[01:27] Any movement on any of the AIs? [01:27] hmm, on my side yes [01:28] got things moving a little on the CSS validator front [01:28] and the privacy/p3p policies are almost ready [01:29] Yves has a tiny bit more time, so I'm helping him pick bugs [01:29] any news on FC4? [01:30] FC like in Fedora Core? FC3 isn't even out yet... [01:30] FC3t3 came out today; any movement on this will happen after FC3 is out. [01:30] (ACTION: Terje, Ville; go pester RH Bugzilla about getting openjade/opensp packages split for FC4) [01:31] ok [01:31] And the obvious way to do it is to get a 1.5.2 release and then start adding comments to the Bug to remind the Owner about it. [01:32] Axel Thimm has done some work in separating OJ and OSP, for FCx, that could probably be (re-)used: http://atrpms.net/dist/fc2/openjade/ [01:32] Perhaps we should dump specfile patches in the bug? [01:33] yes, and hopefully after FC3 the FC CVS opens, and I'd probably be able to even commit something [01:33] Yeah, right. And pigs will fly... :-) [01:34] :) due to the way things are organized, the package split propagates changes into many others too (dependencies) [01:34] (IIRC, that is) [01:35] that's likely [01:35] Hmmm. OJ Requires OSP and BuildRequires OSP-devel. Isn't that fairly self-contained? [01:36] Everything else just depends on OJ, non? [01:36] as a result, things may depend on OJ that should only depend on OSP? [01:36] not that it's a terrible problem [01:37] It's not even something that needs to be solved before execution; it can be be (an similar issues are) fixed as errata later on. [01:38] * xover_ looks questioningly at scop... [01:38] sorry, distracted... [01:38] --- xover_ is now known as xover [01:39] ...yes, I believe you are right [01:39] but we'll see, it's definitely doable and overdue. hopefully OSP 1.5.2 is out by then [01:40] Which means I and Bjoern need to get cracking on that issue. [01:41] by the way, I'm having some issues building latest from OSP 1_5 branch, does someone of "us" have a commit account there [01:41] * xover raises hand... [01:41] xover and bjoern do [01:42] As does Bjoern and possibly Nick; and I'm an admin so can set you up if you want it. [01:42] cool. will ping you guys then (rather than subscribe to yet-another-mailing-list if it's ok with you) [01:42] Sure. [01:42] at this point, I don't think I need commit access there [01:43] ok. on to badges then? [01:43] re- Badges: nothing [01:44] I've spent literally weeks scribing, attending or preparing meetings, haven't had a chance to touch them [01:44] So current status is...? [01:44] ...except functionality missing from HEAD. unintentional, I assume? [01:44] scop: EXPN? [01:44] http://www.w3.org/Bugs/Public/show_bug.cgi?id=915 [01:44] current status is: waiting for yod to merge designs and submit it to parties concerned [01:45] badges not showing up: different issue, though relevant [01:45] I know a few persons that would like that strike to be a neverending one :) [01:47] fwiw, to some extent, I actually do agree with Jukka. favelets-like functionality is better [01:47] which reminds me... should we suggest the use of favelets in the results page? [01:47] the favelets are relatively well hidden [01:47] not on purpose, but still [01:48] Bump visibility of Tools and Resources relevant on the web site in general (including output pages). [01:49] I guess that'd be one goal of a design overhaul [mention of article at apple.com] [01:51] Speaking of that article... "Someone" had an AI to follow up with the author about versions and such... [01:51] I am to ping David whenever we're about to re-release [01:52] Ok, so their current state is stock 0.6.7, no patches? [01:52] yes [01:52] Excellent! [01:53] * xover adds "Ping yod about Pinging David" to his mental Release Checklist... [01:54] oh, one more thing about AIs [01:55] apparently vivien is still receiving reports of "malformed multipart" [01:55] which makes the AI on adding a link to bugzilla more urgent [01:55] Hmmm. Ok. Do we want to make ED500 a CGI that captures browser and referer info? [01:56] xover: ED500? [01:56] ErrorDocument for "500 Internal Server Error". [01:56] that's one idea [01:57] capture referrer, date, browser, query [01:57] That gives us some data to debug any similar problems. [01:57] anything else? [01:57] I have Opera suspected for the malformed multipart stuff. [01:57] (And we can probably lean on Howcome to fix any bugs there that we may uncover) [01:58] I have some hacks around that would add browser/query sanely to _all_ error log messages [01:59] So ACTION yod and scop to whip up something for this then? [02:00] ok [02:00] sounds good [02:01] IIRC those hacks would mean that CGI::Carp needs to go. not a problem, I assume? [02:01] I _like_ CGI::Carp! [02:02] I knew we had run into some kind of catch in the past about this issue [02:03] Ok, 500 is user facing, CGI::Carp is a devel issue... Nuke it and we'll take it from there! [02:03] scop, could you detail how CGI::Carp interferes with the ErrorDocument apache directive? [02:04] I don't remember the details now, will post info to list later [02:05] scop, thanks [02:07] scop, would you mind if we discuss which bugs are on the shortlist for 4.1? [02:07] ok [02:08] There is only one bug (robots.txt cache) with 4.1 as target [02:08] timeout (#896) there as well now [02:08] indeed [02:09] expect a candidate fix for timeout problem in CVS tonight [02:10] most of the other "bugs" in the db are actually enhancement requests [02:10] I'm wondering whether to try and release the bugfix version first and focus on enh later? [02:10] exactly [02:10] ah there is 113 too [02:10] forgot about this one [02:11] regarding robots.txt cache, WWW::RobotRules is buggy: http://www.xray.mpe.mpg.de/mailing-lists/libwww-perl/2004-10/msg00034.html [02:12] I think bug 113 would be better off left until the next major restructured version [02:12] I saw your mail on this issue [02:12] agreed on 113 [02:13] another possibility to "fix" the caching would be to use a global W3C::UserAgent (a RobotUA), but that's even uglier [02:13] re- RobotRules, that would mean we don't have any easy fix for our problem, and would be conditioned to the release of a new LWP? [02:13] hmm [02:14] there are some options, see one above, as well as switching to WWW::RobotRules::AnyDBM_File or writing our own subclass of WWW::RobotRules [02:14] would that allow us to tweak the RobotUA in other useful ways? [02:14] if that would be global? [02:14] yes [02:15] dunno, can't think of any right now. do you have something specific in mind? [02:16] I am still mildly annoyed by the fact that we can't turn off the delay for a specific domain [02:17] and therefore wondering if such a change would give us a possibility to work around [02:18] I think the issues are orthogonal. Seems like control over the delay would be a matter of overriding host_wait() in LWP::RobotUA. [02:20] (plus of course configuration for how long to wait for which hosts etc) [02:23] ok, thanks [02:23] that's about it for 4.1, I guess [02:24] so provided we find a fix for the robots.txt cacheing, we're almost ready for it? [02:24] that's quite good news [02:24] while testing the timeout stuff, I noticed that we don't catch the "user pressed stop" case, which means checklink will continue checking in the "background" [02:24] but since that hasn't caused real problems, it can be fixed later [02:25] is there a way a CGI can know that? [02:25] I confess I never though it was possible [02:25] there are some $SIG{ALRM} related snippets around the web for that. this is gray area for me too [02:26] anyway, definitely not something that must be fixed yesterday [02:27] please record your findings into bugzilla when you know a bit more about it [02:27] will do [02:28] re: the "this may take some time" message in summary mode, feel free to phrase the message and add to CVS [02:28] ok [02:29] * yod would have liked to discuss a bit about fixing some kind of a policy for the sgml-lib [02:29] (niq's input would be welcome for that though) [02:30] precisely, that's why I am wondering whether to take up this item now or not [02:30] I think we could discuss it a little and have the others follow up on the list [nick joins] [02:35] bugs 89 and 860 depend on that discussion. doesn't have to be discussed now though [02:35] * niq undertakes to catch up on what scop did with sgml-lib [02:36] not much, nuked one extra HTML 4.0, added [02:36] ...HTML 3.2 [02:36] ok [02:36] I also removed old empty parts off the repository, upon request from bjoern [02:37] it should make it easier to browse/checkout [02:37] it made me think of whether it's a good idea (TM) to add DTDs from working drafts, though [02:37] re- http://www.w3.org/Bugs/Public/show_bug.cgi?id=89#c5 : do we need REC-xhtml1-20000126? [02:38] FPI is the same so we can nuke the older one. [02:38] re- browsing, everyone's using "cvs up -P", right? [02:38] xover, thanks [02:38] -dP, even [02:39] but in some of my checkouts there are often weird things going on in the sgml-lib area [02:39] probably unrelated [02:39] in any case, I don't think we really need to keep a bunch of WD DTD in the attic forever [02:40] manually nuking them from the repo causes some error messages when people "cvs up". IMHO not big deal [02:41] yod: You're probably seeing the ietf/IETF issue (Mac OS X is case-insensitive). [02:42] ah, that could cause a "foo is in the way" error, yes [02:42] silly FS, if you ask me, but let's not go there [02:44] so, who's taking the AI to nuke the extra DTDs? [02:45] does that AI include the task of defining "extra"? [02:46] I'll take care of removing the XHTML 1.0 1st ed [02:46] I was thinking only of the xhtml1 and HTML4 (?) [02:46] * niq just cvs up-ing [02:47] ah, ok. HTML 4 done, I promised to do XHTML, so AI(scop) [02:47] great [02:48] More pressing in that regard is adding back in MathML, SVG, SMIL et al. [02:49] where are they? removed by mistake in the previous cleanup? [02:50] we're missing some older HTML 2'ish DTDs which are listed in sgml.soc (html-1.dtd, html-0.dtd, html-s.dtd, html-1s.dtd, html-0s.dtd, html-2.1e.dtd). add DTDs or nuke from sgml.soc? [02:50] Removed more-or-less intentionally in the last cleanup. [02:50] scop: I'm not sure. I'm inclined to nuke `em and see if anyone complains. [02:50] fine with me [02:51] * scop adds that to my AI [02:51] * niq can take action to xref it to valet and htmlhelp; check discrepancies and fix if unambiguous [02:54] * yod starting to be tempted to adjourn, proceed with the cleanup, follow-up on list on agreeing upon and documenting "sgml-lib policy" ... Adjourned