00:25 whoa, load average: 63.00, 58.03, 48.66 CRITICAL 00:26 where's that? 00:26 validator-us.w3.org 00:26 ah, okay 00:26 so long as it's not qa-dev, it's not my fault... 00:27 no, and even if it was qa-dev, it wouldn't be a problem 00:28 not a big one at least 00:29 there are interesting load peaks 00:29 rather, yes. are you looking at the grapher now? 00:29 yes 00:30 seems not well balanced between -us and -jp 00:30 that's what bothering me 00:31 the thu-fri peak on -jp in particular 00:31 I wonder whether inverting the entries in the DNS would balance that 00:31 during that time, -us had almost no load... 00:32 in any case, with the horrible bandwidth to -jp, it would be hard to hammer 00:32 it might invert it... 00:33 possible 00:33 -us down to 48, -jp up to 5.7 00:34 + yod pats linux 2.6 and apache2 00:34 doing a pretty good job at hiding the pain 00:35 you know HEAD is, what, 8x as fast... 00:35 + scop [n=scop@...] has joined #validator 00:35 hello 00:35 hello scop! 00:36 sounds like we should start 00:36 hi scop 00:36 bjoern_: 8x? I knew it was much faster... but yeah, I wish its dependencies were a little easier to install, would be great to release it 00:37 8x, HEAD vs 0.7.x? 00:38 according to bjoern_, yes. Not that 0.7.x is extremely fast, anyway 00:38 cool 00:39 well, so it was at some point, I posted the benchmarks to qa-dev 00:39 perhaps compared to 0.7.0 then 00:39 which was very slow 00:40 http://www.w3.org/mid/qqsmf1dfhad4mc1ikar3ucv776ibrt5mta@hive.bjoern.hoehrmann.de has the thread 00:40 by the way, I have FC5 rpms of the 0.8 dependencies, not at all hard to install there :) 00:40 I think I got it from 0.7x r/s to 3.x r/s 00:41 quick question to begin the meeting: I saw that bjoern_ asked about minutes of previous meeting. Do you think it's worth making minutes if we try and stick to giving/receiving/closing/updating action items? 00:41 certainly if someone isn't around and doesn't idle here 00:41 I don't care much about logs vs summary, etc 00:42 ok 00:43 and whoever isn't around in one meeting implicitly volunteers to write minutes in the next one 00:43 on a "BTW" note, /me just received a "go ahead" reply from higher-up on the summer of code idea which niq reminded earler 00:44 + niq looks up again 00:44 I suddenly feel strongly we should just mail the logs... 00:44 hello niq 00:44 yeah, +1 with bjoern 00:44 I'm happy to help drafting something for soc yod 00:45 + yod doesn't really want to spend too much time on minutes unless necessary, but I guess it helps 00:45 summer of code gets useful work done. We have two of last years students who did apache projects and are now valued members of the team 00:45 wow, excellent 00:45 of course, not all of them deliver anything worthwhile. It's a numbers game ... 00:46 one idea for soc which I was thinking of today - tell me if it would make sense, would be making a perl module to have XML::LibXML behave like SPO 00:46 updates to XML::LibXML are really needed 00:46 remind me: SPO? 00:46 make it behave like SPO, not so much 00:47 both should behave like any XML::SAX parser 00:47 (SGML::Parser::OpenSP) 00:47 aha 00:47 but speaking of tools, I released http://lists.w3.org/Archives/Public/public-qa-dev/2006Apr/0022 00:47 enjoy! 00:48 + yod didn't have a chance to test it yet, don't have the right browser to test 00:48 :-) 00:48 well, that falls in several areas. The perl folks are participating 00:49 bjoern_, any reaction to the release on the svg group? 00:49 if you mean the w3c one, then not that I know of 00:49 from several member orgs though 00:49 not necessarily, was thinking of the yahoogroups list you posted it in 00:50 got some feedback on #svg, not much from yahoogroups 00:50 two "cool, thanks" 00:50 heycam made me a Batik profile 00:50 and MacDome is going to make one for webkit 00:50 chaals is looking for better data for opera 00:51 sounds like it's achieving its purpose then 00:51 tor complained about Mozilla not having a documentation and/or qa department 00:51 that's about it, I think 00:51 moz doesn't have a QA dept? 00:52 odd, I thought they had staff at the foundation for that 00:52 anyway 00:52 their users do their qa through bugzilla... 00:52 moz = dual personality. Corp != foundation. At least, AIUI 00:53 well, it's a bit complicated, yes. and yes, they have bugzilla 00:53 reminds me of another complicated org with not much of a QA dept and doing its QA work via bugzilla and public lists 00:53 but I digress 00:54 moving on to... ack? 00:54 adj? 00:54 ville, what do you expect will be the next steps for it? 00:55 I have no specific plans 00:55 has it ever been shown to www-validator, and a larger audience? 00:56 but I think it should give us a taste whether something like that is worth spending more time on 00:56 no, I don't think it's been announced anywhere in public yet 00:56 ok 00:57 I showed it around at the office, got a few "very cool, but if it's not accessible, it's not going to have a great welcome party" 00:57 hm, "not accessible"? 00:58 + bjoern_ gets uninitialized uninitialized uninitialized uninitialized uninitialized ... 00:58 not accessible, as in "what about browsers without js" 00:58 bjoern, which browser? 00:58 or ActiveX disabled... 00:58 the one where you can disable ActiveX 00:58 ah, that one 00:59 bjoern_, which version? 00:59 IIRC it ran fine in 6.0SPx when I tested it some time ago 00:59 yes, but I have ActiveX disabled... 00:59 bjoern_, ah, ok 01:00 so, I think bjoern_ illustrates my point :) 01:00 that said, for people who do have js/activex and so on enabled, it's a very pleasant UI 01:01 (IMHO) 01:01 I'd propose a response like "oh, yeah; how about closing the WebAPI WG then, what do you think?" 01:01 well, these are things that I'm hoping the current version would shed some light on 01:02 we could of course turn all those status indicators into iframes for browsers without activex/js ;) 01:02 JibberJim would like that... 01:02 yum, that's an idea 01:03 we could also if (scriptDisabled) { location.href = "http://checklink.w3.org" } 01:04 hehe 01:04 we shouldn't use proprietary extensions like .innerHTML 01:04 + yod stared at his screen for a while, puzzled 01:06 + bjoern_ has quit ["Quit"] 01:06 woops? 01:07 hmm, I think we were pretty much done, anyway. no sign of xover unfortunately :/ 01:07 .innerHTML is not very cool at all, indeed 01:08 current ack is using that? 01:08 yes 01:08 ah yes, I see 01:09 at one point of time it was using dom3/innerText, but then I wanted markup in the status boxes (eg. for redirects) 01:09 would probably be nicer to use dom methods there 01:09 yes 01:10 ok, so that's one part where progress could be made 01:10 shouldn't be actually that hard, I wonder why I chose to use innerHTML in the first place 01:10 what about the safari issue? 01:11 + scop has forgotten what was the conclusion 01:11 + bjoern_ [n=bjoern@...] has joined #validator 01:11 the bug is fixed in latest snapshot 01:11 good 01:11 but we may still need the workaround, because we don't know how long until said snapshot gets released 01:12 feel free to commit it (with appropriate comments in code) 01:12 will do 01:13 ACTION: yod to commit safari-friendly ack patch, with proper comment in code 01:13 welcome back bjoern_ 01:13 hello yod 01:13 another thing that should probably be done if there is an intention to announce it anywhere 01:13 would be to configure grab.pl to run in mod_perl 01:14 still no mod_perl 2.xx on qa-dev... 01:15 what's the v. on qa-dev at the moment? 01:15 1.99xxxx 01:16 mod_perl/1.999.21 01:16 Package: libapache2-mod-perl2 01:16 Version: 1.999.21-1 01:16 what joke is that? :) 01:17 .g libapache2-mod-perl2 joke 01:17 bjoern_: http://www.softwarelibre.misiones.gov.ar/index.php?option=com_content&task=view&id=2&Itemid=3 01:17 I guess it makes sense, somehow somewhere, but not when looking at numbers 01:18 that's a mod_perl 2 release candidate 01:18 computers handle numbers better than "2.0rc2" 01:18 I see 01:18 for example CPAN would probably treat 2.0rc2 as newer than 2.0 01:19 (you need a backport to get mod_perl2 on sarge, I gather) 01:20 yep, 1.999.21 probably the last 2.0rc which was even remotely API compatible with 1.2x 01:20 uhm 01:20 + yod checks what distrib qa-dev is using 01:21 my own deb box has Version: 2.0.2-2 01:21 there is apt-get dist-upgrade 01:22 hmm, qa-dev is running debian stable 01:22 we could be a bit more adventurous 01:22 we started with testing there, no? 01:22 yeah, when testing was codenamed sarge... then it became stable 01:23 qa#/etc/apt/sources.list the culprit 01:23 + yod will move it to unstable 01:23 it should be fine... debian's unstable is hardly unstable 01:24 testing would be good enough... 01:24 (nor extraordinarily cutting edge either) 01:24 + yod has heard people unhappy with testing a little while ago 01:24 I'll ask around 01:25 and then bump up the dist 01:25 it's not that we do much but running apache there... 01:25 that and a bunch of perl modules 01:25 + bjoern_ wonders what Perl is in unstable 01:26 Version: 5.8.8-3 01:26 should be fine then 01:27 I think so. I'll ask a couple of people dealing with debians all the time, they should have a good enough idea of whether to unstable or testing 01:28 +choose 01:28 anything else for today? 01:29 [silence] 01:29 let's adjourn if not 01:29 good night yod 01:30 nite 01:30 I think Jean-Gui and Damien_ will be starting soon on observer stuff, but I think the ml will be fine for that 01:30 so, adjourned 01:30 thanks all, have a good end of afternoon/evening