[12:46] asac, nss is okay now. pushed [12:53] wow [12:54] Ubulette: maybe you can try to build ffox 2.0 against our new nspr/nss ? [12:54] unless of course you want to go for xulrunner without knowing that these things work at all :) [12:54] Ubulette: i will review those packages tomorrow [12:55] Ubulette: btw, I forgot something right? [12:55] ? [12:55] I ask you ;) [12:55] did you ask me to push new paradiso? [12:56] the java fix would be nice to have for users.. yep I've provided debs to those who asked me [12:56] yet [12:56] ok ... so we want an upload [12:56] i named my stuff so that gutsy takes over automatically [12:57] ok great. [12:57] just appending ~ubulette1 to the debian revision is probably good :) [12:58] after for ff2 with my nss/nspr, should be straight forward as I haven't changed pkg names or bumper .so (yet) [12:58] s/after/as/ [12:58] yeah ... thats good. [12:58] (to test :)) [12:59] bumper, lol, I should read what I type [12:59] welll i understood ... and my irglish is probably even worse ;) [12:59] :) [01:03] if you could also merge my gp/trunk stuff into the mt trees, that would be nice too. If you're happy with my stuff of course [01:04] don't merge nss/nspr just yet [01:07] this wpasupplicant version we have is pretty buggy ... it sometimes appears to send trashed mem to nm [01:07] i still wonder why someone upgraded to 0.6.0 ... which is a development release ... instead of 0.5.8 [01:07] and yes ... let me look at gp now [01:10] hmm ... bazaar.lp appears to drop packages atm [01:11] i'm unpacking a fresh chroot for that ff2 test. Let's not mix things up [01:11] drop packages ? [01:11] yes ... better safe than sorry ;) [01:11] 179 new for ff2 builddep... woww [01:11] yes it stalls my bzr [01:12] hehe [01:12] yeah ;) [01:12] yet i've seen far bigger packages [01:12] Ubulette: https://code.launchpad.net/~fta-launchpad/firefox/granparadiso/ ... is that wrong? [01:12] it's a + [01:12] fta+ [01:12] ah [01:13] i'd like to shorten that but noone did it [01:13] i have a pending question on lp [01:13] does bzr branch https://code.launchpad.net/~fta+launchpad/firefox/granparadiso/ [01:13] work for you atm? [01:14] i just don't get anything [01:14] it just blocks [01:16] asac@hector:~/ubuntu_bzr/firefox-granparadiso$ bzr pull https://code.launchpad.net/~fta+launchpad/firefox/granparadiso/ [01:16] bzr: ERROR: Connection error: curl connection error (couldn't connect to host) [01:16] i think i won'*t be able to do that just now ;) [01:17] worked for me [01:17] right now? [01:17] tried bzr co https://code.launchpad.net/~fta+launchpad/firefox/granparadiso in a trash [01:17] crazy [01:18] ok branching works again :) [01:18] now hope that pull works as well at some point;) [01:19] \o/ [01:25] done [01:28] thx [01:28] ff2 compiling... [01:28] btw, preview works like a charm [01:28] (print) [01:29] at least, configure is happy [01:29] oh [01:29] all gutsy here [01:29] on i386 [01:29] what graphics driver? [01:29] do you use compiz? [01:29] nvidia [01:29] no compiz [01:29] non free [01:29] i have fglrx here [01:30] only thing non free that I tolerate [01:30] but i can bring up my nvidia box tomorrow [01:30] though it needs legacy nonfree drivers [01:30] because its old [01:30] geforce 3 ti200 [01:30] or something like that [01:30] gs7600 here [01:31] maybe some X setting? [01:31] dont know. everything else is ok [01:32] i already hate trackerd [01:32] i got that with the upgrade today [01:32] and its rumbling since 5 hours or so [01:33] FF2 preview is fine [01:33] whatever frontend might exist for that (no clue) ... it must be tremendous to justify this pain [01:33] cool thing is that because we changed app-name we can now run ffox and paradiso at the same time [01:33] nice [01:34] I've removed trackerd.. it sucked my cpu all the time and I keep hearing it cratching my disks [01:34] do you miss the borders as well of location bar drop down? [01:34] Ubulette: yes ... i hope it does this only once [01:35] and then only incremental indexing [01:35] I let it run for a week [01:35] without reboot? [01:35] yes [01:35] and it still indexed? [01:36] and I have ~4 millions files [01:36] from what i read in manpage it just indexes home [01:36] it sucked >1G for the indexes [01:37] i think developers need to exclude the source code trees [01:37] but then ... would be nice to search for functions and symbols :) [01:38] in home, I only my own sources, not sources for stuff I compile or package [01:38] do you have any idea which frontend to use to search these indexes? [01:38] there's one in the menu [01:38] i looked if there is a gnome panel applet or something else (similar to desktop search) [01:38] and there's nautilus [01:39] ah ok [01:39] so seach for files [01:39] ... ok [01:39] nautilus or applications / accessories / tracker search tool [01:40] oh, it indexed my pdf too [01:41] there should be at least a gnome panel applet imo ... i mean it consumes loads of resources and should be added to a prominent place ... otherwise its wasted [01:41] i agree [01:42] gasp, starting the ui restarted trackerd [01:42] hehe [01:43] boom, my 2nd cpu is over the roof [01:43] 1st is on ff2 :) [01:45] its not that bad ... at least it finds symbols in mozilla source pretty fast [01:45] if it would ever stop to index then i would be happy to try this ;) [01:45] :) [01:46] unfortunately the GUI is probably written python or something [01:46] oh, System / preferences / indexing preferences [01:46] so i won't be willing to add features like improved code browsing :) [01:47] you don't like python ? [01:47] well ... its not that i don't like python ... but i love headers :) [01:47] and compile time checks [01:48] and python code is often written in a fashion I don't like too (from OO perspective) [01:48] i tried to implement things in python ... its almost the first language that constantly fails go pop up a decent API documentation in google [01:49] i have no idea how the python community codes ... but for me it appears as if they don't use documentation :) [01:49] seems to me it's impossible to do python without docs [01:50] yet I do perl and C without anything [01:50] well ... try to search in google ... type in the module name or something [01:50] Ubulette: yes in C you have headers [01:50] if you search for the module name the first results are always mailing list posts [01:51] like "mod_python api" [01:51] you won't find the real api documentation on the first page [01:52] do you use http://www.google.com/codesearch ? [01:53] no i only use google ;) [01:53] it works for every other language i used :) [01:53] but python appears to be special :) ... so currently python is for me a battlefield of hackish people that just hack away without designing contracts et al :) (of course exaggerated) [01:54] yet python author now works for google [01:55] yeah [01:55] he's now designing python 3000 [01:55] maybe thats the reason :) [01:55] please give me types ;) [01:56] don't think he is [01:56] sure ... would be a breach of fundamental concepts [01:57] give me ";", I hate forced indentations [01:57] and { } [01:58] and i accept that ... but writing and long-term maintenance of large software products is just nothing for languages like python imo :) ... unless you are a masochist [01:58] lol [01:58] most of ubuntu stuff is in python, no ? [01:58] writing from scratch might really work out well .. but maintenence ... i doubt [01:59] yeah [01:59] but most things are not huge software products [01:59] but take a look at bughelper [01:59] that is already a unberable [01:59] its just about 10 python files ... but still its laden with implicity like hell [01:59] everything is done in constructors et al [02:00] (damn its late ... i will get nightmares because of this) [02:00] :) [02:00] and what people come up with, why they like python is usually: its cool because you don't have to care for pointers :) [02:01] I like pointers [02:01] which says all ... stay away from there because there are too many people fiddling in python code that you will always end up bumping in code that is just too insane to believe [02:01] well if people fear pointers they should stop programming imo [02:02] well, with C, Perl and shell, I have all I want. [02:02] yeah ... depends on what you do with perl [02:02] I've done tons of stuff [02:03] if you start writing application servers with perl you probably suffer from the "one tool to approach them all" symdrom :) [02:04] maybe [02:04] could try something with gp/trunk ? [02:04] +you [02:04] i have no idea if i have everything i want ... but i know that i don't want a scripting language which claims to be designed for huge software architectures ;) [02:04] sure [02:05] try to move over the url bar, without clicking [02:05] what do you see ? [02:05] yeah fonts get ... hmm i don't have the word [02:05] they shake [02:05] i already saw that [02:05] text seems to contract at dots [02:06] or at "i" (narrow letters) [02:06] hmmm for me it just contracts the first . [02:06] thats funny [02:06] try different urls [02:06] ihave the granparadiso start url [02:06] the i in firefox just doesn't move [02:06] while the text before and after moves one or two pixels forth and back [02:07] I have a localhost url (so no dot), it moves at the 'i' for firefox [02:07] well, how to find that in bugzilla... [02:08] Ubulette: look for paradiso blocker bugs [02:08] and my print preview cpu stuff too.. [02:08] how? [02:08] http://wiki.mozilla.org/Firefox3/Blockers [02:09] beta-blockers ;) [02:09] crazy ... all empty [02:09] thats strange [02:09] Zarro Boogs found. [02:09] except Unknown [02:09] well look at all unknown 1.9 and firefox 3 blockers [02:10] i think they move them to the target when they have decided what should go into beta [02:10] ff2 built. cool [02:11] sounds promissing [02:11] at best develop xulrunner against the not-renamed packages for now [02:12] mozilla bug 319196 [02:12] Mozilla bug 319196 in Startup and Profile System "customized toolbar always reset to default on restart, bookmarks and search engines lost, unable to add search engines (localstore.rdf corruption on upgrade or crash). SEE URL FIELD FOR SUPPORT LINK" [Critical,New] http://bugzilla.mozilla.org/show_bug.cgi?id=319196 [02:13] mozilla bug 368116 [02:13] Mozilla bug 368116 in GFX: Thebes "Trunk fonts are occasionally rendered off by 1 pixel, making them appear bolded or italicized" [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=368116 [02:13] that sounds interesting [02:13] ^^^ [02:13] Ubulette: maybe thats it? [02:13] for me it appeared a bit bold while it shifted left [02:14] bug 385215 [02:14] mozilla bug 385215 [02:14] Mozilla bug 385215 in GFX: Thebes "[10.5] major characters width issues" [Major,New] http://bugzilla.mozilla.org/show_bug.cgi?id=385215 [02:17] I don't see anything like 368116 [02:17] mozilla bug 362325 [02:17] Mozilla bug 362325 in GFX: Thebes "[cairo] printing is 82% slower than pre-Cairo" [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=362325 [02:17] Ubulette: does the bug claim that its static? [02:17] i mean that the boldness persists? [02:18] for me when things move left and right while hovering it appears bold for a few milliseconds [02:18] the "..." at the end of long urls is also changing [02:18] well, disappearing [02:20] welll i really think its the same effect somehow [02:21] i can nearly keep it at the "double pixel" rendered state when carefully moving the mouse to the border of the location bar [02:24] oh, forgot to say, we can drop -gnome-support in gp and trunk, deb is empty [02:25] huh? [02:25] are gnome depends in main package? [02:25] (would be a bug) [02:26] yeah they are [02:26] we should fix that [02:29] all *gnome* components should go to -gnome-support [02:29] hmm, install was fine (new ff2 compiled with my nss/nspr, and new nss/nspr installed) [02:29] but run is not fine [02:29] I get a popup [02:29] what kind? [02:29] ./components/libmozgnome.so [02:29] ./components/libnkgnomevfs.so [02:29] could not initialize the browser's security component [02:30] no name [02:30] oh [02:30] well, after that it's okay. no log though [02:30] where does the link in /usr/lib/firefox/ point to? [02:31] (not log) ? [02:31] it's not a link [02:32] lrwxrwxrwx 1 root root 17 Aug 13 08:20 libnspr4.so -> ../libnspr4.so.0d [02:32] lrwxrwxrwx 1 root root 16 Aug 13 08:20 libnss3.so -> ../libnss3.so.0d [02:32] lrwxrwxrwx 1 root root 20 Aug 13 08:20 libnssckbi.so -> ../nss/libnssckbi.so [02:32] lrwxrwxrwx 1 root root 16 Aug 13 08:20 libplc4.so -> ../libplc4.so.0d [02:32] lrwxrwxrwx 1 root root 17 Aug 13 08:20 libplds4.so -> ../libplds4.so.0d [02:32] lrwxrwxrwx 1 root root 18 Aug 13 08:20 libsmime3.so -> ../libsmime3.so.0d [02:32] lrwxrwxrwx 1 root root 20 Aug 13 08:20 libsoftokn3.so -> ../libsoftokn3.so.0d [02:32] lrwxrwxrwx 1 root root 16 Aug 13 08:20 libssl3.so -> ../libssl3.so.0d [02:32] those should be in /usr/lib/firefox/ [02:32] do they all point to a proper file? [02:33] actually ... all these are compatibility links [02:33] the only essential is libnssckbi.so -> ../nss/libnssckbi.so [02:33] iirc [02:33] l /usr/lib/firefox/libns* [02:33] lrwxrwxrwx 1 root root 17 2007-08-18 02:26 /usr/lib/firefox/libnspr4.so -> ../libnspr4.so.0d [02:33] lrwxrwxrwx 1 root root 16 2007-08-18 02:26 /usr/lib/firefox/libnss3.so -> ../libnss3.so.0d [02:33] lrwxrwxrwx 1 root root 20 2007-08-18 02:26 /usr/lib/firefox/libnssckbi.so -> ../nss/libnssckbi.so [02:34] and the oders? [02:34] others? [02:34] $ find /usr/lib/nss -ls [02:34] 6441296 4 drwxr-xr-x 2 root root 4096 Aug 18 02:25 /usr/lib/nss [02:34] 1376261 256 -rw-r--r-- 1 root root 256976 Aug 18 00:43 /usr/lib/nss/libfreebl3.so [02:34] 1376262 296 -rw-r--r-- 1 root root 298872 Aug 18 00:43 /usr/lib/nss/libnssckbi.so [02:34] 1376263 4 -rw-r--r-- 1 root root 476 Aug 18 00:43 /usr/lib/nss/libfreebl3.chk [02:34] does any of the link point into nirvana? [02:34] k [02:35] well strace -f might help then [02:35] otherwise a debug build [02:35] should spit out assertions (hopefully) [02:35] http://pastebin.mozilla.org/185859 [02:37] http://kb.mozillazine.org/Could_not_initialize_the_browser_security_component [02:38] no ssl I guess [02:39] yep, cant go to lp [02:41] SSL has been disabled. [02:41] damn [02:42] hmmm .. thats strange ... either its the patch you modified ... or something else ;) [02:43] yet firefox-bin loaded /usr/lib/libssl3.so.0d [02:43] if you just upgrade the nss/nspr binaries and using the old firefox does it work? [02:43] yes ... there are plenty of components that could lead to this [02:44] trying... I keep my nss/nspr and apt-get install --reinstall ff2 [02:45] yes [02:45] nope, same thing [02:45] ok [02:45] try [02:45] NSPR_LOG_MODULES=all:5 firefox [02:46] and see if you get any output [02:47] you can also set NSPR_LOG_FILE=/tmp/log [02:47] or something [02:48] have you tried with an existing or clean profile? [02:51] nothing interesting with NSPR_LOG_MODULES=all:5 [02:53] lol [02:53] -1222121792[805d548] : GET /update/VersionCheck.php?reqVersion=1&id=ErrorZillaMod@jaybaldwin&version=0.2&maxAppVersion=2.0.0.*&status=userEnabled&appID={ec8030f7-c20a-464f-9b0e-13a3a9e97384}&appVersion=2.0.0.6&appOS=Linux&appABI=x86-gcc3 HTTP/1.1 [02:53] thought it was disabled [02:55] no thats the extension auto-update [02:55] oh, it's an addons [02:55] which is enabled [02:55] we only disable app update [02:55] right [02:56] packaged extensions disable auto update in a different way [02:56] e.g. by a flag in install.rdf [02:56] though i think mozilla could as well check if it has write permissions and don't ask for updates if it doesn't have them to the location the extension is installe [02:56] d [02:57] but install.rdf flag is important ... because users will otherwise run as root and trash their packaged extensions :) [02:57] -1225819248[812ceb0] : nsHttpConnection::OnSocketWritable [this=b6502d48] [02:57] -1225819248[812ceb0] : writing transaction request stream [02:57] -1225819248[812ceb0] : nsSocketOutputStream::Write [this=b6503df8 count=691] [02:57] -1225819248[812ceb0] : calling PR_Write [count=691] [02:57] -1225819248[812ceb0] : PR_Write returned [n=-1] [02:57] -1225819248[812ceb0] : ErrorAccordingToNSPR [in=-12268 out=80004005] [02:57] -1225819248[812ceb0] : nsSocketTransport::OnMsgOutputClosed [this=b6503cf8 reason=80004005] [02:57] -1225819248[812ceb0] : ReadSegments returned [rv=0 read=0 sock-cond=80004005] [02:57] but that happens after the popup [02:58] hmm [02:58] welll ... either we need a debug build then or directly try to build paradiso against this [02:58] no problem from missing or misplaced files at all [02:59] but i have the feeling that paradiso will have the same isseus [02:59] Ubulette: please remove the .chk files manually [02:59] and see if that helps [02:59] e.g. just remove them from your installed binaries ;) [03:00] rm /usr/lib/nss/*.chk /usr/lib/firefox/*.chk [03:03] 80004005 = NS_ERROR_FAILURE [03:04] that kills everything? [03:04] which means nothing as it's the default error message [03:04] nono, i was reading the code [03:04] ah [03:04] have you tried removing the sign files? [03:05] no chk in /usr/lib/firefox/ [03:05] yes ... i expected that [03:05] but in nss [03:05] one [03:05] (just wasn't sure) [03:05] one? [03:05] thats strange [03:05] there should be two [03:05] at least from the error we had [03:05] it signed two files and the second failed iirc [03:05] anyway ... remove that [03:06] ./debian/libnss3-0d/usr/lib/nss/libfreebl3.chk [03:06] ./debian/libnss3-0d/usr/lib/libsoftokn3.0d.chk [03:06] oh right [03:06] yes [03:06] remove them both [03:07] no change [03:07] cool i think trackerd has finished :) [03:08] silence in my room :) [03:09] Ubulette: how does a debdiff look like? [03:09] e.g. new libnss*deb vs. old libnss*deb ? [03:12] i think i will go to bed now ... will branch your nss/nspr branches and see if i can find anything obvious [03:12] $ debdiff /var/cache/apt/archives/libnss3-0d_3.11.5-3_i386.deb /chroots/gutsy-i386-20070812-1-firefox2/src/libnss3-0d_3.12.0~alpha1b-0ubuntu1_i386.deb [03:12] File lists identical (after any substitutions) [03:12] Control files: lines which differ (wdiff format) [03:12] ------------------------------------------------ [03:12] Depends: libc6 (>= [-2.5-5),-] {+2.6-1),+} libnspr4-0d (>= [-1.8.0.10)-] {+1.8.0.10), libsqlite3-0 (>= 3.4.1)+} [03:12] Installed-Size: [-1648-] {+2688+} [03:12] Maintainer: [-Ubuntu Core Developers -] [03:12] [-Original-Maintainer:-] Maintainers of Mozilla-related packages [03:13] Source: [-nss-] {+nss-trunk+} [03:13] Version: [-3.11.5-3-] {+3.12.0~alpha1b-0ubuntu1+} [03:13] otherwise we should just try to build paradiso i think [03:14] that looks (suspicious) good :) [03:15] why does nss use libsqlite? [03:15] no idea [03:15] thats strange ... are there cvs checkout targets in rules? or how do i get the orig? [03:16] oh, maybe nspr was compiled with the wrong nss [03:16] debdiff /var/cache/apt/archives/libnspr4-0d_4.6.6-3_i386.deb /chroots/gutsy-i386-20070812-1-firefox2/src/libnspr4-0d_4.7.0~cvs20070713-0ubuntu1_i386.deb [03:16] File lists identical (after any substitutions) [03:16] Control files: lines which differ (wdiff format) [03:16] ------------------------------------------------ [03:16] Depends: libc6 (>= [-2.5-5)-] {+2.6-1)+} [03:16] Installed-Size: [-300-] {+304+} [03:16] Maintainer: [-Ubuntu Core Developers -] [03:16] [-Original-Maintainer:-] Maintainers of Mozilla-related packages [03:16] Source: [-nspr-] {+nspr-trunk+} [03:17] Version: [-4.6.6-3-] {+4.7.0~cvs20070713-0ubuntu1+} [03:17] Ubulette: no [03:17] hm, no, that's the other way around [03:17] nspr doesn't depend on nss [03:17] but in the end it shouldn't really matter [03:18] upstream claims that they haven't broken ABI for 5 years or so [03:18] (which is not exactly true) ... but thats there policy [03:18] nss devs even force firefox to switch nss branches for a stable firefox branch because they are not willing to maintain multiple branches [03:19] maybe Ive missed a (new) file [03:19] that might be the case [03:19] however ... nss and nspr should be really stable [03:20] and i think that nss and nspr packages don't pick files ... do they? [03:20] oh [03:20] they really pick them [03:20] yet, there are plenty of warning during compilation, unused variables and trivial things like that [03:20] compare the list of debian/... result with what is in packag ehten [03:20] yeah ... but thats normal [03:21] if it was that stable, that should have been fixed long ago, it's just dirty [03:21] they are probably just not used because of arch dependent preprocessor sections or something like that [03:25] btw, I now rsync my sources there: http://www.sofaraway.org/ubuntu/tarballs/ [03:25]