[13:26] <asac> hi ... i am back from the dead
[13:31] <asac> nss hasn't been newed ... too bad
[13:34] <asac> Ubulette: what is in nss-util?
[14:05] <Ubulette> hi
[14:05] <Ubulette> asac, have a look into nss.head/debian/libnss3-1d.symbols
[14:12] <asac_> hmm ... I think we should send the patch to debian
[14:18] <Ubulette> btw, symbols continue to change so i'm not sure what good the so bump is so far
[14:18] <Ubulette> -what+how
[14:33] <asac> Ubulette: well ... we bump it now and respin when its final
[14:39] <asac> Ubulette: do the missing symbols cause issues?
[15:30] <asac> good news ... i might be potent enough to upload to debian again :)
[15:30] <asac>  - updating icedove accordingly :)
[16:36] <asac> !seen bluekuja
[16:36] <asac> hmm ubotu is gone :/
[16:43] <armin76> it's back
[16:43] <armin76> !seen bluekuja
[16:43] <ubotu> Sorry, I don't know anything about seen bluekuja - try searching on http://ubotu.ubuntu-nl.org/factoids.cgi
[16:44] <asac> hmm
[16:44] <asac> stupid bot :)
[16:44] <armin76> anyway, last time, 11:31 this morning, UTC
[16:44] <asac> ah ok
[16:44] <asac> tx
[16:44] <armin76> np
[16:44] <armin76> you test my cairo patch :P
[16:45] <asac> he?
[16:45] <asac> anything we do not yet have in hardy?
[16:45] <armin76> cairo-1.5.4 sigbus on sparc
[16:46] <armin76> cairo included in mozilla trunk >29 nov, includes broken cairo
[16:46] <asac> regression?
[16:47] <armin76> hrm...not sure
[16:47] <armin76> it was fixed yesterday on cairo's tree
[16:47] <armin76> but not on mozilla's
[16:47] <armin76> mozilla bug 409976
[16:47] <ubotu> Mozilla bug 409976 in GFX: Thebes "Cairo in trunk gives SIGBUS on SPARC" [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=409976
[16:48] <armin76> http://rafb.net/p/Z0YZI825.html
[16:48] <armin76> so i made a patch for mozilla
[16:49] <asac> alignment issue?
[16:49] <armin76> yup
[16:49] <asac> (sparc64) ?
[16:49] <asac> ok
[16:49] <armin76> yup
[16:50] <asac> did you include the new patch for align from david miller on 2.x branch yet?
[16:50] <asac> mozilla bug 161826
[16:50] <ubotu> Mozilla bug 161826 in Layout: Fonts and Text "nsTextFrame::MeasureText()'s fast text measuring codepath crashes on RISC machines" [Critical,New] http://bugzilla.mozilla.org/show_bug.cgi?id=161826
[16:52] <armin76> nope
[16:52] <asac> he claims that with that patch the Set/GetUnichar business won't be needed anymore
[16:53] <asac> i couldn't test it, but just got a negative comment from someone that applying just the patch doesn't cure it (though it should)
[16:53] <asac> the patch needs to be adapted though as it depends on the previous one (which got review-)
[16:54] <asac> would be perfect if the patch would be done properly against the pristine 2.x codebase ...without the other one
[16:55] <armin76> oh
[16:55] <armin76> http://sources.gentoo.org/viewcvs.py/gentoo/src/patchsets/mozilla-firefox/2.0.0.11/003_firefox-bus-error.patch?view=markup
[16:57] <asac> armin76: yes ... the idea is that all that isn't needed anymore once we have the proper Miller patch :)
[16:58] <asac> hmm yours is different somehow
[16:58] <asac> oh you have an old one
[16:59] <asac> https://bugzilla.mozilla.org/attachment.cgi?id=222992
[16:59] <armin76> yah, looks like it
[16:59] <asac> its the " Complete fix for alignment problems   "
[16:59] <asac> we have " Slightly modified fix"
[17:00] <armin76> does that affect cairo anyway?
[17:00] <asac> nope ... just another alignment issue
[17:00] <asac> for 2.x
[17:01] <asac> i hoped you might untangle the patch and test as you appear to be able to test on sparc :)
[17:01] <asac> if not ... don't bother, i will just punch the latest on top of the other cruft :)
[17:04] <armin76> Linux bender 2.6.23-gentoo-r3 #1 SMP Fri Dec 7 17:10:47 UTC 2007 sparc64 sun4v UltraSparc T1 (Niagara) GNU/Linux
[17:04] <armin76> looks like it :P
[17:04] <armin76> so what do i have to test exactly?
[17:04] <armin76> just the last patch?
[17:06] <asac> armin76: replace the one you have with just the last one
[17:06] <armin76> okay
[17:06] <asac> you need to unthrottle the bits because some hunk already has the code lines changed by the previous most likely
[17:06] <armin76> on 2.0 branch, right?
[17:07] <asac> yes
[17:07] <asac> 3.0 should be fixed without patch
[17:07] <armin76> yeah
[17:07] <armin76>  Mozilla/5.0 (X11; U; Linux sparc64; en-US; rv:1.9b2) Gecko/2007122719 Firefox/3.0b2
[18:08] <CheGuevara> Ubulette: new nss fails
[18:08] <CheGuevara> dpkg: error processing /var/cache/apt/archives/libnss3-1d_3.12.0~cvs20071221t1057-0ubuntu1~fta1_i386.deb (--unpack):
[18:08] <CheGuevara>  trying to overwrite `/usr/lib/nss/libfreebl3.chk', which is also in package libnss3-0d
[18:14] <Ubulette> hmm
[18:15] <Ubulette> ok, i'll have a look, it's from the merge i did with debian..
[18:15] <Ubulette> damn
[18:16] <Ubulette> oh, .1d is installed before the old .0d is updated
[18:18] <Ubulette> CheGuevara, did you cherry picked it or was it a regular upgrade ?
[18:19] <asac> missing Replaces?
[18:21] <armin76> asac: http://rafb.net/p/G313Cc48.html
[18:21] <armin76> that should be the patch that applies
[18:22] <asac> does it fix the crash?
[18:22] <asac> armin76: ?
[18:22] <armin76> didn't start to compile yet :P
[18:22] <armin76> unpacking...
[18:22] <asac> we still need SetUnicode?
[18:22] <asac> aeh Unichar
[18:23] <armin76> nfc :P
[18:23] <Ubulette> asac, http://paste.ubuntu.com/3078/
[18:23] <asac> e.g SetUnichar was only implemented in the previous patch :)
[18:24] <asac> Ubulette: is that right?
[18:24] <asac> anyway  ... a Replaces: should fix the install / remove order
[18:24] <armin76> compiling, bbl
[18:24] <Ubulette> asac, yes. there's no replace in 0d, not even in your nss-3.12.0~1.9b2+nobinonly
[18:25] <asac> Ubulette: strange somehow ... ffox 2 still works here
[18:25] <Ubulette> it works for me too
[18:25] <asac> yeah ;)
[18:26] <asac> but add a versioned Replaces to 1d so its not unpacked while -0d is still installed
[18:26] <asac> -0d << previous
[18:26] <Ubulette> 0d is not empty
[18:26] <asac> yes, thus a versioned Replaces:
[18:27] <Ubulette> is contains links for .so.0d -> .so.1d
[18:27] <Ubulette> it
[18:27] <asac> yes ... i know
[18:28] <asac> Replaces: doesn't remove -0d if its properly versioned ... it just tells dpkg to remove old 0d first before unpacking -1d
[18:28] <asac> (at least i hope:))
[18:31] <Ubulette> oh, i see the problem. it's the f*g  3.12.0~1.9b1 vs 3.12.0~cvs
[18:33] <asac> hehe
[18:34] <asac> ~~cvs
[18:38] <[reed]> how often does firefox-trunk get updated?
[18:40] <asac> [reed]: its a dummy package for firefox-3.0
[18:40] <[reed]> ah
[18:40] <asac> so atm it gets updated on new RC releases that appear good
[18:41] <asac> rc2 is still waiting for new nss ... which ended up in some queue which isn't really processed over christmas :)
[18:41] <asac> aeh beta2 :)
[18:46] <Ubulette> [reed], I also daily build trunk
[18:47] <Ubulette> and I push on snapshot to my ppa about once a week
[18:49] <Ubulette> so far, my ppa is stuck to b2 until 1/ hardy has b2 and 2/ b3pre works again for miro
[18:51] <asac> is trunk completely broken for gtkmozembed atm?
[18:51] <asac> Ubulette: does epiphany start with your b3pre package?
[18:52] <asac> well it won't unless you don't use system-ns{s,pr}
[18:53] <asac> but a respin should work if its not completely broken :)
[18:54] <[reed]> Ubulette: is there a bug filed on miro?
[18:55] <[reed]> have you requested blocking?
[19:05] <Ubulette> [reed], it's just that because of the mass backout, some of our patches (python) no longer applies
[19:06] <[reed]> mass backout of gtkmozembed stuff?
[19:08] <CheGuevara> Ubulette: regular apt-get update && apt-get dist-upgrade
[19:09] <CheGuevara> (sorry was out shopping)
[19:10] <Ubulette> [reed], yes
[19:18] <Ubulette> asac, it's far too late for ~~cvs. we should not have let ~1.9b in.
[19:18] <asac> ... for next bX we should fix the dailies
[19:19] <asac> i think it should be ~1.9X~cvs :)
[19:19] <asac> for Xpre
[19:20] <Ubulette> it's too late. ~c > ~1
[19:20] <asac> he?
[19:20] <asac> what version do you have now?
[19:21] <asac> 3.12.0~1.9b2+nobinonly
[19:21] <asac> thats the version in archive (NEW)
[19:21] <Ubulette> 3.12.0~cvs20071221t1057-0ubuntu1~fta2
[19:22] <Ubulette> (binonly is gone)
[19:22] <asac> so yeah ... you should just switch the versioning and append to the archive version
[19:22] <asac> if you are ahead:
[19:22] <Ubulette> ?
[19:22] <asac>  3.12.0~1.9b3~cvs
[19:22] <asac> if you feel more like you are apost
[19:22] <asac> 3.12.0~1.9b2+cvs
[19:22] <asac> but since we consider our builds "pre-releases" i would go for ~cvs
[19:24] <asac> but i see the issue
[19:25] <Ubulette> i don't understand why debian used ~1.9 in the 1st place. nss has its own version. it's 3.12.0 alpha 2b
[19:25] <asac> point is it isn't tagged with an nss tag anymore
[19:25] <asac> they use the FIREFOX tag
[19:25] <Ubulette> yep
[19:26] <asac> so there is no clear versioning except that its 3.12.0~something
[19:26] <asac> mike choose the 1.9 from the firefox
[19:26] <asac> aeh xul
[19:27] <Ubulette> i'm stuck with 3.12.0~cvs unless i use an epoch
[19:28] <asac> plesae keep it below 3.12.0 (without epoch) ... whatever comes
[19:28] <asac> you could use 3.12.0~+cvs
[19:29] <asac> its bad as it won't fit into archive versioning, but at least we have a chance to join the right versioning when 3.12.0 is final
[19:30] <Ubulette> it's too late for 3.12.0, until final
[19:31] <asac> which archives have the problem? your ppa?
[19:31] <asac> mt ppa?
[19:31] <Ubulette> both
[19:32] <Ubulette> and my own repo but it doesn't matter
[19:34] <asac> i don't get why you run into this replaces issue because of this ... cvs > 1.9 ... so the upgrade should work out well
[19:35] <asac> i really think the above problem is due to a missing Replaces:
[19:35] <armin76> asac: nsTextFrame.cpp:5144: error: 'SetUnichar' was not declared in this scope
[19:35] <asac> so as long as we keep ~cvs in mt in sync with what is shipped in hardy ... and keep your nss at least ahead we shouldn't have any issues
[19:35] <asac> armin76: right ... read what i said above: SetUnichar is defined in the patch you dropped :)
[19:35] <armin76> what should be there instead?
[19:35] <armin76> oh
[19:35] <asac> which is why i thought that we should not use SetUnichar
[19:36] <asac> if we need it we have to add the functio n... but i thought a memcpy should work well too
[19:36] <asac> (because the code takes care for the right alignment)
[19:36] <Ubulette> asac, mike used a Conflicts: << ~1.9b1
[19:37] <Ubulette> not Replaces
[19:37] <asac> yes, but conflicts is not really the right thing to express that the other package has to be removed before its unpacked
[19:37] <asac> replaces is the right way to express that there is a file conflict
[19:38] <asac> conflicts == doesn't work together ... replaces: == ships same files
[19:38] <asac> but anyway ... in this definitly the versioning that bites us here ... right.
[19:44] <Ubulette> CheGuevara, i've just pushed a new nss to my ppa, it should be ready in ~1h
[19:46] <Ubulette> next, I'll push a new xul, to use the new nss/nspr so that "depends" are correct (useful for those cherry picking in PPAs)
[19:48] <Ubulette> asac, I've packaged the long due thunderbird 2.0.0.9 for hardy: http://paste.ubuntu.com/3079/
[19:49] <asac> * Use --disable-pedantic as glib 2.15.0 is not clean enough - update debian/rules
[19:49] <asac> please file a bug about that
[19:50] <asac> (against glib)
[19:59] <Ubulette> bug 179119
[19:59] <ubotu> Launchpad bug 179119 in glib2.0 "glib 2.15 not clean with -pedantic" [Undecided,New] https://launchpad.net/bugs/179119
[20:01] <asac> the thing that makes me shiver is that it fails nowhere but in a .cpp file developed by mh
[20:02] <asac> or did oyu see this error with xul/ffox trunk?
[20:02] <asac> they use the mime thing as well ... but of course without this patch
[20:02] <asac> but there should be other places in moz code where they include glib.h i guess
[20:02] <asac> strange
[20:03] <armin76> asac: i have nfc about C, so if you could please replace what should go there? :)
[20:03] <asac> he :)
[20:03] <asac> mozilla bug 161826
[20:03] <ubotu> Mozilla bug 161826 in Layout: Fonts and Text "nsTextFrame::MeasureText()'s fast text measuring codepath crashes on RISC machines" [Critical,New] http://bugzilla.mozilla.org/show_bug.cgi?id=161826
[20:04] <armin76> what?
[20:05] <asac> wait a second
[20:06] <asac> *cp2-- = PRUnichar(*cp1--);
[20:06] <asac> armin76: ^^ ?
[20:06] <armin76> okay, what was before, then
[20:07] <armin76> http://mxr.mozilla.org/mozilla/source/layout/generic/nsTextFrame.cpp <- why it doesn't exist?
[20:07] <asac> it doesn't exist because no patch has been landed at all
[20:07] <asac> it was review-
[20:07] <asac> which is why i would like to get it right :)
[20:08] <armin76> err
[20:08] <armin76> https://bugzilla.mozilla.org/attachment.cgi?id=286275
[20:08] <asac> from what i understand the new patch fixes the alignment before the SetUnichar is run
[20:08] <armin76> that's the only thing i need?
[20:09] <armin76> or i need the other patch as well?
[20:09] <asac> so it shouldn't be needed
[20:09] <asac> armin76: the idea is to just use that patch and don't use SetUnichar
[20:09] <asac> if that crashes ... we have to pull in the SetUnichar implementation from the previous patch
[20:10] <asac> the patch unfortunately was developed on top of the review- patc h... i asked mr. miller for clarification but he refused to give any answer saying "this all is monkey business" :)
[20:10] <armin76> okay, but why nsTextFrame.cpp doesn't exist, if its in the tarball?
[20:10] <asac> because it has been moved on trunk
[20:10] <asac> you need to use branc
[20:11] <armin76> oh
[20:11] <asac> http://mxr.mozilla.org/mozilla1.8/source/layout/generic/nsTextFrame.cpp
[20:12] <armin76> thanks
[20:19] <Ubulette> asac, what about mozclient, would you sponsor it or what ? revu is dead at the moment, or dying.. from inactivity or from boredom
[20:22] <asac> hehe
[20:22] <armin76> http://rafb.net/p/Qrhu9p52.html
[20:23] <asac> looks like what i would think
[20:23] <asac> if it crashes we should add the SetUnichar
[20:23] <asac> to the patch
[20:25] <armin76> applies fine, starting to compile, will take 30mins or so
[20:44] <armin76> http://blog.netscape.com/2007/12/28/end-of-support-for-netscape-web-browsers/
[20:52] <Ubulette> I remember when i've built the 1st release of netscape
[20:53] <Ubulette> i was using mosaic before
[20:57] <Ubulette> well, there was no other choice at that time, mosaic was the 1st, then came netscape
[21:04] <armin76> asac: Mozilla/5.0 (X11; U; Linux sparc64; en-US; rv:1.8.1.11) Gecko/20071228 Firefox/2.0.0.11
[21:06] <armin76> i'm going to build it without the patch
[21:08] <Ubulette> asac, I don't get it.. you're doing icedove 2.0.0.9 ? what about tb 2.0.0.9 then ?
[21:59] <asac> who says that i am not doing tbird?
[22:00] <asac> i just want to sort out some pressing debian things first
[22:15] <armin76> asac: rofl, without any patch it works anyway
[22:16] <armin76> oh no
[22:16] <armin76> it sigbus right now
[22:16] <armin76> took a while, though
[22:19] <armin76> lol
[22:19] <armin76> it just gives randoms sigbus :D
[22:20] <Ubulette> armin76, http://home.kairo.at/blog/2007-12/the_saga_continues
[22:35] <Ubulette> it's everywhere: http://mozillalinks.org/wp/2007/12/aol-announces-end-of-support-for-netscape/
[23:32] <armin76> asac: nah, same with the patch
[23:49] <Ubulette> asac, please review & sponsor tb
[23:52] <Ubulette> http://bazaar.launchpad.net/~mozillateam/thunderbird/thunderbird.dev and http://www.sofaraway.org/ubuntu/tarballs/thunderbird_2.0.0.9+nobinonly.orig.tar.gz
[23:52] <Ubulette> rev #51