Jump to content

Wikipedia talk:AutoWikiBrowser

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

This is an old revision of this page, as edited by Bluemoose (talk | contribs) at 13:48, 30 August 2006 (→Buglet). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

Archive
Archives

Frequently asked questions

  • When I start it up I get error "The application failed to initialize properly (0xc0000135). Click on OK to terminate the application."
    This error means your computer does not have the .NET framework version 2 installed properly. Download
  • Will it ever work on linux?
    Probably not.
  • Does AWB work on other projects/languages?
    Many WikiMedia projects and languages are supported, see the "Select language and project" option in the file menu. Other languages will be added on request, though at the moment the interface is always in English.
  • I don't like or use Internet Explorer, please use FireFox instead.
    AWB does not use Internet Explorer, it does however happen to use the same web browser control that Internet Explorer does, the equivalent FireFox component does not provide the needed functionality.
  • How do I open the page in another browser if I can't use the one in AWB?
    Right click on the edit box in the bottom right side of your screen. Select "Open page in browser"
  • How do I edit a page that doesn't exist?
    Uncheck "Ignore non existing pages" in the "Skip articles" box.
  • How do I skip certain articles?
    Use the "Skip if contains" and "Skip if doesn't contain" in the "(2) Set options" tab
  • Can't you leave up a "stable" version, so I don't have to download new versions?
    It is important to keep people up to date with the latest versions, because their use of the software doesnt just affect them, but the whole of wikipedia. As any bugs that remain will be trivial, hopefully releases won't be so frequent anyway.

Discussion

Formatting templates

While I'm waiting to see what I can do on Wikinews with AWB, I'm trying out the MWiki-Browser. How would I go about formatting all occurrences of a template with it?

From

*{{source|url=http://somewhere.at.example.com|title=This is not the news you are looking for|author=|pub=Example.com|date=August 25, 2006}}

To

*{{source|url=http://somewhere.at.example.com

|title=This is not the news you are looking for

|author=

|pub=Example.com

|date=August 25, 2006}}


And, as a (hopefully) minor feature request, can the protection of articles be performed from AWB. This is probably only of use to Wikinews where all articles 10 days old are protected. --Brianmc 08:19, 26 August 2006 (UTC)[reply]

AWB now works with wikinews. I think the best way to do that would be using the advanced find and replace to only do replacments inside templates, though i haven't really done anything like it before so i'm not sure what would be the best way. Protection of articles would currently be quite technically difficult. Martin 08:37, 26 August 2006 (UTC)[reply]

Kingbotk has had a few false positives, where I've tagged talk pages for articles deleted since I built my list. Looking at the message displayed above the edit box for new talk pages, I think AWB could very easily trap and avoid this.

This is a currently redlink talk page and redlink article HTML: 

please verify that a page called <i><a href="/w/index.php?title=There%27s_nothing_to_see_here%2C_move_along&action=edit" class="new" title="There's nothing to see here, move along">There's nothing to see here, move along</a></i> exists.

This at the time of writing is a blue link article but red link talk page HTML: 

"please verify that a page called <i><a href="/wiki/About_a_Book_Club_%28Hope_%26_Faith_episode%29" title="About a Book Club (Hope & Faith episode)">About a Book Club (Hope & Faith episode)</a></i> exists"

It's a silly message really, because Mediawiki has had to look up whether the article exists or not!!! Anyway, it seems that class="new" is responsible for making the link red.

Presuming that the message comes from the Mediawiki namespace somewhere and not from the PHP code directly, we can also leave a message on the Mediawiki talk page asking to be alerted of changes or of course just watchlist it :) --kingboyk 12:46, 26 August 2006 (UTC)[reply]

Stupid me... much easier way is to just check if the "article" tab is red or not. --kingboyk 10:11, 28 August 2006 (UTC)[reply]

OS X?

I'm not sure if this has been asked before, but is there a version of Mac OS X? Geoking66 02:40, 27 August 2006 (UTC)[reply]

For there to be such a version the OS would need to support the .NET framework version 2 and have Internet Explorer. In other words, I doubt it. --kingboyk 09:59, 27 August 2006 (UTC)[reply]

Most (Mboverload) typos in one article?

Possibly Japanese settlers in New Guinea during 1919-1940 times "comunity → community (12), regulary → regularly, autorities → authorities (2), autority → authority, belived → believed (2), colaborators → collaborators (3), condemmed → cond" then the edit summary fins out of space. Approx 32 replacements. Rich Farmbrough 08:51 27 August 2006 (GMT).

Wow, that's mighty impressive, shame the article is probably going to get deleted! Martin 10:15, 27 August 2006 (UTC)[reply]

New version ignoring option?

The new version seems to be enabling "add replacements to edit summary" even though I have this option disabled in the settings. It's just doing it anyway. Am I missing something or is this a bug? --W.marsh 13:19, 27 August 2006 (UTC)[reply]

It is a new bug, I improved the edit summary system and missed something, the older version is ok. thanks Martin 13:25, 27 August 2006 (UTC)[reply]

here [1] it only added a summary with no change, link [USA] should be changed like in here [2] pl:gregul 06:30, 28 August 2006 (UTC)

It is most likely a bug in what ever find and replace strings were used, AWB has an option to ignore articles when no replacement was made anyway. Martin 10:42, 28 August 2006 (UTC)[reply]

Nested square brackets bug

I quote Timeshifter who I think explains the symptom well

The bot is causing some images to not show up. The image is Table 1 in the section called "Frequency of Incarceration." SmackBot deletes one of the brackets at the beginning of the external link in the caption. It also deletes the 3rd bracket at the end of the caption and link. It obviously is not recognizing the stacked brackets due to the combination of link and image coding ending at the same place. This is a serious problem because it is common to put a source link at the end of a sentence. Here is the correct image coding:

[[Image:USA._Prisoners_1995_to_2005.gif|right|frame| Table 1 from [http://www.ojp.usdoj.gov/bjs/abstract/pjim05.htm "Prison and Jail Inmates at Midyear 2005".] A U.S. Bureau of Justice Statistics report. The totals do not include people held in juvenile facilities. According to a 2006 OJJDP (Office of Juvenile Justice and Delinquency Prevention) report there were 97,000 held in juvenile facilities as of October 22, 2003. [http://ojjdp.ncjrs.org/ojstatbb/nr2006/html/chp7hl.html]]]

SmackBot ends up with this coding below, and it causes the image to not show up:

[[Image:USA._Prisoners_1995_to_2005.gif|right|frame|Table 1 from [http://www.ojp.usdoj.gov/bjs/abstract/pjim05.htm "Prison and Jail Inmates at Midyear 2005".] A U.S. Bureau of Justice Statistics report. The totals do not include people held in juvenile facilities. According to a 2006 OJJDP (Office of Juvenile Justice and Delinquency Prevention) report there were 97,000 held in juvenile facilities as of October 22, 2003. [[http://ojjdp.ncjrs.org/ojstatbb/nr2006/html/chp7hl.html]]



Clearly a somewhat rare occurrence to have two external links in an image link, but there it is. Rgds, Rich Farmbrough 09:22 28 August 2006 (GMT).

Ok, thanks I'll look into it. Martin 09:38, 28 August 2006 (UTC)[reply]

Tip - Disappearing space on C:

This might be worthy of a mention on the project page.

Since I started using AWB in anger, free space on C: has become an issue. I cleared 2GB of space and soon that was gone too. I searched for large files, cleared my internet cache regularly, but still the drive would be full. Well, exploring in Cygwin it would seem that IE creates a sh*tload of files in the Temporary Internet Files folder which don't get cleared even if you tell IE to clear it's cache. What's more, they're hidden in Explorer even if you have it configured to show hidden files. Why they would do this I don't know - it's slightly sinister if you ask me - but, anyway, if you find that C: is gobbling up the gigs with no apparent cause this might be it.

It's covered in more detail here. The best tip seems to be:

  • Click Start, click Run, type the following command and click OK:
Shell:Cache\Content.IE5

--kingboyk 09:53, 28 August 2006 (UTC)[reply]

I haven't noticed a problem, but if it does does exist for some people it will only become noticed if they were doing 10s of 1000s of edits. The cache would probably clear itself after a period of time anyway, unless there is a massive flaw in how internet explorer works. Martin 10:49, 28 August 2006 (UTC)[reply]
I'd call 386,156 files in just one of the subfolders - having clicked "clear cache" multiple times - a massive flaw! I'm only part of the way through clearing this crud and I've clawed back gigabytes. This is insane! Anyrode, I hope the tip is useful to somebody, it's sure helped relieve my disk space issues. --kingboyk 10:55, 28 August 2006 (UTC)[reply]
But it should be noted that your bot has made enourmous number of edits in a relatively short period of time, for the average user doing a few 1000 edits this will never be an issue. Martin 11:09, 28 August 2006 (UTC)[reply]
It's not that I'm blaming you Martin (unless you work on the MSIE team :)) but I think that a "clear cache" button which leaves several gigs of cached files in place, to the extent that a user's C drive is full and programs start crashing, is quite manifestly broken. Perhaps we'll have to agree to disagree on this point :) --kingboyk 19:38, 28 August 2006 (UTC)[reply]
For me clearing the cache of IE (in IE) clears "Shell:Cache\Content.IE5" (accessed by the procedure you've shown above). On a second note, AWB just uses the BrowserControl which is shared with IE. So we are rather bound to that with all the drawbacks/bugs. If you can provide a better control for browsing (or refer to one) or a tweak to the AWB code, you are of course very welcome :-). --Ligulem 08:48, 29 August 2006 (UTC)[reply]
I know that - why does everyone seem to think I'm complaining?! Can't a guy indulge in a bit of gentle Microsoft bashing? :) --kingboyk 09:11, 29 August 2006 (UTC)[reply]
For your curiousity: The mozilla control which iirc can be downloaded here, while nice, does not provide the critical functionality that AWB needs. Martin 08:54, 29 August 2006 (UTC)[reply]
Ooh. Interesting, thanks. --kingboyk 09:11, 29 August 2006 (UTC)[reply]

small bug

There's a small bug if a URL is embedded in a image description link where AWB messes up the brackets: for example [3], you have [[image:x| ... [url...]]], and awb removes one of the right square brackets and adds a left square bracket before the url. Tim! 10:45, 28 August 2006 (UTC) This is the same as two sections up, sorry :) Tim! 10:47, 28 August 2006 (UTC)[reply]

Fixed in newest release. Martin 08:43, 29 August 2006 (UTC)[reply]

Weird bug with talk pages

I have a weird bug with AWB when prepending information to talk pages. When I want to prepend a msg to a list of talk pages, the diff blanks the whole page with just a "Modified" word. Here is the screenshot: (clicky).

Is it me or is there something weird?

Thanks, Grafikm (AutoGRAF) 11:34, 28 August 2006 (UTC)[reply]

I'm trying to release a new version at the moment that cleans up a few issues, but sourceforge is giving me an "internal server error", I'm sure it will work soon. Martin 11:45, 28 August 2006 (UTC)[reply]
Done it now, hopefully that will be the last release for a while. Martin 11:48, 28 August 2006 (UTC)[reply]
Woot, thanks!!! :) -- Grafikm (AutoGRAF) 12:06, 28 August 2006 (UTC)[reply]

Auto-update?

How about a one button "upload new version" to make re-installing a snap when an existing version is superseded? bd2412 T 15:26, 28 August 2006 (UTC)[reply]

Not really possible while it is hosted at sourceforge, hopefully the frequency of releases will slow down now anyway. Martin 08:43, 29 August 2006 (UTC)[reply]

Prepending to talk pages reloaded

While testing my bot that currently delivers newsletters, I encountered a weird thing. While prepending to an un-existing page, well, nothing gets prepended. The diff is loaded but is not saved. And of course, the option "Ignore non-existing pages" is unchecked, as can be seen. And it works for a blank article talk page too. Obviously, it is only true for auto-mode.

Is it a bug or a feature?

Thanks, Grafikm (AutoGRAF) 16:57, 28 August 2006 (UTC)[reply]

Hhhhmm, I tihnk it's fixed now, hopefully this really will be the last release for a while. Martin 08:43, 29 August 2006 (UTC)[reply]

(Hopefully very easy) feature request

Hi Martin: can you create a way to add the contents of first-level subcategories to the article list? For example: when you make a list from a category, you get all the pages in the category plus the subcategories as part of the list. Do you think it would be possible to be able to double-click on the category in the list of articles to add the contents of that category to the list? Or something of the like...? It would be much easier than copy/paste, especially if you wanted the pages in all the subcategories of something with a huge amount of subcategories, such as Category:Saints by country. —Mets501 (talk) 21:24, 28 August 2006 (UTC)[reply]

If you highlight the categories in the list and then open the context menu "Add selected to list..." and then "from category", this will get all the articles from those categories. Martin 21:34, 28 August 2006 (UTC)[reply]
/me slaps himself in the head. Thanks :-) —Mets501 (talk) 21:49, 28 August 2006 (UTC)[reply]

Unicode bug

When I was operating WinBot in 3.0.2.3 I was told that the unicodify in this edit was a bad one. So I am wondering if there is a bug in here? Many thanks. --WinHunter (talk) 03:44, 29 August 2006 (UTC)[reply]

IOW, the bug discussed at Wikipedia_talk:AutoWikiBrowser/Archive_8#AWB_bots_are_breaking_non-BMP_articles appears to be back. EdC 07:50, 29 August 2006 (UTC)[reply]
Ok, fixed in the newest verison. Martin 08:43, 29 August 2006 (UTC)[reply]

Linux?

Can I run it on Linux, using Mono and Wine. --Mihael Simonič 07:13, 29 August 2006 (UTC)[reply]

Afraid not. Martin 08:43, 29 August 2006 (UTC)[reply]

What about this: I have .NET farmework and IE6 instalated with Wine, mscoree.dll is also included in Mono. --Mihael Simonič 13:50, 29 August 2006 (UTC)[reply]

You like to make things complicated, I see :)) -- Grafikm (AutoGRAF) 13:59, 29 August 2006 (UTC)[reply]
Lol. It's not reasonable to expect Martin to support, erm, "esoteric" configurations like that. If you can get it to work, great - please report back - but the official line I suspect will remain the same :) --kingboyk 15:41, 29 August 2006 (UTC)[reply]

Plugins

Feature requests

  • (Being looked into) Access to XML settings. (If this isn't available I might try to go in the opposite direction, having the plugin control AWB's settings)
  • HasChanged boolean byval argument to ProcessArticle(), so that the plugin doesn't tell AWB to skip when AWB has made a change; or (easier) ignore the plugin's Skip value if AWB made a change
  • If the plugin has set a valid edit summary, have AWB not complain about empty edit summary box (but if in point above the Skip value is ignored and plugin returns an empty summary, AWB must use it's own summary)

--kingboyk 15:49, 29 August 2006 (UTC)[reply]

Reading and writing AML settings is a possibility, actually changing the AWB settings in a definite no. Having the "HasChanged" variable would be tricky, I know things like that look easy but they are not, largely because it is fundamentally against how AWB works, i.e. if you are doing job x, then the article either needs job x doing (so save it) or it doesn't (so ignore it), also, anything that is done with find+replace wuithin AWB can be done easily in the plugin anyway. Martin 16:50, 29 August 2006 (UTC)[reply]
True, true. Thanks. My plugin is working now anyway and hopefully will go into service later today - I have a backlog to catch up on! :) --kingboyk 17:15, 29 August 2006 (UTC)[reply]

I'm going to be a pain in the arse now and suggest that - given your (well argued) line about moving all work to the plugin and not using AWB's skip/find/replace features at all, the code which calls the plugin ought to be moved back to where it was :) i.e. AWB gets article text, sends it to plugin for processing, and if plugin says skip that's the job done. What do you think?

I took your advice by the way and moved my "skip this article" regex into the plugin. I hardly need worry about AWB settings now, it's all blank settings except for "make from file" and an edit summary of "Bot". --kingboyk 12:51, 30 August 2006 (UTC)[reply]

Well, you're the only person making a plugin, so i don't see the harm in moving it back. Also, I have added 3 methods to the interface for reading/writing XML and reseting the settings. It is a bit limited because of the complexity in dealing with plugins, I'll update my example when I have released the newest version. Martin 12:57, 30 August 2006 (UTC)[reply]

Plugins in testing

I have a plugin ready (User:Kingbotk/Living persons biography plugin) and am testing it, should anybody be interested :) --kingboyk 15:49, 29 August 2006 (UTC)[reply]

Small categorisation bug

When recategorising articles, could you please fix it so that links that start [[:Category are ignored, or at least create an option for this in settings? I'm not sure this is a very good explanation of what I mean, but hopefully it's understandable. Thanks, — FireFox (talk) 16:01, 29 August 2006

Why would you not want to change links like that? they will also need to be changed surely? Martin 16:42, 29 August 2006 (UTC)[reply]
Well not if they're in discussion, changing someone else's comments. — FireFox (talk) 16:44, 29 August 2006
I don't see a problem with changing a link in someone's comments if leaving it would result in the link being incorrect. This doesn't apply to automated bots of course, as a human needs to check the context. --kingboyk 10:19, 30 August 2006 (UTC)[reply]
Well I am mainly talking about an automated bot anyway, FFBot, so an option here would be helpful. — FireFox (talk) 13:00, 30 August 2006

Another small categorisation issue

When categories have been foolishly placed in the middle of text it seems that their removal leads to extra carriage returns being inserted. If followed by spaces this can change formatting, e.g. this edit. Regards, Rich Farmbrough 10:01 30 August 2006 (GMT).

new feature

1) can this be implemented: change from i.e. [laser|lasery] to [laser]y ? to simplify the links

2) why can't general fixes/unicodify (like removing underscores) be done (by AWB) before "find and replace" ? this causes a need of consideration all unicode/special characters into the regexes used into "find and replace" (special wikipedia's characters can't be matched by "find" until you know them exactly, but it's hard to cosider them everywhere!) gregul

1) Already a feature under "Apply general fixes" Harryboyles 12:53, 30 August 2006 (UTC)[reply]
read again, it's done after "find and replace" so in fact it wont work when matched string contains those characeters, general fixes will change [laser|laser] to [laser] and not moregregul
It has to be done after the find+replace or it can cause some complications. Links aren't simplified any more than they currently are because this can often lead to some strange looking links. Martin 13:14, 30 August 2006 (UTC)[reply]
Actually, in this case it is fairly easy to allow an option to apply before or after the general fixes. Martin 13:27, 30 August 2006 (UTC)[reply]

Buglet

I thought my plugin had a bug, because it was skipping red-link talk pages. However, in debugging mode I found nothing wrong... then noticed that when I click "Auto save" AWB then automatically checks "Skip articles when no change made", which for some reason is causing those pages to skip.

I've no idea why AWB would auto-check "Skip articles when no change made" but if there's no vital reason for doing so could you turn that off pse Martin? :) --kingboyk 13:43, 30 August 2006 (UTC)[reply]

Unfortunately it's not a bug, it's a defensive feature, otherwise some people have a nasty habit of setting a bot loose without it on and making a series of trivial edits. Martin 13:48, 30 August 2006 (UTC)[reply]
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia_talk:AutoWikiBrowser&oldid=72818287"