Jump to content

Wikipedia talk:AutoWikiBrowser/Bugs

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
< Wikipedia talk:AutoWikiBrowser

This is an old revision of this page, as edited by 84.234.42.68 (talk) at 14:35, 18 April 2008 (→Will not find {{example}} In Template Rule / Sub rule). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

AutoWikiBrowser 6.3.1.1
Shortcuts
Archive
Archive 1
Archive 2
Archive 3 (Version 4.0)
Archive 4 (Version 4.1)
Archive 5 (Version 4.1)
Archive 6 (Version 4.2-4.3)

Please only use this page to report bugs in the current version of the software. Please update to the most recent version, check to make sure your bug has not been reported already on this page, and read the FAQ before posting a bug here. See How to Report Bugs Effectively for advice on how to write bug reports.

File a new bug report
But check for already existing or fixed bugs first!
This Is Mandatory!

Before posting anything related to non-Wikimedia wikis, please make sure that this site has a recent SVN version of MediaWiki with enabled Bot API. Older versions of MediaWiki, or wikis without the Bot API are not supported. Please be sure to mention the exact URL of your wiki.

Any discussion can follow as normal.

This format allows the developers to keep track of bugs and to make sure the bugs are fixed as fast as possible. Once the bugs have been fixed, they will be moved to the /Archive 6.

If you cannot find your Bug on this page, please check the /Archive 6, as it may have been fixed for the next release!

Undone typo still in edit summary

Status New
Description Typo undone by double clicking the diff line still appears in the edit summary, as in this edit. mattbr 15:48, 2 July 2007 (UTC)[reply]

I noticed than even if i undo the orphan tag and/or the stub tag, in the summary box is stated that AWB added these tages. Check here. -- Magioladitis (talk) 21:43, 5 January 2008 (UTC)[reply]

To duplicate: As described
Operating system Windows Vista
.NET FW Version Unknown
AWB version Alpha 3.9.3.2
Workaround Unknown
Fixed in version Unknown


I agree - I think it would be helpful to other users reviewing AWB edits if undone typos didn't appear in the edit summary. Rjwilmsi 17:50, 23 September 2007 (UTC)[reply]
I'm also supporting this, since I'm getting comments that the summary does not match my change.. - déRahier talk+contrib 17:10, 9 December 2007 (UTC)[reply]
I agree -- I wouldn't like for people to think from my edit summaries that I had changed things that were not actually typos! KathrynLybarger (talk) 04:11, 20 December 2007 (UTC)[reply]
SVN Rev 2028 is still having that issue. Is there going to be a fix for this eventually? Maybe generating the difference after the user hits Save instead of generating the list of changes before the user begins to review the changes? - Jameson L. Tai talk ♦ contribs 04:01, 22 January 2008 (UTC)[reply]
I agree, I ran into this too and would expect the summary to not list typos I did not apply. WilliamKF (talk) 03:29, 29 February 2008 (UTC)[reply]
Comment - This bug is still not fixed, as evidenced here. In this case, "Beng" is not a typo, so I removed the change, and yet it still appeared in the edit summary. —  Tivedshambo  (t/c) 18:24, 13 April 2008 (UTC)[reply]
As evident by it not being marked as fixed... —Reedy 18:35, 13 April 2008 (UTC)[reply]

Logged in user at bottom right still not updating

Status New
Description The bottom right corner with the user logged into AWB still doesn't always update when switching between profiles. This was said to have been fixed, but I'm running the version where that fix was supposedly fixed. --(Review Me) R ParlateContribs@ (Let's Go Yankees!) 02:21, 3 July 2007 (UTC)[reply]
To duplicate:
Operating system Windows Vista
.NET FW Version Unknown
AWB version 3.9.3.2
Workaround Exit and restart.
Fixed in version Unknown


I've just been testing it just now. Most of the time it does work, there just seems to be a few cases where it doesnt always update properly... Reedy Boy 14:47, 3 July 2007 (UTC)[reply]
Tried to improve this a bit... But still not much better... Reedy Boy 16:23, 5 July 2007 (UTC)[reply]
It seems, when its polling for the username, to check what it is, its still getting the old username... And therefore doesnt always update... Need to have a look at the way its doing it... And potentially move it around or something Reedy Boy 14:41, 1 October 2007 (UTC)[reply]

Its caching the username somewhere... UpdateWikiStatus/UserNameChanged events arent fired... So the form isnt getting updated.. MaxSem, any chance you could have a look? Presumably, its the webbrowser login... Its not changing pages.. or getting the updated session? —Reedy Boy 00:13, 28 December 2007 (UTC)[reply]

My preliminary analysis shows that WebControl.Wait() returns immediately in WebControl.Login() and thus we call UserProperties.UpdateWikiStatus() before we actually log in. The easiest way would be to add a static delay, but we should try find out the very root of this problem first. MaxSem(Han shot first!) 20:13, 28 December 2007 (UTC)[reply]
Would seem this is causing a problem relating to loading the checkpage. OsamaK added the Typo: bit to the ar.wp checkpage, but it wasnt loading it. When i breakpointed it, it only got as far in the code to load that on the 2nd attempt at logging in. The first seemed get kicked out in the if (!LoggedIn)... —Reedy Boy 18:23, 13 January 2008 (UTC)[reply]

Auto checkbox switches itself off

Status New
Description Editing in bot mode, sometimes I find AWB is sat there because the auto-save checkbox has mysteriously become unchecked. Possible clue: we have bad weather here, causing a bad connection. kingboyk 12:23, 20 July 2007 (UTC)[reply]
To duplicate:
Operating system Vista
.NET FW Version Unknown
AWB version svn
Workaround
Fixed in version Unknown


Maybe, i think it may be set to turn off/uncheck if there is problems... Reedy Boy 13:43, 20 July 2007 (UTC)[reply]
Actually... I dont think it is... Reedy Boy 22:54, 20 July 2007 (UTC)[reply]
It's always been like this. I agree it should stay on. You also can't save auto on in the settings file. This mean every time you get logged out, when you log back, you must re-check the box, and if you don't, the same thing that happens to Kingboyk happens to me. --(Review Me) R ParlateContribs@ (Let's Go Yankees!) 03:20, 21 July 2007 (UTC)[reply]
This is, nonetheless, a useful safeguard sometimes. Rich Farmbrough, 12:50 21 July 2007 (GMT).
Useful safeguard against what exactly? I agree that it's a setting which should be turned on by the operator and not saved in the xml,
That's what I thought you meant. Rich Farmbrough, 07:44 11 August 2007 (GMT).
but if it's turned off by AWB just because the connection was lost I'm not sure what's gained. Having a bad connection doesn't make my bot go wild, and even if in fact my bot does go wild it's easily shut down by an admin blocking it or somebody replacing my talk page redirect with a message. --kingboyk 14:25, 21 July 2007 (UTC)[reply]
Why shouldn't it be saved in the xml file? Now that we have profiles, you can associate one that always has "auto on" with your bot account only. --(Review Me) R ParlateContribs@ (Let's Go Yankees!) 16:47, 21 July 2007 (UTC)[reply]
I guess. The concept has always been this is a feature you must turn on, once you're happy with your settings. I wouldn't argue against becoming a little more liberal in that regard (current implementation can be a bit of a pain) but others might. --kingboyk 21:45, 24 July 2007 (UTC)[reply]
Simplish to fix.. If we decide which way it wants to be... Stay on? —Reedy Boy 00:27, 18 November 2007 (UTC)[reply]

IRCM doesn't show any actions with ar.wp

Status New
Description When I use IRCM with Arabic Wikipedia, I can't see any actions (moving, deleting, un-deleting, etc..) OsamaK 16:18, 12 September 2007 (UTC)[reply]
To duplicate: Unknown
Operating system Windows XP
.NET FW Version Unknown
AWB version SVN
Workaround None
Fixed in version Unknown


Does it on any other wiki's? We know en.wp works.. Reedy Boy 20:20, 13 September 2007 (UTC)[reply]
Just was shown a restore on commons... Reedy Boy 20:40, 13 September 2007 (UTC)[reply]
So the problem wasn't in non-English language.--OsamaK 19:10, 15 September 2007 (UTC)[reply]

Bold letters

Status Bug currently being investigated/fixed
Description Check here. AWB was supposed to convert the first two words in brackets into bold. For some reason, it didn't work. Magioladitis 03:31, 4 November 2007 (UTC)[reply]
To duplicate:
Operating system Windows XP Service pack 2
.NET FW Version Unknown
AWB version 4.1.4.0
Workaround None
Fixed in version Unknown


Presumably cause its linking back to itself.. Probably doesnt cater for these cases, and therefore didnt embolden. Reedy Boy 13:08, 4 November 2007 (UTC)[reply]
In other cases (here one but without the brackets and here one with the brackets!), I've seen AWB fixing it. It's a common mistake people to put [[]] instead of bold letters and AWB was working fine with that. -- Magioladitis 14:59, 4 November 2007 (UTC)[reply]
Hmm.. Whatever way, it needs sorting. Reedy Boy 15:03, 4 November 2007 (UTC)[reply]
It doesn't work when there is no other bolded text in the article because it's incorrectly marked as "NoChange" in BoldTitle(). I'm working on it. I wish Wikibluemoose made more comments in the source code :( Jogers (talk) 19:26, 8 November 2007 (UTC)[reply]
Just to report a normal case that didn't change to bold here. -- Magioladitis 02:34, 9 November 2007 (UTC)[reply]

PAGENAME

Status New
Description Inconsistent modification of {{PAGENAME}}. Sometimes the pipe is after the PAGENAME template in some infoboxes. Sometimes it changes it and sometimes it does not. The page "Venae cavae" is an example of one I had to change manually today. Snowman (talk) 19:24, 10 December 2007 (UTC)[reply]
To duplicate: Normal operation of AWB,
Operating system Wix XP pro Sp 2
.NET FW Version Unknown
AWB version 4142
Workaround write in the page name manually
Fixed in version Unknown


Bugs in general fixes

Status Bug currently being investigated/fixed
Description 1) changes [[1936 ]]roku to [[1936]]roku instead of [[1936]] roku

2) does this [1]

3) [2] - one hint for this, at the end of the article there was "[[Kategoria:Kościoły w diecezji tarnowskiej|Sz]]] "

--84.234.42.68 (talk) 00:35, 13 January 2008 (UTC)[reply]

To duplicate:
Operating system XP SP2
.NET FW Version Unknown
AWB version rev1974
Workaround
Fixed in version Unknown


I think 1st should be simple to fix, the 2nd is probably connected somehow to # in the link --84.234.42.68 (talk) 13:59, 23 January 2008 (UTC)[reply]
  1. Fixed in rev 2305.
  2. Disabled some time ago.
  3. Not sure what should we do about such misformatting. Anyway, in MW it looks no less ugly both before and after such change. MaxSem(Han shot first!) 18:36, 21 March 2008 (UTC)[reply]

Very large lists, can't pull the scrollbar down to the bottom

Status New
Description With a very large list (over 100,000 articles) it doesn't seem possible to hold down the listmaker scrollbar and scroll to the bottom. Every time it's moved to the bottom the list expands and instead of being at Z I jump back to say E. I can supply a zipped large list and more details if needed (and if I don't fix this myself in the meantime). kingboyk (talk) 20:53, 15 January 2008 (UTC)[reply]
To duplicate:
Operating system Vista
.NET FW Version Unknown
AWB version SVN
Workaround
Fixed in version Unknown


Just tried with 150,000 generated "articles" and it works fine... —Reedy Boy 20:05, 7 March 2008 (UTC)[reply]
Status New
Description AWB won't fetch from special pages. Reedy Boy also tested links and image pages, which also incorrectly return no results. – Mike.lifeguard | @en.wb 23:52, 18 January 2008 (UTC)[reply]
To duplicate: Attempt to create a list from any special page etc.
Operating system XP SP2
.NET FW Version Unknown
AWB version SVN 2011
Workaround manual list creation :(
Fixed in version Unknown


I think some work. Just not all/that image one... —Reedy Boy 00:07, 19 January 2008 (UTC)[reply]
4.2.0.2 still has this bug. You may well be correct; I've only tried to fetch a a list from Special:Unusedimages. – Mike.lifeguard | @en.wb 19:04, 28 January 2008 (UTC)[reply]
Same issue here. Tried Allpages and Prefixindex (both of which worked before); neither return any results. --TMF Let's Go Mets - Stats 02:42, 17 February 2008 (UTC)[reply]
4.3.0.0 still has this problem.
(Ru-Wikipedia)
Unusedimages still does not work. Gives 0.
But Contributions&namespace=6&limit=50&target=Alex+Spade has start to give out something - but its results are strange - it give out 102 pages, one of them is from Участник-space (User-space), more over I didn't edit the respective page. Alex Spade (talk) 21:28, 5 March 2008 (UTC)[reply]
Special:Newpages is consistently returning around 85 pages, only 20 of which are mainspace, no matter what parameters I use. A random sampling of those pages show that none of them are new. For example, Woodland, California, a page created in 2002, is one of the mainspace pages returned just now. --Fabrictramp (talk) 22:06, 5 March 2008 (UTC)[reply]

Make list confused by unpatrolled changes?

Status New
Description When I try to make a list of the new pages on the Finnish Wikipedia (http://fi.wikipedia.org/wiki/Toiminnot:Uudet_sivut), the resulting list is incorrect. The correct list currently begins with "Østlandet", "Gustav Peter Blom" and "Anatoli Fomenko", but the AWB-generated list looks different: "Rahva Hääl", "Viron sosialistinen neuvostotasavalta", "Brian Joubert" etc.

The AWB-generated list now seems to include only the patrolled changes and pages linked from their edit summaries (which is definitely a bug). This used to work correctly. Silvonen (talk) 05:50, 28 January 2008 (UTC)[reply]

To duplicate: Set your project to Finnish Wikipedia, select "Make from: Special page", type "Uudet sivut" in the Toiminnot box and press "Make list".
Operating system XP
.NET FW Version Unknown
AWB version 4.2.0.1
Workaround
Fixed in version Unknown


References-2column not replaced with 2 argument to reflist

Status New
Description The div class of "references-2column" is not handled. WilliamKF (talk) 00:13, 2 February 2008 (UTC)[reply]
To duplicate: Run AWB on a page with this in it:

==References== <div class="references-2column"> <div class="references-small"> <references/> </div> </div>
It should become:
==References== {{reflist|2}}

Operating system Windows XP
.NET FW Version Unknown
AWB version 4.2.0.1
Workaround Manual edit
Fixed in version Unknown


Probably... Im sure max put that in... But its not normal for it to have -small and the -2coloumn.. Probably the problem... —Reedy Boy 00:29, 2 February 2008 (UTC)[reply]

As per [3] —Reedy Boy 00:31, 2 February 2008 (UTC)[reply]

"using AWB" broken on custom wikis

Status New
Description "Using" is missing from "Using AWB" on custom wikis. Example.kingboyk (talk) 16:05, 4 March 2008 (UTC)[reply]
To duplicate:
Operating system Vista
.NET FW Version Unknown
AWB version SVN
Workaround None
Fixed in version Unknown


This is intentional. We localize edit summaries if possible on preset wikis, but can't do that on custom wikis, so we attemt to use as less unknown language as possible. MaxSem(Han shot first!) 16:09, 4 March 2008 (UTC)[reply]

Thanks Max. It looks a bit crappy though so let's leave the bug report open for now please - I'll have a think about how we might improve it. We could, for example, allow language selection of custom wikis, or allow customisation of the AWB string. I know folks could just alter their edit summary, but my observations (especially now we have some log data allowing me to look at custom wikis where AWB is being used) say that they don't. --kingboyk (talk) 18:48, 4 March 2008 (UTC)[reply]

Possible issue sending user name to stats server

Status New
Description Not sure if this is a bug (e.g. with the sending of Cyrillic characters, which I thought I'd fixed) or whether somebody has hacked their AWB (as is their right) but 10 sessions (and rising) don't have a username stored (the username is blank in the POST field, at least according to MySQL). Last time I checked these were all from RU wiki and probably just the one user. Could I get a second opinion on whether there's a bug or not? (AWB devs who want the FTP and SQL passwords and who don't have them yet should ask me or Sam btw). kingboyk (talk) 19:54, 8 March 2008 (UTC)[reply]
To duplicate:
Operating system
.NET FW Version Unknown
AWB version
Workaround
Fixed in version Unknown


AWB encountered XmlException

Status New
Description
Exception:XmlException
Message:'�', hexadecimal value 0x07, is an invalid character. Line 1, position 100.
Call stack:
   at System.Xml.XmlTextReaderImpl.Throw(Exception e)
   at System.Xml.XmlTextReaderImpl.Throw(String res, String[] args)
   at System.Xml.XmlTextReaderImpl.ThrowInvalidChar(Int32 pos, Char invChar)
   at System.Xml.XmlTextReaderImpl.ParseAttributeValueSlow(Int32 curPos, Char quoteChar, NodeData attr)
   at System.Xml.XmlTextReaderImpl.ParseAttributes()
   at System.Xml.XmlTextReaderImpl.ParseElement()
   at System.Xml.XmlTextReaderImpl.ParseElementContent()
   at System.Xml.XmlTextReaderImpl.Read()
   at System.Xml.XmlTextReader.Read()
   at WikiFunctions.Lists.GetLists.FromCategory(Boolean subCategories, String[] categories)
   at WikiFunctions.Controls.Lists.ListMaker.MakeList2()

bd2412 T 05:57, 13 March 2008 (UTC)[reply]

To duplicate: ListMaker Text: Noms communs en anglais
Operating system Microsoft Windows NT 5.1.2600 Service Pack 2
.NET FW Version Unknown
AWB version 4.3.0.0
Workaround None
Fixed in version Unknown


System.InvalidOperationException

Status New
Description System.InvalidOperationException: List that this enumerator is bound to has been modified. An enumerator can only be used if the list does not change. 
  at System.Windows.Forms.ListBox.ItemArray.EntryEnumerator.System.Collections.IEnumerator.MoveNext()
  at WikiFunctions.Controls.Lists.ListMaker.loadArticlesInBrowser() in C:\Documents and Settings\RJW\Desktop\AWB_SVN\AWB\WikiFunctions\ListMaker.cs:line 1257
  at WikiFunctions.Controls.Lists.ListMaker.openInBrowserToolStripMenuItem_Click(Object sender, EventArgs e) in C:\Documents and Settings\RJW\Desktop\AWB_SVN\AWB\WikiFunctions\ListMaker.cs:line 1252
  at System.Windows.Forms.ToolStripItem.RaiseEvent(Object key, EventArgs e)
  at System.Windows.Forms.ToolStripMenuItem.OnClick(EventArgs e)
  at System.Windows.Forms.ToolStripItem.HandleClick(EventArgs e)
  at System.Windows.Forms.ToolStripItem.HandleMouseUp(MouseEventArgs e)
  at System.Windows.Forms.ToolStripItem.FireEventInteractive(EventArgs e, ToolStripItemEventType met)
  at System.Windows.Forms.ToolStripItem.FireEvent(EventArgs e, ToolStripItemEventType met)
  at System.Windows.Forms.ToolStrip.OnMouseUp(MouseEventArgs mea)
  at System.Windows.Forms.ToolStripDropDown.OnMouseUp(MouseEventArgs mea)
  at System.Windows.Forms.Control.WmMouseUp(Message& m, MouseButtons button, Int32 clicks)
  at System.Windows.Forms.Control.WndProc(Message& m)
  at System.Windows.Forms.ScrollableControl.WndProc(Message& m)
  at System.Windows.Forms.ToolStrip.WndProc(Message& m)
  at System.Windows.Forms.ToolStripDropDown.WndProc(Message& m)
  at System.Windows.Forms.Control.ControlNativeWindow.OnMessage(Message& m)
  at System.Windows.Forms.Control.ControlNativeWindow.WndProc(Message& m)
  at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)

This occurred when I chose 'open article in browser' from article list (article was Gertrude Stein). Firefox is my default browser. Thanks Rjwilmsi (talk) 10:05, 15 March 2008 (UTC)[reply]

To duplicate:
Operating system XP
.NET FW Version Unknown
AWB version SVN 2270
Workaround
Fixed in version Unknown


Hmm, weird.. Its not modifying anything, so not sure why that happened.. Did you click on something else/remove something just after opening it..? As the code itself is not modifying the list.. —Reedy Boy 12:41, 15 March 2008 (UTC)[reply]
Yes, I clicked Ignore straight afterwards before Firefox had opened. Rjwilmsi (talk) 09:20, 16 March 2008 (UTC)[reply]

AWB encountered NullReferenceException

Status New
Description
Exception:NullReferenceException
Message:Object reference not set to an instance of an object.
Call stack:
   at WikiFunctions.Browser.WebControl.Login(String username, String password)
   at WikiFunctions.Profiles.AWBProfilesForm.browserLogin(String Username, String Password)
   at WikiFunctions.Profiles.AWBProfilesForm.browserLogin(String Password)
   at WikiFunctions.Profiles.AWBProfilesForm.login()

A4 (talk) 13:58, 20 March 2008 (UTC)[reply]

To duplicate: Unknown
Operating system Microsoft Windows NT 5.1.2600 Service Pack 2
.NET FW Version Unknown
AWB version 4.3.0.2
Workaround None
Fixed in version Unknown


Could you elaborate what exactly did you do to encounter this bug? Can you reproduce it? MaxSem(Han shot first!) 14:17, 20 March 2008 (UTC)[reply]

I'd love to know what causes this. It happens now and again, randomly.. And theres a null check on the web browser, but its gotta be something in the Login() method —Reedy Boy 14:21, 20 March 2008 (UTC)[reply]

DEFAULTSORT with spaces

Status New
Description This is more of a bug with DEFAULTSORT that needs to be recognized by AWB. {{DEFAULTSORT: foo}} does not include the space in the sort parameter. [4] is the diff where AWB did it. NE2 01:28, 23 March 2008 (UTC)[reply]
To duplicate: Run general fixes on 1964 state highway renumbering (Washington).
Operating system XP
.NET FW Version Unknown
AWB version 4.3.0.2
Workaround
Fixed in version Unknown


{{DEFAULTSORT: Renumbering}} is what its doing, but it needs to be {{DEFAULTSORT:Renumbering}}, trimming the leading space? —Reedy Boy 19:00, 23 March 2008 (UTC)[reply]

Sometimes people intentionally use space character instead of asterisk to sort a page before the others. MaxSem(Han shot first!) 19:05, 23 March 2008 (UTC)[reply]
It needs to not convert to DEFAULTSORT unless the MediaWiki bug is fixed. --NE2 20:53, 23 March 2008 (UTC)[reply]

focus lost when in 'hide bottom panel' mode

Status New
Description When using the mode 'hide bottom panel', the focus is lost from the preview window after saving each edit. This means that pressing Control+S to save when a new edit loads has no effect, focus has to be returned to the preview window by a mouse click. I think this occurs in normal (panel shown) mode too. Rjwilmsi (talk) 10:14, 24 March 2008 (UTC)[reply]
To duplicate:
Operating system XP
.NET FW Version Unknown
AWB version 4.3.0.2
Workaround
Fixed in version Unknown


I can't reproduce this. It works well for me. —METS501 (talk) 21:32, 17 April 2008 (UTC)[reply]

Null reference

Status New
Description
Exception:NullReferenceException
Message:Object reference not set to an instance of an object.
Call stack:
   at AutoWikiBrowser.MainForm.LoadSuccess() in C:\Documents and Settings\RJW\Desktop\AWB_SVN\AWB\AWB\Main.cs:line 719
Thread:Main thread

Rjwilmsi (talk) 17:51, 29 March 2008 (UTC)[reply]

David Gerard (talk) 14:58, 27 March 2008 (UTC)[reply]

To duplicate: Unknown
Operating system Microsoft Windows NT 5.1.2600 Service Pack 2
.NET FW Version Unknown
AWB version 4.3.0.3
Workaround None
Fixed in version Unknown


Which SVN revision? MaxSem(Han shot first!) 17:54, 29 March 2008 (UTC)[reply]

Not sure exactly what revision I last built but about rev 2310. Rjwilmsi (talk) 17:57, 29 March 2008 (UTC)[reply]

Ah, it's most likely string HTML = webBrowserEdit.Document.Body.InnerHtml; MaxSem(Han shot first!) 18:03, 29 March 2008 (UTC)[reply]

The error is now in line 730 of Main.cs (SVN 2338). Thanks Rjwilmsi (talk) 10:20, 6 April 2008 (UTC)[reply]

(1) unicode substitutions mangle Typographic articles
(2) workaround will mangle multi-paragraph Blockquotes in future

Status New
Description
  1. !NoWiki'd text– can mangle Typographic articles: inappropriate Unicode substitutions
  2. Temporary fix may prevent wiki-standard Blockquotes from displaying multi-paragraph quotes

Saltation (talk) 15:57, 31 March 2008 (UTC)[reply]

To duplicate: see below -- alternate-spacing unicode characters can be inappropriately replaced, and the workaround could cause future AWB edits to destroy the paragraph-separations in multiparagraph blockquotes
Operating system N/A (all)
.NET FW Version Unknown
AWB version (all)
Workaround workaround creates a further issue: removing paragraph markup in Blockquotes. See below.
Fixed in version Unknown


i've just been informed that a key problem a bot (User:SmackBot) has with typographically-significant unicode-representing html codes (specifically: replacing them with raw unicode chars and thereby, in some small but important areas of wikipedia, seriously and significantly reducing the ability of editors to use wikipedia usefully) is likely to recur due to AWB.

i hadn't heard of AWB before and frankly i'm not clear on its relationship with this bot (or vice versa) but the assertion by the bot's writer that "Be aware, however that the next AWB gen fixes to hit that page will probably remove those <p>s you have put in the blockquotes" suggests that this devel community should be aware of the problem sooner rather than later.

rather than risk clouding the issue by trying to paraphrase the problem, i will simply re-post below my initial flagging of the problem to the bot's writer and his response: (emphasis added)

SmackBot: conversion of HTML char-codes to raw Unicode: issue & consequent suggestion

greetings rich. first off, although you've probably heard it many times before, congrats and thanks for smackbot. it does good work, with few errors.

but (you knew there was going to be a but, didn't you), there's an exception:

it passed through French spacing and mungled the examples of different-width spacing in the unicode section. see the 2nd mod block HERE -- note the html charcodes &#8201; and &#8239; have been converted to their raw unicode equivalents, and in the latter example user-content becomes effectively invisible. even given my own knowledge of what should have been there (i created this section (by the miracle of copy-paste)), i thought the code had been completely deleted until i went to replace it and discovered by accident that there was an invisible essentially-zero-width character still between the last word and the exclamation mark.

now i'm in 2 minds as to doing this sort of thing anyway.

PRO: it's technically slightly purer for those on perfectly updated systems using standards-aware (typically paid-for) tools.
CON: it makes the article uneditable offline for anyone without access to unicode-compliant tools, which are much less common than most people think.
CON: some of the unicode codes' characters are not directly creatable on our current keyboards/OSs —that is, they CAN NOT be manually entered, and in some cases can not even be clearly or even visibly observed–can YOU determine at a glance in an edit-window that a gap in the text is not a normal space but actually a non-breaking space?— and in some key cases are not even visible: once they've been forced from being display-time characters to also being edit-time characters, most people lose the ability to directly enter them, and in some cases (eg varying-width spaces) may lose the ability to even see they are there.

but clearly i need to keep SmackBot away from the typographic examples.

i looked at {nobots}, and was about to exclude poor ole smackbot despite his sterling service and well-meaning edits. then i stopped and thought.

i do NOT want to forever exclude smackbot from hoving to with dustpan and brush, monkeywrench and oilcan, tidying up and improving common errors that may be inserted in future by later editors (none of us will be here forever {existential angst} )

but i DO need him not to munge my spacing examples.


then a penny dropped.

SUGGESTION:
modify SmackBot to have a context-sensitive ruleset.
CASE: where an article is tagged Category:Typography, he does NOT execute the html-code-->unicode-char conversion rules.

"simple as that", he says, secure in the knowledge that it's not him that would have to be making them changes...


LESS ARCHITECTURALLY-ARDUOUS SUGGESTION (which in many ways is superior to the above) :
you the developer do one single personal manual pass through the code's config data for the unicode chars smackbot will seek to convert, and remove at devel-time any character with typographic significance (perhaps best defined as: invisible behaviour different from a typewriter font; in particular: spaces).

for now, i'm going to {nobotno,badbot,down!} the article. but it'd be nice not to have to.

i remain,
yours in mutual futile pursuit of perfection,
Sal
Saltation (talk) 11:46, 31 March 2008 (UTC)
OK, I have a solution, but it's not perfect.
First realise that SB is almost exclusively run on WP:AWB, so excluding SB's "general fixes" will likely only buy you time (as another AWB bot will arrive).
There is however an option to ignore nowiki'd text, and a bunch of other stuff. SB is supposed to have this turned on, but sometimes I turn it off to get to picture captions - and forget to turn it back on.
So I've nowiki'd the unicode, which will probably keep most AWB bots away from them, including SB, provided I don't forget an leave the switch off.
Be aware, however that the next AWB gen fixes to hit that page will probably remove those <p>s you have put in the blockquotes.
Rich Farmbrough, 13:49 31 March 2008 (GMT). (tweaked 14:54)
It's not the workaround that "mangles" the paragraph tags, they should in theory go, but they are themselves a workaround to a browser problem, as I understand it. Rich Farmbrough, 19:21 31 March 2008 (GMT).
fair enough -- the need for raw html within blockquotes to mark paragraph breaks did always strike me as a workaround itself.
regardless: i have no problem with a bot or editor altering the markup syntax -- i DO care that multi-paragraph blockquotes continue to be presented to readers in their original (quoted) format: multi-paragraph
Saltation (talk) 23:15, 31 March 2008 (UTC)[reply]

See related and earlier bug &emsp; 

// Line 57, Parsers.cs
            //look bad if changed
            RegexUnicode.Add(new Regex("&(ndash|mdash|minus|times|lt|gt|nbsp|thinsp|shy|lrm|rlm|[Pp]rime|ensp|emsp);", RegexOptions.Compiled), "&amp;$1;");
            //IE6 does like these
            RegexUnicode.Add(new Regex("&#(705|803|596|620|699|700|8652|9408|9848|12288|160|61|x27|39);", RegexOptions.Compiled), "&amp;#$1;");

            //Decoder doesn't like these
            RegexUnicode.Add(new Regex("&#(x109[0-9A-Z]{2});", RegexOptions.Compiled), "&amp;#$1;");
            RegexUnicode.Add(new Regex("&#((?:277|119|84|x1D|x100)[A-Z0-9a-z]{2,3});", RegexOptions.Compiled), "&amp;#$1;");
            RegexUnicode.Add(new Regex("&#(x12[A-Za-z0-9]{3});", RegexOptions.Compiled), "&amp;#$1;");

            //interfere with wiki syntax
            RegexUnicode.Add(new Regex("&#(126|x5D|x5B|x7b|x7c|x7d|0?9[13]|0?12[345]|0?0?3[92]);", RegexOptions.Compiled | RegexOptions.IgnoreCase), "&amp;#$1;");

// line 316, Parser.cs
        [Test]
        public void DontChangeCertainEntities()
        {
            // http://en.wikipedia.org/wiki/Wikipedia_talk:AutoWikiBrowser/Bugs/Archive_3#.26emsp.3B
            Assert.AreEqual("&emsp;", parser.Unicodify("&emsp;"));
        }

From w3schools - HTML Symbol Entities

Result Description Entity Name Entity Number AWB Escapes
  en space &ensp; &#8194; Yes
  em space &emsp; &#8195; Yes
  thin space &thinsp; &#8201; Yes
‌ zero width non-joiner &zwnj; &#8204; No
‍ zero width joiner &zwj; &#8205; No
‎ left-to-right mark &lrm; &#8206; Yes
‏ right-to-left mark &rlm; &#8207; Yes

Preferable AWB convert &#8201; into &thinsp;. Table of spaces for the complete list of space marks. System should be able to handle &#x and &#X ways of writing HTML entities. — Dispenser 00:32, 9 April 2008 (UTC)[reply]

Suggestion on fixing:
  • Change //interfere with wiki syntax… to escape all character codes bellow 128 (plain ASCII) [&#32; wasn't on the list]
  • After Unicodify, change all non-print characters about 128 to their HTML equivalents (named preferred) [fix pages to be more readable in the text editor window, would have fixed Template:Italian Military eventually]
BTW, Saltation could you provide the diff of the other edit in question? — Dispenser 02:47, 9 April 2008 (UTC)[reply]
Status New
Description doesn not work well - [5] --84.234.42.68 (talk) 15:52, 2 April 2008 (UTC)[reply]
To duplicate: Unknown
Operating system XP SP2
.NET FW Version Unknown
AWB version rev. 2327
Workaround None
Fixed in version Unknown


BD-SORT

Status New
Description I 've been doing a replaced of BD-SORT (a redirect) with {{Lifetime}} when this occurred. It's funny because until this replacement AWB does is not happening all the time but sometimes, (for example check that in this case the replacement i requested occurred), moreover it doesn't create a DEFAULSORT as it should. If you really want to replace BD-SORT you have to create a birth category, a death category and defaultsort. Moreover, this edit shows that the replacement, sometimes, created THREE instead of two categories. -- Magioladitis (talk) 02:19, 6 April 2008 (UTC)[reply]
To duplicate: No idea
Operating system XP SP2
.NET FW Version Unknown
AWB version 4.3.0.2
Workaround
Fixed in version Unknown


Call me crazy but I suspect it has something to do with the header capitalization. It seems this bug is triggered only after a head capitalization. Check here as well. Moreover, I would like to inform you that BDSORT doesn't exist anymore, BD-SORT is now empty and BIRTH-DEATH-SORT will be soon empty as well. -- Magioladitis (talk) 10:57, 6 April 2008 (UTC)[reply]

The same bug occurs with BIRTH-DEATH-SORT as well. The only active templates in this category is Lifetime and BD (redirect of Lifetime). I replaced all the others with Lifetime. I think, it's better to completely remove the replacement completely.
Check this. It's very interesting! Some bullets appeared as well. -- Magioladitis (talk) 22:34, 6 April 2008 (UTC)[reply]
Btw, can an administrator unprotect and delete BD-SORT? It's empty and completely unnecessary redirect. Thanks -- Magioladitis (talk) 23:11, 6 April 2008 (UTC)[reply]
Done that. —Reedy 18:30, 7 April 2008 (UTC)[reply]
Cheers! I think we can just remove any reference to these templates from AWB. Lifetime works just fine. -- Magioladitis (talk) 22:43, 7 April 2008 (UTC)[reply]

In this example I make no replacement. I just apply general fixes, but the bug appears again. -- Magioladitis (talk) 09:31, 12 April 2008 (UTC)[reply]

Cannot connect to interwiki or load typos

Status New
Description When trying to load AWB for the last 2 days I have been getting several errors and I cannot use tha application. Here are the errors that I get in order of appearance.

1) The interwiki list didn't load correctly. Please check your internet connection and then restart AWB. 2) I get the REGEX attention screen 3) Load from English wikipedia?: No list of typos was found. Would you like to use the list of typos from the enlish wikipedia? Only choose yes if this is english wiki. I select Yes 4) There was a problem loading the list of typos 5) Cannot load version checkpage from wikipedia. Please verify that youve connected to the internet. Thats it and I can connect to the internet with everything else. Also, it seems to happen when its trying to load Main form:load prefs ().--Kumioko (talk) 12:57, 9 April 2008 (UTC)[reply]

To duplicate:
Operating system Windows XP
.NET FW Version Unknown
AWB version
Workaround
Fixed in version Unknown


Can you access Wikipedia from Internet Expolorer? Did something change in your network settings recently? MaxSem(Han shot first!) 09:00, 12 April 2008 (UTC)[reply]

Yes I can access IE, No nothing changed. I found a workaround and that is if I open IE and login to Wikipedia AWB will work. If I logout or close IE then it will not work. Its only with the newest version of AWB also.--Kumioko (talk) 13:55, 14 April 2008 (UTC)[reply]
Well, I spoke to soon. I was able to access it for 1 day and now I can't access it again. Is it possible someone suspended my access accidentally or something?--Kumioko (talk) 13:19, 17 April 2008 (UTC)[reply]

Will not find {{example}} In Template Rule / Sub rule

Status New
Description Will not find {{example}} In Template Rule / Sub rule. The regex tester finds and replaces it. It will find {example} (single curly brackets) but not double. It works fine as a regular rule, but not as a In Template Rule / Sub rule pete 17:29, 17 April 2008 (UTC)[reply]
To duplicate:
Operating system Microsoft Windows NT 5.1.2600 Service Pack 2
.NET FW Version Unknown
AWB version 4.3.1.0
Workaround manually remove outer most curly brackets and reparse
Fixed in version Unknown


ignore external/interwiki links, images, nowiki, math...

Status New
Description that option ignores also the link description like [http://www.sejm.gov.pl/poslowie/posel6/316.htm strona sejmowa posła VI kadencji], didn't ignore it some some ago --84.234.42.68 (talk) 14:35, 18 April 2008 (UTC)[reply]
To duplicate:
Operating system XP SP2
.NET FW Version Unknown
AWB version rev. 2345
Workaround
Fixed in version Unknown


Retrieved from "https://en.wikipedia.org/enwiki/w/index.php?title=Wikipedia_talk:AutoWikiBrowser/Bugs&oldid=206496325"