Wikipedia. If the worst result of linking headings is that "Depending on settings, some users may not see them clearly", then clearly removing the link altogether is a worse result - in that case *no one* sees the link. Do you see my point? Stevage 02:03, 12 December 2006 (UTC)

I'm not in a place where I can setup and use AWB, so I'll throw this request out there to see if someone else can help. The article Christian Coalition was recently renamed to Christian Coalition of America due to the fact that there are several other groups out there with the name "Christian Coalition". The originall CC page is going to be changed into a disambiguation page. Unfortunately, there are sevral dozen articles still pointing towards the short name page. I'd like to request someone with AWB make a quick sweep of the links to CC and change them to CCoA. This will get things out of the way so I can setup the disambiguation page without messing up a bunch of article links. Thanks. --StuffOfInterest 12:49, 11 December 2006 (UTC)

Im making a start changing direct links Christian Coalition to Christian Coalition of America|Christian Coalition and Christian Coalition to Christian Coalition of America. Based on a google search using AWB. Im not going to get every link, as that would be impossible, but it should help turning it into a disambiuation page, without destroying thousands of links. Hope this helps you out =)

Reedy Boy

17:32, 11 December 2006 (UTC)

Oh, as well as doing regextypo's and the normal AWB changes to the pages that im changing the links on

Reedy Boy

17:33, 11 December 2006 (UTC)

Just FYI

Reedy Boy

17:38, 11 December 2006 (UTC)

Hey, After doing those edits/replaces, 106 edits, 893 listed as ignored, based on googling Christian Coalition

Reedy Boy

20:49, 11 December 2006 (UTC)

Thanks a lot for the help! I'll tackle the disambiguation page on the old site now. --StuffOfInterest 20:55, 11 December 2006 (UTC)

I'm working on uncat stubs with AWB, and there seems to be a flaw (either in AWB or this user! *grin*) When AWB encounters a page that has a {{stub}} tag and no categories, it wants to put in {{Uncategorizedstub|{{subst:CURRENTMONTHNAME}} {{subst:CURRENTYEAR}}}}, even if the article already had an uncatstub tag such as {{Uncategorizedstub|December 2006}}. I noticed that if it just has the {{Uncategorized|December 2006}} tag (ie without the "stub"), AWB doesn't try to add another uncat tag to stub articles.

I assume this behavior is not intentional... Thanks for all the great work to everyone who has worked on it!

Fixed in the next release, see above. —Mets501 (talk) 21:24, 12 December 2006 (UTC)

Many thanks!

Kathy A.

22:42, 12 December 2006 (UTC)

Rewrote this function so it could be used in practice. It gives bloody results like this. Does anyone think that such edits should be tolerated? :) MaxSem 17:40, 8 December 2006 (UTC)

Because this is against

policy, I'm a bit skeptical. It is a cool tool though. —Mets501 (talk

) 22:00, 8 December 2006 (UTC)

Fortunately, this function is extremely slow to use all the time. For example, edit I've used as example took over minute to complete and contents of over 100 articles to load. MaxSem 06:58, 9 December 2006 (UTC)

Bypassing redirects is a fairly complex issue. It should *not* be done when the redirect refers to a more general article. Say there was a small music group, The Foos with four members, one of whom is [John Foo]. Since they're small, [John Foo] actually redirects to [The Foos]. You wouldn't want to bypass that redirection directly to [The Foos] - one day, there may actually be an article about [John Foo]. And there are plenty of examples like this. Stevage 01:58, 12 December 2006 (UTC)

To avoid some problems, I've restricted this function to admins only. MaxSem 11:06, 15 December 2006 (UTC)

When I signed the request list, I signed as "blm07" and now I am on the approved list but I can't use the software. My guess is that I need to be approved as "Blm07" instead, is this correct? --blm07 15:29, 14 December 2006 (UTC)

Fixed. All usernames must be entered in uppercase on the checkpage. --

Ligulem

15:58, 14 December 2006 (UTC)

How would I use AWB to replace something along the lines of "[[Image:X.png|123px|This is a pic of X]]" with "X.png" where the image name, filetype, size and descriptions are all variables? /Lokal_Profil 03:28, 15 December 2006 (UTC)

In the Faye Wong article, AWB removed "th" from [[December 26th]], which is fine. However, it apparently did not detect it in [[May 27]]th and left that as it was.

(Wiki is clever, anyway: I have set my Dates preference to appear the other way round e.g. 27 May, and while [[May 27]] displays as 27 May, [[May 27]]th displays as May 27th, not 27 Mayth.) Fayenatic london 18:22, 15 December 2006 (UTC)

I would find it very useful if I AWB could just remove non-matching pages from the list and leave matching ones. This way fixing rare errors in a large list would be easier — leave AWB to go through the entire list without saving anything, and then come back to work the list that has the few positive matches only. --Hautala 13:16, 16 December 2006 (UTC)

Interesting idea. I'm not sure if it's worth all the resources to load all pages twice, though. —Mets501 (talk) 16:19, 16 December 2006 (UTC)

Absolutely right. Processing even short lists with ~100 items can take considerable amount of time, and the better internet connection you have, the faster you'll process those articles at expence of server load. And processing of huge lists will make Wikimedia personnel mad. There's a better possibility: you can download database dumps and scan them for errors, instead of the live site. MaxSem 16:32, 16 December 2006 (UTC)

lot of "empty" edits occurs after scanning database as we later ignore external links, interwiki, nowiki, graphics, etc., database scanner can't ignore those in its searches

unsigned comment was added by 84.234.42.68 (talk

) 16:46, 16 December 2006 (UTC).

1. bug [3]
2. 3.0.6.1 again moves stub tags on bottom, in 3.0.6.0 it was fixed as on .pl we expect stubs to be before categories
3. split [4] into lines like categories
4. run general fixes before alike after "find and replace"

• first case (general fixes done after "find and replace") - find and replace will fix wrong "15 kwiecień" to "15 kwietnia" [[15 kwietnia|15 kwiecień]] -> [[15 kwietnia|15 kwietnia]], after that general fixes changes it to [[15 kwietnia]] (this option won't let some regexes to match if they contain _ or such characters
• second case (general fixes done before "find and replace") - [[15 kwietnia|15 kwiecień]] will be fixed only to [[15 kwietnia|15 kwietnia]], so to fix this completely 2 edits are needed

-- gregul

1. Fixed.
2. I'm investigating this.
3. These variants aren't absolutely identical, so better leave layout things to human editors.
4. There IS such option ("Apply after general fixes, otherwise before" checkbox on the Find & replace form). MaxSem 16:18, 16 December 2006 (UTC)
   • there is because I was the one who was postulating this, but before OR after is too less as i prooved, could be AND (general fixes running 2 times, always some edits are missed in case of contrary setting) -- gregul

Hi :-) I have some suggestions for improve AWB (cool tool !!).

1.After some changes in a article, while in preview mode, let to copy text (mouse menu or Ctrl+C) from the above window. This would help in restore some part of articles with innecesary replacements.

2.Add an attribute to RegExTypoFix like secure="yes|no" for substitutions that don't need or yes human verification. Then add a checkbox in AWB to use the full list of substitutions or only the secure ones...

<Typo word="à" find="á" replace="à" secure="no" /> 

3.Don't check for typos between « and » nor inside html tags, or add a configuration section/tag on Typo pages, configuring replacement zones.

For example;

<!-- Make changes only in image filenames and template filenames -->
<SearchZone description="Image filenames" start="[[Image:" end="|" > 
<SearchZone description="Template names" start="{{" end="|" > 
   <Typo word="Space to underscore in images" find=" " replace="_" /> 
   ...
</SearchZone>
</SearchZone>

<!-- Make changes anywhere except between  « and » and inside bibliography section-->
<NoSearchZone description="« and »" start="«" end="»" > 
<NoSearchZone description="HTML TAG" start="<" end=">" > 
<NoSearchZone description="Bibliography" start="==(\s*)Bibliography(\s*)==" end="==" > 
<NoSearchZone description="References" start="==(\s*)References(\s*)==" end="==" >
   <Typo word="Xip" find="\bChip\b" replace="Xip" />
   ...
</NoSearchZone>
</NoSearchZone>
</NoSearchZone>
</NoSearchZone>

Joanjoc 23:01, 17 December 2006 (UTC)

Im not sure of the relevance of the secure yes or no. All edits are to be manually checked, unless its a bot account, but that is not allowed to do typo fixing automated. However, is suppose, it could be used in that case, so the bot account could use some changes

The Ctrl + C option for the web browser part would definately be a good idea, i know what you mean, things where its changed things you dont want it to, and have to manually type it out

For the HTML tag not changing links, i believe, that AWB has recently implemented the changes so it didnt change html links and such, but this could still be in the SVN version...

Reedy Boy

10:08, 18 December 2006 (UTC)

Can people with less than 500 mainspace edits join the project? Culverin? ^Talk 02:11, 18 December 2006 (UTC)

Usually, no, but there can be exceptions. Would you like to use it for a specific purpose? —Mets501 (talk) 03:28, 18 December 2006 (UTC)

I would just like to join. But no. I don't have 500 edits. sob, sob, sob. Culverin? ^Talk 23:46, 18 December 2006 (UTC)

Why don't you just wait it out if you want to use it for general purposes. If there was a specific purpose, then perhaps, but for general use, it shouldn't be too much of a problem waiting a bit. :-) —Mets501 (talk) 01:48, 19 December 2006 (UTC)

Hi, Max, would it be possible, as a check thing, that when people are entering custom wiki links into preferences, that if they have included http://, then it isn't prepended

In VB...

If InStr(txtCustomProject.Text, "http://") = False Then txtCustomProject.Text = "http://" & txtCustomProject.Text

But theres no InStr in C#, its the IndexOf... Sure you can work it out =)

Ie if in the text there is no "http://", then add it ;)

Done, by simply removing "http://" from the user's entry with regexes. —Mets501 (talk) 20:31, 18 December 2006 (UTC)

Fair Enough - Thankyou =)

Reedy Boy

20:32, 18 December 2006 (UTC)

Wikipedia:List_of_Wikipedians_by_number_of_edits - Shows that martins bot has done the most edits on en.wiki, over 70,000 more than the next bot

And Bluemoose itself, is ranked in 130th place

I didnt realise Martin did so many edits!!

Hi Max, On the latest SVN commit, DabControl.cs is missing

=)

Just FYI

Uh, me bad:) I always forget to add new files. Fixed now. MaxSem 19:51, 17 December 2006 (UTC)

Phew, thought I had screwed up already :-) —Mets501 (talk) 20:33, 17 December 2006 (UTC)

Hehe - Good you sorted it quickly. The SVN softwre should ask you if you want to add new files to the SVN if they are added after the last commit/update. Thanks

Reedy Boy

21:14, 17 December 2006 (UTC)

Why is the old version disabled every single time a new one comes out? It gets very frustrating for me as I'm (currently) on dial-up, and only use AWB to make a few minor but tedious edits. In the time it to download a new copy I can usually do the edits myself by hand. The old versions shouldn't be disabled for anything but the most critical bugs. +

A lot of the changes aren't critical bugs, but they can be fairly important changes. Between many versions, there are loads of bug fixes and new little feature implementations...

Reedy Boy

10:01, 19 December 2006 (UTC)

I agree, there's absolutely no reason to disable the previous version. If there's a "critical bug" which could damage Wikipedia, fair enough. But for any other bug, let us, the users decide if we want to upgrade or not. Stevage 04:17, 20 December 2006 (UTC)

Quite often these "bugs" will affect less then 1% of the userbase, and some of us don't need nifty cutting-edge features, AWB is great as-is. +

t

) 06:13, 20 December 2006 (UTC)

I understand that edits made with AWB are the responsibility of individual editors, but I have now undone the unthinking capitalisation of guernseys by four different users using AWB. Is there a way to make this sort of error less likely/alert all users to it? JPD (talk) 12:43, 19 December 2006 (UTC)

With it being a relatively small article, you could put a comment on the page like

<!--- Attention AWB Users: All references to guernsey in this article, are supposed to be capitalised as such - Please do not change them --->

Sorry, I should have said that all the edits were actually to Australian rules football. I suppose it might be worth including a comment, but I don't know how often this, or something similar happens in other places as well. Not everything that AWB suggests is an improvement. JPD (talk) 18:15, 19 December 2006 (UTC)

It's a symptom of a growing problem. AWB is implementing more and more of these rules that are right 95-99% of the time. The other 1-5% they're pretty annoying. Some serious discussion on this is warranted. Stevage 04:15, 20 December 2006 (UTC)

I've removed guernsey --> Guernsey from the typo list. —Mets501 (talk) 04:23, 20 December 2006 (UTC)

I seem to keep loosing "subrules" from advanced Find and Replace rules after saving and loading settings. Whenever I load settings the rules are stripped down to just the top level rules, with all subrules missing. After a bit of experimenting it seems like the subrules are never actualy saved in the first place as there is never any sign on them in the config.xml file after saving. This is with version 3.0.6.0. Can anyone comfirm this? --Sherool (talk) 13:25, 21 December 2006 (UTC)

I didn't test this on AWB, but I created that whole rule/subrule thingy for

Ligulem

13:48, 21 December 2006 (UTC)

Fixed. MaxSem 16:44, 21 December 2006 (UTC)

Quick-tested (r562 [5]. Fix works. Thanks. Although that's

    ..
    [System.Xml.Serialization.XmlInclude(typeof(Rule))]
    [System.Xml.Serialization.XmlInclude(typeof(TemplateParamRule))]
    [System.Xml.Serialization.XmlInclude(typeof(InTemplateRule))]
    public abstract class IRule : ICloneable
    ..

not the intention of an abstract class anymore (but who cares :-). Seems to be the price of the new storage design by Martin... --

Ligulem

17:10, 21 December 2006 (UTC)

When did AWB change to sorting interwiki links below stub templates? Was there some discussion and if so, where can I find it? TIA HAND —Phil | Talk 16:49, 21 December 2006 (UTC)

In was there. MaxSem 17:21, 21 December 2006 (UTC)

Is their anyway to get AWB to remove

• Red links

and

• Dates like
  10 January ie dates that are pointless (Gnevin
  22:05, 21 December 2006 (UTC))

So the first idea is feasible, i doubt it for the 2nd though.

Reedy Boy

22:49, 21 December 2006 (UTC)

I thought that some red links are potentially useful. Snowman 23:33, 21 December 2006 (UTC)

Red links definitely serve a purpose; they invite people to create articles. It would be very easy to remove them (they can be matched by the class="new" in the HTML), but in my opinion, they should not be removed anyway. —Mets501 (talk) 23:52, 21 December 2006 (UTC)

Date-linking is a tetchy subject. It would be most unwise to start delinking dates in large numbers without a very very VERY good rationale: this has IIRC been the subject of an ArbCom case in the past and people have been blocked, etc. In any case, the reason for linking dates (like

I basicly edit GAA articles such as Offaly_Senior_Football_Championship and would like to use AWB to remove the linked years . Can some provide he regex for this ? (09:50, 22 December 2006 (UTC))

Also i feel the red links could be removed from

Cork_Senior_Club_Hurling_Championshipwith out much problem but its a secondary concern the dates is the main issue (Gnevin

10:00, 22 December 2006 (UTC))

PPS User:Phil Boswell i'm surprised you can hear us normal users all the way up their on your high horse (Gnevin 10:05, 22 December 2006 (UTC))

Gnevin, I've removed you from the list of approved users per your apparent lack of understanding what AWB is supposed to be used for. Using AWB for mass delinking of dates and redlinks is against consensus. --

Ligulem

11:00, 22 December 2006 (UTC)

So instead of saying AWB cant be used for that and check my contribs with AWB you removed me . Well thats a bit extreme . Wiki has more policy than i care to think about and AWB is part of that huge policy set. I've never heard of someone being banned for asking a question so much for

WP:GAA will alot poorer without a user like me to run AWB on the stub, category etc (Gnevin

11:52, 22 December 2006 (UTC))

Well, if you confirm not to use AWB for

Ligulem

12:37, 22 December 2006 (UTC)

I dont know why you jumped to the conclusion i would use it for mass removing of dates you could get my contribs to see my AWB edits .I was asking the question because i was unsure if it could be done and if it was allowed to be done .So yes i can confirm that i wont use to remove dates .

My reply to Phil was such as their are two ways of asking the user if they know what their doing one assume their an idiot and tell them to go away till they know better while speaking down to user the whole time the phil way . The second is to say how its done and say is that what your doing (Gnevin 12:50, 22 December 2006 (UTC))

I've readded you per your confirmation above. --

Ligulem

12:59, 22 December 2006 (UTC)

How do I add notices (ie: new section, then a template) to a talk page. For example, the way

I think you're looking for the append text option (it's at the top of the second options tag. —Mets501 (talk) 11:55, 22 December 2006 (UTC)

I know this has been touched on before, but I'm getting increasingly concerned about all the useful information that is being wantonly discarded by the automatic removal of links from headings. Many piped links, carefully crafted to avoid dab pages, are simply vaporized by this tool. I know it's recommended that words in titles should not be linked, but it is not exactly an error and currently these actions of AWB degrade rather than enhance the usefulness of an article.

Please can we have a change to this functionality, so that the link is preserved by inserting a line below the title, before any more valuable inter-page linking is destroyed. Euchiasmus 10:10, 23 December 2006 (UTC)

There are some pages (mainly pages of lists) with dozens of links in the headings and the page depends on this, although blue headings do not look very good in my opinion. I think that AWB could look for the words within a heading-link in the paragraph that follows and link the first instance in that text. AWB could only suggest a removal of the link in the heading if it can be replaced in the text that follows. I think that this could be done with regular expressions, some "if..then" lines and some other lines of code. I have only being removing links in headings if I can manually replace the link in the following text, even if I have do to a minor rewording of the text to include the link. Snowman 12:14, 23 December 2006 (UTC)

Nice idea, I'll try to implement this. MaxSem 12:24, 23 December 2006 (UTC)

Perhaps this could include headings that contain more than one link. Snowman 12:49, 23 December 2006 (UTC)

Hi Barberio. Was there an exact point when J.smith and Spartaz were asked to stop, and they refused? (I don't really have time to read the pages and pages of dialogue at the link you provided). I completely trust you, but I need to make sure before I remove their AWB access. —Mets501 (talk) 04:01, 23 December 2006 (UTC)

Yes, there's plenty of discussion on their talk pages, and even biggest problem is that buggy regexes are being used. For example, in this edit youtube link remaint untouched, and valid link to CNN was removed instead. I suggest immediate removal of acess. MaxSem 08:02, 23 December 2006 (UTC)

All right, I'll do so, because MaxSem (a regular here) has requested it and given an example of a bad bot-like edit (akin to blocking an errant bot). Also, it breaks the "Don't do anything controversial with it" rule of use. They can probably be readded to the list when all this calms down. Titoxd^(?!?) 08:09, 23 December 2006 (UTC)

This is ridiculous but hardly surprising from an editor who hasn't properly checked what we are doing and who instructed Dmcdevit to stop deleting youtube links when he hasn't been doing this. Our use of AWB is for list sorting on a project that has been discussed and agreed on AN. I wouldn't mind if people actually looked at my edits and pointed out the ones that are incorrect. I have already modified my bahaviour in response to discussion at EL and there seemed no real objections left over after that. Then a RFC rises that is related to Barrington Hall and subsequently there has been a concerted campaign to criticise my work. I have had nothing to dow ith Barrington Hall and the main complainer does not seem to have taken any time to review what I have done personally. The editor who made the controversial changes to Barrington Hall is not having his use of AWB challenged but I have had it revoked without even a discussion. Strikes me as a case of double standards.

My use of AWB is not controversial. The RFC acknowledges that blatant copyvios cannot be linked to from Wikipedia. Thats pretty much all I'm doing with AWB. Please reconsider or at the very least discuss what I'm doing wrong with me first. That's only fair. --

Spartaz

09:29, 23 December 2006 (UTC)

The issue is not as much the controversial edits, (it's part of it, but not the most important part). The issue is that the copyvio link was left, and a good link was removed, which indicates a problem in the routine being used to remove those links. Again, as I indicated on your talk page, when it is fixed, I see no reason why the access can't be returned. Try asking for suggestions here as to how to fix them. Titoxd^(?!?) 09:35, 23 December 2006 (UTC)

What good link was removed and what copyvio was left. First I have heard of this.

Spartaz

09:41, 23 December 2006 (UTC)

Ok, then you folks are good to go, I've restored the access now. Titoxd^(?!?) 09:43, 23 December 2006 (UTC)

Thanks.

Spartaz

09:45, 23 December 2006 (UTC)

Wow... a comedy of errors. lol Mind actually looking at my contributions or asking us for a copy of our regex before removing access next time? ---J.S ^(T/C) 17:11, 23 December 2006 (UTC)

Sorting people into categories

Hi! Today is my first day as an AWB user. I'm enjoying it very much, thank you for a great piece of software.

However, no matter how Featured an article is, it can always be improved: and so can AWB. My humble suggestion would be fixing the problem illustrated by this diff. AWB accurately wanted to list Jan Björklund in Category:Living People and I agreed. However, AWB missed out on that when categorizing persons with Swedish letters in their names (å ä ö), these should be replaced (by a a o, respectively). Is this a general fix feature we may be able to see in the next release? Jobjörn (Talk ° contribs) 22:54, 23 December 2006 (UTC)

RegExTypoFix

I have been thinking about this subject and have posted some ideas at

^{Talk Contributions}

13:53, 24 December 2006 (UTC)

Login

I launch Microsoft IE and log in under my bot account (DomBot), then launch AWB. I always have to use File>Log in; I am never automatically logged in to AWB if I'm logged in via Microsoft IE. When I choose the "Log in" menu option, I get the "Problem" dialog box.

First, according to the manual and the dialog box, I shouldn't have to do this if I'm already logged in via Microsoft IE. Second, if I've chosen the "Log in" menu option, why should I see anything other than the login page? The box is basically telling me what I already know (that I'm not logged in). While the information on how to avoid having to manually log in in the future is helpful, it seems to be incorrect for me, at least.

Am I doing something wrong? Help?—DomBot ^talk ; Chidom ^talk, owner/operator. 07:39, 18 December 2006 (UTC)

Are you checking the "remember me" box when logging on in IE? —Mets501 (talk) 19:42, 19 December 2006 (UTC)

No; and I'd rather not. If that's required, it should be part of the instructions. Thanks.—DomBot ^talk ; Chidom ^talk, owner/operator. 02:19, 20 December 2006 (UTC)

Update. I just tried that; it doesn't fix the problem. I still have to log in separately in AWB.—DomBot ^talk ; Chidom ^talk, owner/operator.

Have you allowed cookies? ("Tools --> Internet Options --> Privacy" and check the cookie settings) —Mets501 (talk) 03:42, 20 December 2006 (UTC)

My cookies work fine everywhere else (bank sites, credit cards, shopping, etc.); my setting is at "Medium", which blocks cookies using personally identifying information, etc.; I think I'm just resigned to having to log in twice. The pain is that AWB doesn't support RoboForm, either, and I'm lazy. <grin> Thanks for the effort, though.—DomBot ^talk ; Chidom ^talk, owner/operator. 10:54, 27 December 2006 (UTC)

DB Dump Scanner

Hi, Noticed to save list of links in the DB scanner, that you have to create a wikified list...

Also, ive noticed this time that when scanning, when it finds a result, that it is not adding it to the AWB listbox, aswell as its own result list. Whereas it did before...

I only noticed the first thing this time, as the 2nd was working before, and i saved the list in AWB

Any ideas?

Reedy Boy

19:37, 18 December 2006 (UTC)

Oh, i've just noticed, if you do, make from Database Dump, and then scan it that way, it puts it into the listbox in AWB

Reedy Boy

23:34, 22 December 2006 (UTC)

Yeah, I realized that too. I'll take a look, but no guarantees on me fixing it. —Mets501 (talk) 01:15, 24 December 2006 (UTC)

Cool, in theory, a copy and paste of the make list button pressed when the source is DB Dump..? Thanks

Reedy Boy

21:10, 24 December 2006 (UTC)

DB Dump Success Rate

Regextypo on the lates en.wiki dump - first 599 results, resulted in 20 edits...

It seems like the DB scanner is throwing more false positives than it should. Cause i cant believe that such a high proportion of the articles taht the scan pulled to scan on the live wiki, had the problems fixed on them...

Any ideas?

Reedy Boy

20:50, 18 December 2006 (UTC)

I guess it's time to make DB scanner to ignore same things as normal processing does. E.g. comments, templates and so on for some types of fixes. MaxSem 16:50, 24 December 2006 (UTC)

If you would, i and im sure others would appreciate it =)

Reedy Boy

21:12, 24 December 2006 (UTC)

Block Quotes

</blockquote><blockquote>

Can we disable Regextypofix on these areas? Cheers. --

Reedy Boy

11:46, 21 December 2006 (UTC)

Would be easy enough, but do we want that? There can be typos in quotes. —Mets501 (talk) 11:59, 21 December 2006 (UTC)

Well, it depends if its typos from the actual quote - ie if its a written quote, then really, it should stay like that, but if a user has introduced the error, it should be fixed...

Reedy Boy

18:52, 21 December 2006 (UTC)

Hmmm.... maybe we can get a third opinion on this? Anyone? —Mets501 (talk) 19:11, 21 December 2006 (UTC)

The third opinion says: "Make it an option, with a default of not typofixing inside </blockquote><blockquote>" :-). --

Ligulem

19:23, 21 December 2006 (UTC)

Sure, we can make an option for it. But there's already an option creep, users forget to set this option, that option, and as a consequence they produce unexpected edits. Frankly, I'm not sure about that. MaxSem 19:45, 21 December 2006 (UTC)

Yeah, true, but then we get an inconsistency in edits of this... It really needs to be all or none... But all IMHO would give more bad edits...?

Reedy Boy

19:37, 21 December 2006 (UTC)

I'd vote in favor of not making edits inside blockquotes; if there are typos, it should be the responsibility of an editor to check to see if the typo is or isn't part of the quote (ah, for good old [sic] days).—DomBot ^talk ; Chidom ^talk, owner/operator. 10:57, 27 December 2006 (UTC)

API

Error	3	'LoggedIn' is not a member of 'WikiFunctions.Browser.WebControl'.	F:\Work\Programming\AWB Plugin\AWB Plugin\PluginManager.vb	712	16	AWB Plugin

My plugin won't compile because this member, WebControl.LoggedIn, has been removed. Is there any reason why it's gone?

Also, how do these two work, which replaced ReadXML()/WriteXML()?

        Public Sub LoadSettings(ByVal Prefs() As Object) Implements WikiFunctions.Plugin.IAWBPlugin.LoadSettings

        End Sub

        Public Function SaveSettings() As Object() Implements WikiFunctions.Plugin.IAWBPlugin.SaveSettings

        End Function

--kingboyk 13:20, 23 December 2006 (UTC)

1. It was replaced with auth code in Variables.cs. You can use WebControl.UserName to check this: if user isn't logged in, this function will return empty string.
2. SaveSettings should return array of serializable variables/classes. Sorry, I don't know VB, but in C# it should look like this:

[Serializable]
public class MyPrefs
{
....
}

public class MyPlugin : IAWBPlugin
{
....
   MyPrefs myPrefs = new MyPrefs(blah);

   public object[] SaveSettings()
   {
      object[] Prefs = new object[1];
      Prefs[0] = myPrefs;

      return Prefs;
   }
}

MaxSem 13:35, 23 December 2006 (UTC)

Cheers, i noticed that GetLogInStatus() returns a boolean for logged in status, and that looks to be a replacement for what we had currently. So we'll use that for the moment, unless any problems occur. I did notice that Username() would return, good to know thats theres another way to do it!

Reedy Boy

13:52, 23 December 2006 (UTC)

Thanks Max. I wonder if it would be simple to restore ReadXML() and WriteXML() but mark them as deprecated? Or did Martin make some significant changes under the hood? It's not so nice when an interface breaks and I don't see any huge benefit from changing (if it isn't broke etc etc...) If it's an irreversible change that's fine but if it's not I'd appreciate having the old way available too :) --kingboyk 11:18, 24 December 2006 (UTC)

Unfortunately, XmlSerializer class does not support writing custom data, afaik. Only objects to serialize. Config files created by older versions are supportd in read-only mode, and support for outut is fully dropped. So, returning back to XmlWriter means rolling back a lot of code and losing compatibility with new format. As of benefit, it's really nice: you can make an option saveable much more easily. MaxSem 16:44, 24 December 2006 (UTC)

We've got some work to do then Steve! Can't really see it being too bad, just a bit of a faff!

Reedy Boy

21:17, 24 December 2006 (UTC)

Yep! --kingboyk 02:41, 25 December 2006 (UTC)

Multiple wiki-links

When the multiple wiki links list is so long that it needs a scroll bar it jumps to the top every time you hit a link down the bottom can you make it so that the link your selecting stays selected (Gnevin 01:20, 24 December 2006 (UTC))

Fixed, and find/replace improved a bit (it was horrible before!). —Mets501 (talk) 16:12, 24 December 2006 (UTC)

Lock Summary

Hi, Would it be possible for the lock summary check box to be included in saved/default settings?

Ie if i lock the summary, and then save as default, it will be marked as locked for my default settings?

Cheers

Reedy Boy

13:00, 24 December 2006 (UTC)

Wouldn't you want the default to be unlocked, though, so you can set it at the beginning of each task you do? —Mets501 (talk) 16:13, 24 December 2006 (UTC)

Done. MaxSem 16:24, 24 December 2006 (UTC)

For me, i tend to usually only do typo fixing and the other normal stuff, so its my default. For the times where im doing other stuff aswell, its not difficult to change. If people want it unlocked, just make sure it is not ticked when you save default! Cheers Max =)

Reedy Boy

21:09, 24 December 2006 (UTC)

AWB Not Signed In When I Am Signed In

Because my connection is through a satellite, I have real problems staying signed in. Perhaps it's shifting IP or something but to stay signed in, the addresses have to start with https://secure.wikimedia.org/wikipedia/en. What this translates to is AWB often (and recently always), after checking that I'm logged in and finding that true, suddenly finds that I'm not logged in and gives me the signin screen. Unfortunately for me, this is the normal signin, not the secure signin I need to stay signed in. I've tried numerous combinations of opening IE 7 and going to the secure version of the WP main page before loading AWB, loading lists after logging in, loading them before logging in, nothing seems to work. "Custom" under the "User and project prefs" would work except it's locked into "http://" and I probably need an "https://" for it to work. Any suggestions for a work-around would be greatly appreciated. I'd be sad to abandon AWB but it just isn't working well with my situation. Help? --

talk • contribs

17:38, 25 December 2006 (UTC)

AWB won't install

I have been approved for AWB and I have Win2000Pro and .NET framework version 2. When I try to install AWB, I get the following error signature: AppName: autowikibrowser.exe AppVer: 3.0.6.1 AppStamp:4582812d ModName: kernel32.dll ModVer: 5.0.2195.7099 ModStamp:4498ec9e fDebug: 0 Offset: 0002bc81. Does this offer any clue as to where my problem is, or should I look somewhere else? Hoverfish Talk 21:54, 25 December 2006 (UTC)

hrm.... I'm 100% sure you're approved (we already had this discussion on my talk page). Have you tried on any other machines? I realize that's a bit of a cop out, but I figured I'd reply if nobody else is. alphachimp. 04:38, 26 December 2006 (UTC)

I'll try another machine when possible. Thanks for the good idea. Hoverfish Talk 21:25, 26 December 2006 (UTC)

https: access?

Hi! Due to my internet connection, I have to connect to Wikipedia using the https: route. Can someone who knows hwo to write C# submit a fix to AWB to allow users to connect via https? --

is me

03:26, 26 December 2006 (UTC)

You might not have noticed but I asked for a fix for the same problem two subjects up this talk page. Except I didn't know enough to specifically ask about a C# solution. You might want to watch my question as well for any input on this matter. --

talk • contribs

04:30, 26 December 2006 (UTC)

I'm going to sleep now, but I'll look into this tomorrow. —Mets501 (talk) 04:32, 26 December 2006 (UTC)

Hmm...the problem is that the base site, https://secure.wikimedia.org/wikipedia/en/, is forbidden, confusing AWB's checks. MaxSem can probably fix that; he's a much better programmer than I am :-). —Mets501 (talk) 15:21, 26 December 2006 (UTC)

Gee, this wikidumb MaxSem can't figure out how to work with secure.wikimedia.org, he gets either Wiki does not exist or You don't have permission to access / on this server. errors. MaxSem 18:09, 27 December 2006 (UTC)

A wikidumb steward! The world is coming to an end! You need to go to https://secure.wikimedia.org/wikipedia/en/wiki/ to access the wiki, and index.php is at https://secure.wikimedia.org/wikipedia/en/w/index.php. —Mets501 (talk) 22:35, 27 December 2006 (UTC)

Possible flaw?

I just finished a lengthy CFD run using AWB (deleting a category from more than 200 articles, thanks for AWB!!!!). It did great, but when I checked to see if the category was empty, there was still a page listed. It turns out that the article had the category added as "CAtegory:..." instead of "Category:...". Apparently Wikipedia doesn't seem to care much about case, but it looks like AWB does. Any way to fix this? Or this is not what was wrong?—DomBot ^talk ; Chidom ^talk, owner/operator. 11:06, 27 December 2006 (UTC)

Fixed in the next release. —Mets501 (talk) 14:46, 27 December 2006 (UTC)

Eh... This function is used in a numerous pieces of code, and some testing is required to make sure that this change won't break anything. For example, while namespaces are case-insensitive, page names are. And CaseInsensitive is used for matching image names, for example. MaxSem 17:45, 27 December 2006 (UTC)

I see, I thought it was only used for namespaces. I'll fix that now. —Mets501 (talk) 18:07, 27 December 2006 (UTC)

OK, all fixed with a new AllCaseInsensitive Function. —Mets501 (talk) 18:35, 27 December 2006 (UTC)

"Do you really want to do that?"

I have not used the delete or move buttons on the AWB GUI, but it is near to the ignore button and I guess that it could be pressed accidentally. Does AWB display a notice and is a second click needed to actually action larger changes? Snowman 21:19, 27 December 2006 (UTC)

I didnt know myself, so i just checked. When you press Delete, you get asked for a summary, and then you have to press delete again. For Move, you are asked for a summary and a new title, and then you have to press move again

Reedy Boy

21:52, 27 December 2006 (UTC)

Download?

I have requested approval, and since I'm in good standing, I expect to be approved. So... I'm just wondering: Do you download the AWB, because I don't see a download link anywhere? TeckWiz^TalkContribs_@ 21:22, 27 December 2006 (UTC)

Reedy Boy

21:49, 27 December 2006 (UTC)

Help Menu Item in AWB

If you go on Help --> Help, it takes you too [6]

Can you change it to [7]

Saves users having to click on a link!!

Reedy Boy

22:52, 27 December 2006 (UTC)

Done —Mets501 (talk) 23:44, 27 December 2006 (UTC)

Cheeers!

Reedy Boy

13:55, 28 December 2006 (UTC)

Remove double links

Is there a way to automatically take out double links. All I see is a list of them, but no way to take it out automatically, --(trying to get 1,000 edits in 1 month!) TeckWiz^TalkContribs_@ 03:38, 28 December 2006 (UTC)

I don't think things like that could be automated very well, because it is hard to "guess" which link should remain; if you want to remove double links, you can click on an item in the list to see where it is used in the article. —Mets501 (talk) 04:10, 28 December 2006 (UTC)

See the suggested rules in item 1 in my wishlist above. I think it'd be pretty straightforward to automate removal of most duplicate links. Colonies Chris 09:51, 28 December 2006 (UTC)

I appear on the check page, but get an error message saying I'm not enabled

I appear on the check page, but when I try to use AWB, I get an error message saying "Crocodile Punter is not enabled to use this". How can I resolve this? Thank you! --Crocodile Punter 14:15, 28 December 2006 (UTC)

Should be fixed now. —Mets501 (talk) 14:17, 28 December 2006 (UTC)

Thanks! --Crocodile Punter 07:41, 29 December 2006 (UTC)

Select all

Is their a way to select all the contents of a list quickly ie a select all button if not can it be added (Gnevin 00:07, 29 December 2006 (UTC))

Not so quickly, but if you use the "Save" button it prompts you for a name for a text file and saves the contents of the list there. It is formatted with wikicode for a numbered and wikilinked list:

# [[Creuse]]

# [[Gard]]

# [[Gironde]]

Pasted into a sandbox, that would result in:

1. Creuse
2. Gard
3. Gironde

This could be a handy thing or a nuisance, depending upon your intended use. It's easy enough to do a global search to remove the wikilink characters, however.

Alternatives for selecting the list:

Click the first item at the top of the list, scroll to the bottom of the list, hold down the shift key and click the last item in the list.

Or click the first item at the top of the list, hold down the Shift and Ctrl keys and touch the End key (commonly, Ctrl-Shift-End)

Both of these methods highlight the entire list; then right-click in the list and you'll get a list of options, one of which is to save the list as a file.

In other words, use the "Save" button and then employ a text editor if you don't want the numbered list of links.—DomBot ^talk ; Chidom ^talk, owner/operator. 04:44, 29 December 2006 (UTC)

False Positive

What's a false positive? TeckWiz^TalkContribs_@ 14:07, 29 December 2006 (UTC)

In general, a false positive is something that is not supposed to match a criteria, but does by accident. For examples, if you wanted to replace "hell" with "heaven", and "hello" was replaced with "heaveno", that would be a false positive (sorry, that's the best example I can come up with :-)). I still can't figure out how it applies to AWB, though. —Mets501 (talk) 15:28, 29 December 2006 (UTC)

WP:CFD/W

and Bot Friendly Lists

I hope you're all having a really great and pleasant holiday season.

I'm looking for a new bot-oriented feature in AWB. Ideally, it could completely eliminate the backlog in

WP:CFD/W

. Once the decision is made to delete or move a category, it is listed on this page in a bot-readable format. Bots run through and depopulate the categories one by one. Is there any interest in implementing the capability to read and act on the following type of list (this is an example from the current page):

• Wikipedia:Categories for deletion/Log/2006 December 14

  • Category:Acorn Archimedes software to Category:RISC OS software
  • Category:Musicals stars to Category:Musical theatre actors
  • Category:Recipients of Iron Cross to Category:Recipients of the Iron Cross
  • Category:People from Fukuoka to Category:People from Fukuoka, Fukuoka
  • Category:People from Konan to Category:People from Konan, Aichi

Thanks! alphachimp. 15:40, 26 December 2006 (UTC)

I'll try to implement this in a separate plugin. Hope, nobody minds if I'll work on backlog here while debuggin' it? (not in bot mode, all edits will be manually confirmed) MaxSem 18:06, 27 December 2006 (UTC)

As

talk • contribs • Bot

) 18:24, 27 December 2006 (UTC)

First results seem to be OK (except for incorrect link to discussion, but that's my fault, not plugin's). I will release it a bit later, after some improvements. The procedure is the following: you copypaste contents of section to the plugin window (available at Plugins → Recategorise per CFD), it extracts categories to be removed/replaced and loads their contents. Then you only need to set a proper edit summary and press "Start". MaxSem 13:40, 29 December 2006 (UTC)

Brilliant :-) —Mets501 (talk) 13:48, 29 December 2006 (UTC)

I love it :). This should really help. alphachimp. 20:59, 29 December 2006 (UTC)

I've committed this plugin to SVN, need some testing before release. MaxSem 10:13, 30 December 2006 (UTC)

Looks great! I don't see any problems. —Mets501 (talk) 15:24, 30 December 2006 (UTC)

WP:AWB/CFD. Shall we ship it in the main distro, and if we shall, shall it be enabled by default? MaxSem

16:09, 30 December 2006 (UTC)

I would. It's not too large, so it wouldn't affect download times or whatever for people who don't use it, and I think a lot of people will want to use it. It doesn't make a difference if it's enabled by default and you just don't enter anything into the backlog box, right? I would enable it, because otherwise people might accidentally forget to check the box to enable it. —Mets501 (talk) 16:23, 30 December 2006 (UTC)

No, I meant that it could be shipped in a subdirectory, and therefore inactive, because most users don't need it anyway. Those who need it can copy it to the main AWB dir, thus enabling it. MaxSem 17:49, 30 December 2006 (UTC)

OK, that sounds good. —Mets501 (talk) 18:17, 30 December 2006 (UTC)

Suggestion: Skip certain pages and sections

Could AWB be made to skip certain pages that contain intentional spelling errors or unusual words that might be mistaken for spelling errors. This feature could be turned on or off by a user option.

The need is particularly urgent should an AWB typo hunt accidentally hit a page such as:

• Wikipedia:Lists of common misspellings
• List of common misspellings in English
• Wikipedia:AutoWikiBrowser/Typos
• Wikipedia:WikiProject Fix common mistakes

And probably a good many more...

The problem with such pages is that if AWB gets to work on them with a long list of regexes the whole system gets locked up as AWB tries and tries to make hundreds of pointless changes.

More obvious applications include protecting vebatim quotations etc.

Perhaps pages or sections could be marked in the original text by a standardised flag comment such as: :<!--Automatic spellcheck off-->''Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.''<!--Automatic spellcheck on--> Gaius Cornelius 13:27, 29 December 2006 (UTC)

There is such option, at the end of

Mathematica. Admins, please update! MaxSem

14:14, 29 December 2006 (UTC)

Done. But that says it protects it against general fixes; will that also protect it against RegexTypoFix? —Mets501 (talk) 15:26, 29 December 2006 (UTC)

Ow, no, it doesn't. It prevents following things: unicodification, auto tagger and general fixes. Shell I disable RETF too? MaxSem 18:16, 29 December 2006 (UTC)

Probably. Either that or we should create a new section for ignoring RETF. —Mets501 (talk) 20:11, 29 December 2006 (UTC)

I made the new section and coded it to ignore pages on that list, but it's not working, so I don't want to commit it to SVN :-( I don't see what the problem is. —Mets501 (talk) 17:42, 30 December 2006 (UTC)

I'm missing the section of our Checkpage for adding that type of pages. Can somebody point me to it? Thanks, alphachimp. 21:00, 29 December 2006 (UTC)

You have to view it via the edit page (it's commented out at the bottom) —Mets501 (talk) 21:46, 29 December 2006 (UTC)

A more complete list - can somebody add it?

• Wikipedia:Lists of common misspellings
• Wikipedia:Lists of common misspellings/A
• ...
• Wikipedia:Lists of common misspellings/Z
• Wikipedia:Lists of common misspellings/Homophones
• Wikipedia:Lists of common misspellings/Repetitions
• Wikipedia:Lists of common misspellings/For machines
• Wikipedia:AutoWikiBrowser/Typos
• Wikipedia:WikiProject Fix common mistakes
• Wikipedia:Common mistakes
• List of common misspellings in English
• Spelling reform
• Misspelling

3.0.6.2

Can someone enable it for me?

Cheers

Reedy Boy

22:47, 30 December 2006 (UTC)

Done —Mets501 (talk) 22:56, 30 December 2006 (UTC)

Thanks =)

Reedy Boy

23:07, 30 December 2006 (UTC)

More list features

I would like to see the following features:

1. Deep category content
2. Intersection of lists
3. relative complement

Eli Falk

08:13, 31 December 2006 (UTC)

Particularly lists intersection would be a useful feature for me. Right now this has to be done externally to AWB. --Lysy^talk 08:24, 31 December 2006 (UTC)

1) "deep category content"? You mean subcategories? 2,3) Set operations are supported, just press Filter. MaxSem 10:43, 31 December 2006 (UTC)

Set operations are supported through external (saved) file, which is fine of course as it adds more flexibility (and requires only few more clicks). --Lysy^talk 13:27, 31 December 2006 (UTC)

I can make it work with copy/paste... MaxSem 13:30, 31 December 2006 (UTC)

About deep category content, I mean that if I want to see all stubs in

Eli Falk

16:18, 31 December 2006 (UTC)