Log file opened at: 2004-05-26 00:27:12

[00:31:58:] <yod> Agenda: http://esw.w3.org/topic/QaDev
[00:38:51:] <bjoern_> fine, we should probably start with css validator then
[00:39:58:] <yod> -  Remove platform/os fields in Bugzilla, if possible?
[00:40:09:] <yod> - AssignedTo in Bugzilla to Yves?
[00:40:18:] <yod> -  www-validator-cvs as QA contact in bugzilla?
[00:40:36:] <yod> - ACTION: Bjoern to modify priorities in ?CssValidator's bugzilla
[00:41:11:] <yod> ... for the (default) QA Contact, I guess it makes sense, it's just a matter of setting it up
[00:41:27:] <yod> unless anyone sees a problem with it, I'll do it
[00:42:11:] <yod> ACTION: olivier to set up QA contact for CSS validator components
[00:42:21:] <bjoern_> http://www.w3.org/Bugs/Public/buglist.cgi?product=CSSValidator
[00:42:34:] <bjoern_> I've modified the priorities and severity of various bugs
[00:42:47:] <bjoern_> not yet finished with all of them though
[00:43:27:] <yod> looks good... I see the problem with the defaul owner too
[00:43:55:] <bjoern_> it basically makes little sense to bug plh with lots of mails
[00:44:10:] <bjoern_> there are some bugs that i think are already fixed, http://www.w3.org/Bugs/Public/show_bug.cgi?id=294 for example
[00:44:23:] <bjoern_> __Yves, could you confirm that 294 is fixed?
[00:44:46:] <bjoern_> (the css validator ignored charset parameter in content-type header)
[00:45:09:] <yod>  /me could take default ownership of some components
[00:45:40:] <yod> I am not sure the current components division is the best possible, though
[00:47:07:] <bjoern_> regarding components, I guess it would make sense to discuss this per mail
[00:47:58:] <yod> sounds good... do you think the division could be better, too?
[00:48:48:] <bjoern_> I agree that it is not perfect, not sure how to improve it though. some components are missing, e.g. CSS 3.0
[00:49:17:] * yod notes he has not yet completed action item related to CSS3 support
[00:49:22:] <yod> will do urgently
[00:49:40:] <yod> ok, we'll discuss that further by e-mail then
[00:49:59:] <bjoern_> regarding my test cases action item, all my test cases should be in bugzilla, either in comments or in the URL field
[00:50:54:] <bjoern_> I would suggest you just go through the reports and take them as you think is appropriate
[00:51:03:] <bjoern_> unlikely that I'll get around to do that soon
[00:51:38:] <yod> hmm ok
[00:52:00:] <yod> next (or previous) on the list is
[00:52:09:] <yod> -  long style sheets in textarea vs. browser limitations on URI length, can we use POST instead?
[00:52:21:] <bjoern_> this is http://www.w3.org/Bugs/Public/show_bug.cgi?id=177
[00:52:42:] <bjoern_> browers support only a limited URI length, so if you have a long style sheet, e.g. IE will not submit the form at all
[00:52:51:] <bjoern_> which is quite confusing for users
[00:53:13:] <bjoern_> (the GET interface is nevertheless very useful and should be kept)
[00:55:18:] <yod> bjoern_: any idea how to have both to coexist?
[00:55:24:] <bjoern_> along with http://www.w3.org/Bugs/Public/show_bug.cgi?id=152 it is probably what most users complain about
[00:55:38:] * bjoern_ not that deep into the code
[00:55:51:] <bjoern_> though it should be simple
[00:56:11:] <yod> 152 should really not be difficult either
[00:56:26:] <yod> at least I don't see why
[00:57:13:] <niq> but 152 won't resolve included stylesheets unless they have abs url ...
[00:57:39:] <bjoern_> yes, there are certainly some usability issues
[00:57:51:] <niq> so fixing it generates a new problem with confsed users
[00:58:07:] <bjoern_> well, you would mention that as documentation near the form
[00:58:35:] <niq> and when has that ever stopped users complaining?
[00:58:46:] <bjoern_> it might also be a good first step to say more explicitly that markup is not supported for these forms
[00:59:05:] <bjoern_> they complain because they do not understand the resulting error messages
[00:59:20:] <bjoern_> you get "Unrecognized <!DOCTYPE ..." 
[00:59:20:] <yod> markup not supported: can do
[00:59:45:] <yod> though as usual we have the translated interface getting more and more out of sync
[00:59:46:] <niq> hmmm ...
[01:00:23:] <bjoern_> I guess you can maintain the .fr version a bit and i the .de version
[01:00:44:] <bjoern_> and feel free to take the .jp version, too :)
[01:01:09:] <niq> how many languages?
[01:01:12:] <yod> yes... and I have volunteers to maintain es and it
[01:01:46:] <yod> niq : some parts have more languages than other
[01:02:10:] <yod> total : en fr it de es ja zh-cn zh-tw ru nl 
[01:02:15:] <bjoern_> niq, .zh-cn, .nl, .ja, 
[01:02:35:] <bjoern_> no "ru" actually
[01:02:43:] <yod> I have to upload them
[01:02:49:] <bjoern_> aha, ok
[01:02:51:] <yod> someone sent me a translation a while ago
[01:02:57:] <yod> same for zh-tw
[01:02:58:] <bjoern_> only the interface or the messages/etc. also?
[01:03:04:] <yod> just the interface
[01:03:34:] <yod> but I can contact the same people - not that I have their e-mail
[01:03:36:] <yod> 0:)
[01:04:30:] <bjoern_> in doubt just cry for help on w3c-translators :)
[01:04:39:] <bjoern_> so i don't think this is too much of an issue here
[01:04:58:] <yod> indeed, probably OK
[01:05:40:] <yod> so ACTION olivier to mention non support of HTML in upload/textarea in a less subtle way
[01:05:51:] <yod> (until they're actually supported)
[01:06:43:] <bjoern_> ok, what about "Remove platform/os fields in Bugzilla, if possible?"?
[01:07:11:] <bjoern_> for the css validator these make little sense 
[01:07:24:] <yod> not sure that's something we can configure with bugzilla
[01:07:47:] <bjoern_> could you have a look?
[01:07:57:] <__Yves> sorry, got an emergency here
[01:08:06:] <__Yves> 294 is fixed
[01:08:42:] <bjoern_> ok, marked as fixed
[01:08:49:] <bjoern_> thanks __Yves
[01:09:07:] <__Yves> I sent some test cases to Olivier already, might be good to write a test for that as well :)
[01:09:14:] <niq> shouldn't take away platform: it's relevant to people installing.  Add an "online@w3" platform for the online services.
[01:10:13:] <yod> platforms are hardcoded in bugzilla I reckon
[01:10:20:] <niq> hmm
[01:10:28:] <niq> bugzilla--
[01:10:32:] <yod> well, yeah
[01:11:02:] <bjoern_> __Yves, would you take an AI to look into http://www.w3.org/Bugs/Public/show_bug.cgi?id=177 ?
[01:11:14:] <yod> anyway, I'll check whether/how this can be tuned, but I have little hope
[01:11:17:] <bjoern_> (textarea POST/GET)
[01:11:39:] <__Yves> regarding earlier comment, yod when you send me files, give me the encoding as well, so that I can update it on the server
[01:12:21:] <yod> Yves : yes. Should we agree on a default of utf8 for everything I send, or is it better to mention anyway?
[01:12:52:] <__Yves> bjoern: 177, maybe it already work if the form is changed to use POST instead of GET, but not sure as it might interact with file upload, but I will try to take a look at it
[01:13:07:] <__Yves> yod: always good to mention, otherwise I may forget to change
[01:13:15:] <yod> okay
[01:13:34:] <bjoern_> __Yves, cool
[01:13:56:] <bjoern_> what do we do for the default owner?
[01:14:11:] <bjoern_> currenly plh is default owner for everything
[01:14:16:] <__Yves> qa team as the default owner? olivier perhaps?
[01:14:40:] <__Yves> if olivier is ok to prioritize things and harass me afterwards :)
[01:14:55:] <yod> I'm fine with that, though I suspect I will do little more than 'dispatch'
[01:14:57:] <bjoern_> that would work for me :)
[01:15:03:] <__Yves> dispatching is fine
[01:15:15:] <yod> fine then
[01:15:50:] <bjoern_> Pre-compiled packages for the css validator
[01:16:00:] <bjoern_> we get an increasing number of requests for that
[01:16:01:] <yod> ... and packages, period
[01:16:12:] <yod> (i.e uptodate tarballs)
[01:16:18:] <__Yves> I was thinking of specialized jigsaw distrib packages (like the old proxypackage)
[01:16:43:] <__Yves> so maybe having a pre-configured css validator + server might be ok
[01:17:17:] <__Yves> perhaps a .war at some point even if I don't like that
[01:17:29:] <scop> __Yves, why not?
[01:17:34:] <bjoern_> I posted a win32 batch file that would pull stuff from cvs and compile to a .zip file
[01:17:34:] <bjoern_> http://www.w3.org/mid/4003283f.159972538@smtp.bjoern.hoehrmann.de
[01:17:49:] <__Yves> the format it not abstract enough, very heavily linked to apache config
[01:17:52:] <__Yves> the web is NOT apache
[01:18:15:] <bjoern_> the package would be ready to stuff it into your fav. web server
[01:18:21:] <bjoern_> or to run from the command line, java -cp validator-snapshot.zip org.w3c.css.css.StyleSheetCom http://www.example.org/
[01:18:26:] <__Yves> having a jar with a main would be good as well
[01:18:38:] <__Yves> java -jar cssvalidator.jar url 
[01:18:47:] <bjoern_> even better, yes
[01:18:48:] <__Yves> it depends on the usage
[01:19:25:] <__Yves> I can work on that
[01:19:38:] <bjoern_> ha, he took another action item :)
[01:19:41:] <__Yves> (well after I did my 4 publications)
[01:20:23:] <yod> what would be the priority of this vs bug fixes?
[01:20:37:] <yod> (the ones on top of the list at least)
[01:20:47:] <__Yves> depends on people's will
[01:21:05:] <bjoern_> Since bug fixes affect more users, those should go first
[01:21:45:] <bjoern_> and then there is http://www.websitedev.de/css/validator-faq
[01:22:02:] <bjoern_> most of these are fatal error where the validator just emits a line of error message
[01:22:12:] <yod> agreed re-bugs, though I suspect the "market" for a well distributed css validator is bigger than we think
[01:22:13:] <bjoern_> I am not sure where to integrate something more useful
[01:22:43:] <bjoern_> I basically would like to have foo.bar entries in the property files for these
[01:22:43:] <__Yves> yeah post-processing errors needs improvements
[01:23:06:] <__Yves> agreed
[01:23:17:] <bjoern_> we have the system for normal error messages, but I think these fatal errors are not currently integrated into that
[01:23:26:] <__Yves> code will be easy
[01:23:34:] <bjoern_> cool :)
[01:23:36:] <yod> having them in property files ++
[01:23:59:] <yod> if the syntax makes it possible
[01:24:28:] <bjoern_> it should. once we got the identifiers, yod and I can work on more useful messages
[01:24:35:] <__Yves> I have to drive back home (before the next 2 teleconferences), see you in ~20mn
[01:24:36:] <yod> should we enter that as a new "bug"?
[01:24:44:] <yod> later Yves
[01:25:10:] <bjoern_> we've got one on this, like http://www.w3.org/Bugs/Public/show_bug.cgi?id=232
[01:25:26:] <bjoern_> I am not too happy with using a Bug tracker for feature requests or a todo list
[01:25:47:] <bjoern_> I think this is not needed for this particular issue
[01:27:19:] <bjoern_> we could just mark it as an AI for __Yves to look into that (and for us to work on the messages once __Yves is done)
[01:28:14:] <yod> ACTION: Yves to work on way to add error message explanations
[01:28:39:] <yod> ACTION: Bjoern/Olivier to work on error messages for CSS validator



[01:30:08:] <yod> =============
[01:30:10:] <yod>  ACTION: Olivier to harmonize checklink's layout with ?MarkUp Validator -> proposal 
[01:30:27:] <bjoern_> proposal at http://lists.w3.org/Archives/Public/public-qa-dev/2004May/0044.html
[01:30:38:] <yod> This is (AFAIK) the last blocker for a new beta of checklink
[01:30:59:] <scop> not quite, the cookie fixes aren't in yet...
[01:31:29:] <bjoern_> scop, did you have a look at yod's proposal?
[01:31:43:] <scop> bjoern_, looking into it now
[01:33:23:] <scop> as not all of the pages are dynamic, dunno if it's a good idea to make anything "configurable" at all
[01:33:32:] <scop> (even though I did suggest it)
[01:35:10:] <bjoern_> my concern is just that it currently does not integrate nicely, 
[01:35:29:] <bjoern_> if you go through the validator navbar you get lost on checklink
[01:35:33:] <yod> then basically the current checklink file structure does not really allow us to have "nice" navigation
[01:36:20:] <yod> unless we fall back to a simple solution of having links hardcoded (defaulting?) to the documents/service at v.w.o
[01:36:22:] <bjoern_> it would work for me if we have just the markup validator header there
[01:36:38:] <scop> yes, needs to integrate better.  h/c'ing for 3.9.3 is fine with me
[01:36:43:] <yod> just the header, not the navbar you mean?
[01:37:20:] <bjoern_> the header would need to read link checker; the navbar too of course
[01:37:48:] <yod> ok
[01:38:14:] <yod> we seem to agree on the (temporary) course to follow
[01:38:38:] <yod> I'll give this another shot as soon as CVS is usable again
[01:40:49:] <scop> navbar link suggestions for wlc 3.9.3:
[01:40:55:] <scop> homepage: http://validator.w3.org/checklink
[01:41:00:] <scop> download: http://search.cpan.org/dist/W3C-LinkChecker/
[01:41:05:] <scop> (...or http://www.w3.org/2000/07/checklink#install)
[01:41:08:] <scop> docs: http://www.w3.org/2000/07/checklink
[01:42:30:] * yod notes http://www.w3.org/2000/07/checklink redirects to v.w.o
[01:43:35:] <scop> yep, but 2000/07/checklink is what's used elsewhere for linking to docs
[01:43:44:] <yod> oh OK
[01:43:56:] <scop> does not really matter to me, as long as it's consistent
[01:43:59:] <yod> then maybe I should move the redirect the other way around?
[01:44:21:] * yod no strong opinion
[01:44:44:] <scop> http://validator.w3.org/docs/checklink looks slightly better, and avoids a redirect...
[01:44:58:] <yod> OK
[01:45:13:] <yod> I'll finish my action item then
[01:45:58:] <bjoern_> ok, next?
[01:45:59:] <yod> next action item not discussed yet...
[01:46:15:] <yod> mod_validator: that's done, right?
[01:46:22:] <bjoern_> niq?
[01:47:08:] <niq> mod_validator installed but some problems.  Will post to list.
[01:47:18:] <bjoern_> ok
[01:47:35:] <bjoern_> I did not send my reservations yet, will do soon.
[01:47:48:] <bjoern_> niq, scop, you got a public w3c account?
[01:50:02:] <bjoern_> niq, http://cgi.w3.org/MemberAccess/Public
[01:50:22:] <yod> public accounts are web accounts useful for... stuff
[01:50:38:] <yod> they don't give particular access right though, only a way to be identified
[01:50:58:] <niq> ah, separate?
[01:51:09:] <bjoern_> yes, for strawpolls and stuff


[01:55:41:] <yod> ===============
[01:55:46:] <yod> Markup Validator
[01:56:50:] <yod> transparency of icons... why was I thinking this had been fixed long ago?
[01:57:05:] <yod> (useful state of denial, perhaps)
[01:57:13:] <bjoern_> that i do not know, but on http://www.w3.org/2000/09/vsimg/transparency-test.html it is clearly broken
[01:57:26:] <bjoern_> for current and legacy
[01:57:31:] <niq> is that really wrong, or is it just tickling a browser bug?  Looks OK here ...
[01:58:03:] <yod> looks OK with my setup too
[01:58:09:] <bjoern_> in IE the HTML 4.01 PNG badge is different from the other PNG badges
[01:58:42:] <bjoern_> it might be using alpha transp. while the others use the simpler transp. means in PNG
[01:58:55:] <yod> possible
[01:59:07:] <niq> what's the simpler means?
[01:59:20:] * niq has been reading PNG spec:-(
[01:59:21:] <bjoern_> color $color means transparent
[01:59:29:] <niq> ah
[01:59:36:] <scop> vcss.png has a thin white vertical line on the right side
[01:59:59:] <bjoern_> yeah, right, that is borken too, good catch
[02:00:06:] <scop> ditto for all "valid css" png's
[02:00:28:] <scop> s/all/the other/
[02:00:57:] <bjoern_> http://www.bjoernsworld.de/temp/badge-trans.png
[02:01:36:] <yod> bjoern_: ugh
[02:01:37:] <bjoern_> it would also be nice if the page gets updated to include e.g. the XHTML Basic 1.0 icon
[02:01:58:] <niq> so who "does" graphix, or should contribs be invited?
[02:02:10:] <bjoern_> Comm Team
[02:02:30:] <yod> icons are the property of W3C brand defense division
[02:02:59:] <yod> Dom took responsibility for liaison with the Comm a wile ago, (basically after I gave up)
[02:03:08:] *** Dorward (~david@dorward.demon.co.uk) has joined the channel
[02:03:14:] <scop> ISO-HTML icon to transparency-test too?
[02:03:17:] <yod> I could ping him againand get this sorted out... perhaps
[02:03:42:] <yod> ACTION: olivier add new icons to transparency test page
[02:04:16:] <yod> ACTION: dom get new (transparent, no alpha) versions of HTML401 and XHTMLBasic10 icons
[02:04:43:] <bjoern_> s/XHTMLBasic10/CSS/
[02:04:51:] <bjoern_> I think the basic icon is not broken
[02:04:59:] <bjoern_> but the css png icon is
[02:05:06:] <yod> isn't the basic a quick hack?
[02:05:23:] <bjoern_> it certainly looks like one, yes
[02:05:44:] <bjoern_> http://validator.w3.org/images/vxhtml-basic10.gif
[02:05:54:] <yod> yes
[02:06:06:] <bjoern_> and we need a XHTML Print 1.0 (currently in CR) Icon
[02:06:08:] <yod> and it's not in http://www.w3.org/Icons
[02:06:18:] <scop> XML 1.1?
[02:06:37:] <bjoern_> aaahh, wishlist :-)
[02:06:47:] <bjoern_> http://images.google.com/images?q=%22valid+xhtml%22&imgsz=icon
[02:07:18:] <yod> bjoern :)
[02:07:42:] <bjoern_> this is indeed an issue, people want better icons and since W3C resists they violate w3c's icon usage guidelines...
[02:08:09:] <bjoern_> but that is probably out of scope for this meeting...
[02:08:23:] <yod> not an issue *we* are likely to resolve
[02:08:30:] <bjoern_> yes
[02:08:48:] <bjoern_> hmm, at a closer look, I also note that the W3C blue is different in these icons
[02:08:54:] <bjoern_> some have the old blue, some the new
[02:09:10:] * yod not giving a very good example and using such a "custom" icon on personal HP
[02:09:13:] <bjoern_> HTML 3.2 and 4.0 icons
[02:09:22:] <bjoern_> hehe
[02:10:13:] <yod> indeed, different blue's
[02:10:21:] <bjoern_> anyway, I think we should move on for now
[02:10:27:] <yod> methinks the whole batch should be regenerated
[02:10:35:] <yod> I'll talk with Dom about it
[02:10:41:] <bjoern_> cool
[02:10:46:] <bjoern_> SVG Icons! :)
[02:11:27:] <yod> the banana is too big...
[02:11:46:] <bjoern_> drop PNGs and use JPEG?
[02:11:49:] <niq> and .. erm ...
[02:12:03:] <_Yves> use jpeg and the blue will be different on mac and pcs
[02:12:39:] <bjoern_> there is not much blue in the bana images (http://validator.w3.org/ background images)
[02:12:55:] <_Yves> ah ok those ones
[02:13:17:] <_Yves> same issue, jpeg2000 or png with gamma chunk, otherwise you won't have a consistent rendering
[02:13:26:] <_Yves> but that's out of scope I guess :)
[02:14:01:] <bjoern_> http://validator.w3.org/images/header.png
[02:14:06:] <bjoern_> http://validator.w3.org/images/header.jpg
[02:14:10:] <yod> jpeg would solve size issue
[02:14:11:] <bjoern_> 30kb vs 6kb
[02:14:36:] <bjoern_> the problem is that the server would deliver PNG if supported...
[02:14:55:] <bjoern_> http://validator.w3.org/images/footer.png is 52k withput JPEG alternative
[02:15:07:] <yod> not if CSS states .jpg
[02:15:18:] <bjoern_> yes, of course
[02:15:52:] <yod> given that there are cache issues with them, I suppose that's the only reasonable option, actually
[02:16:06:] <bjoern_> agreed.
[02:16:23:] <yod> a 80k download for every page is, uh...
[02:16:45:] <yod> IE-- for the dumb cache, and for the record :)
[02:17:36:] <yod> Re - "Malformed multipart POST"
[02:17:36:] <bjoern_> so we use JPEGs instead?
[02:17:42:] <yod> yes
[02:17:58:] <bjoern_> ok, and I guess you will create the JPEGs then :)
[02:18:11:] <yod> RESOLVED: create and use jpegs for validator's stylesheet
[02:18:52:] <bjoern_> [14:28]<bjoern_> "Malformed multipart POST",
[02:18:53:] <bjoern_> [14:28]<bjoern_> The browser is Internet Explorer 6.0.2800.1106.xpsp2.030422-1633
[02:18:53:] <bjoern_> [14:28]<bjoern_> Operating system is Windows XP Professional SP1
[02:19:04:] <bjoern_> [14:28]<bjoern_> It was just a basic upload through the validator which failed first
[02:19:05:] <bjoern_> [14:28]<bjoern_> time, but appeared to work on the second attempt straight after.  Then
[02:19:05:] <bjoern_> [14:28]<bjoern_> it failed again in a new browser session, I think, but again, the
[02:19:05:] <bjoern_> [14:28]<bjoern_> subsequent attempts in that instance were okay.
[02:19:08:] <bjoern_> [14:28]<bjoern_> I can't actually remember which document I was checking, but it did turn
[02:19:08:] <bjoern_> [14:28]<bjoern_> out to be XHTML 1.0 valid on the next attempt.
[02:19:08:] <bjoern_> [14:28]<bjoern_> Sorry I can't be of any more help!
[02:19:10:] <bjoern_> [14:29]<bjoern_> "
[02:19:20:] <bjoern_> (from one of the users I talked to)
[02:19:27:] <niq> bjoern_: can you run that through cg-eye live?
[02:19:44:] <bjoern_> niq, hm?
[02:19:46:] <yod> oh, I somehow thought it was an issue with opera
[02:20:11:] <bjoern_> not according to the logs scop posted to the list
[02:20:40:] <bjoern_> I guess someone will have to figure out what changed in CGI.pm and how to restore the old behavior...
[02:20:48:] <yod> opera identifies itself as IE by default, but I may have been completely mistaken
[02:20:51:] <niq> Oh, scratch that, it doesn't support file upload 
[02:20:51:] <bjoern_> it seems to be a race condition though
[02:21:28:] <scop> opera identifies as IE, but has the Opera string somewhere in User-Agent
[02:21:37:] <yod> scop: ok, thanks
[02:22:04:] <yod> bjoern_: yes, have to investigate further
[02:22:23:] <yod> it doesn't seem possible to just go back to the old CGI.pm though, as it is tied to the version of perl used
[02:22:43:] <bjoern_> yeah, we probably want to coordinate with l.d.s
[02:22:51:] <scop> from CGI.pm 3.05 changes:
[02:22:58:] <scop> # Fixed incorrect processing of multipart form fields that contain embedded quotes. There's still the issue of how to handle ones that contain embedded semicolons, but no one has complained (yet). 
[02:23:31:] <scop> http://search.cpan.org/src/LDS/CGI.pm-3.05/cgi_docs.html#new
[02:24:23:] <yod> interesting
[02:24:43:] <yod> seems like a good lead! :)
[02:25:39:] <bjoern_> maybe we should complain then...
[02:25:50:] <bjoern_> though I am not sure whether that's the problem here
[02:25:54:] <yod> to? LStein?
[02:26:27:] <bjoern_> http://search.cpan.org/src/LDS/CGI.pm-3.05/cgi_docs.html#bugs
[02:26:39:] <bjoern_> so, yes
[02:27:34:] <yod> We could/should try 3.05 first, though
[02:28:26:] <yod> except that it's not easy to reproduce
[02:28:27:] <bjoern_> another "release"?
[02:29:05:] <yod> not sure I understand?
[02:29:07:] <bjoern_> we should notice whether this addresses the issue trough our logs
[02:29:18:] <bjoern_> well, how would we try 3.05?
[02:29:44:] <scop> btw there are already 138 "Malformed multipart POST"s today (only 13å¸ hours on v.w.o)
[02:30:41:] <bjoern_> due we run anything that would give us the actual request? ScriptLog or something?
[02:30:48:] <yod> we could install 3.05, bind :8001 to it and get some people to try it
[02:31:09:] <bjoern_> maybe we could eval { } CGI.pm and store the request to a file?
[02:31:20:] <bjoern_> in case it fails?
[02:33:07:] <yod> hmm
[02:33:09:] <yod> dunno
[02:33:53:] <yod> beyond the technical aspect, I'm not sure our privacy policy covers this
[02:34:12:] <yod> and this is a tricky subject
[02:34:57:] <yod> I suggest we continue by e-mail
[02:35:28:] <yod> so that we can finish the meeting reasonably soon
[02:35:31:] <yod> agreed?
[02:35:35:] <bjoern_> ok
[02:35:52:] <scop> ok
[02:35:55:] <bjoern_> "replacement for web-human@ in case of fatal errors"?
[02:36:14:] <scop> some related bits to the previous: i've a couple of warning-silencing patches waiting for dev
[02:36:43:] <scop> as well as been toying with installing better signal handlers to get better warning/error log entries
[02:36:58:] <yod> scop: cool
[02:37:36:] <yod> bjoern_: I'm not too fond of shifting from general policy of having webmaster dispatch from web-human, but if he wants it, we can't really refuse
[02:38:13:] <yod> we could create an alias or list for that purpose
[02:38:34:] <bjoern_> Would it be possible that we mention www-validator@ in addition to web-human and point out that this is pulic/archived and much more likely the place to get help from?
[02:39:03:] <bjoern_> Custom ErrorDocument maybe?
[02:39:51:] <yod> should be feasible, yes
[02:40:23:] <bjoern_> ok
[02:40:26:] <yod> the archive approval system already takes care of telling people that the list is public
[02:40:30:] <yod> but that would not hurt
[02:40:41:] <yod> I'll check the config for ErrorDoc 500
[02:40:49:] <bjoern_> Pointer to the archive to avoid duplicates and such
[02:40:55:] <scop> custom errordoc may not play nice with CGI::Carp /fatalsToBrowser
[02:41:18:] <scop> but then again, maybe that'll be scratched anyway if we install signal handlers of our own
[02:41:54:] <bjoern_> we could get rid of fatalstobrowser...
[02:42:09:] <scop> and since the "web-human@" error doc is shown nowadays, it means that fatalsToBrowser isn't working in all cases anyway
[02:42:38:] <bjoern_> yeah, in fact, we are only talking about that case
[02:42:57:] <bjoern_> other such cases should be of less concern
[02:43:06:] <scop> right
[02:43:37:] <bjoern_> so I think that is one more AI for yod :-) and we should adjourn...
[02:44:18:] <yod> Action: olivier to modify ErrorDocument 500 to also mention w-v (public)
[02:44:56:] <yod> adjourning++, I have a user with a ssh problem to help before I can hope to go to bed
[02:45:18:] <scop> adjourn++, thanks
[02:45:51:] <bjoern_> yod will send stuff to list and we wil update the wiki, etc.
[02:46:14:] <yod> ok, thanks all for joining
[02:46:19:] <yod> adjourned
[02:46:22:] <bjoern_> next meeting is probably 8th june