[00:10] Now I get the error "E: Couldn't find package telepathy-mission-control" [00:44] StevenK, HappyCamp: This method is not working. so I will create a crownbeach-full image and then download the PPA deb package and install it on the target. I hope it should work. [01:06] Jayc_: If you do so, you'll need to mount /boot manually before you install it. [01:08] well, The initial image doesn't boot to stat with :-( === davidm_ is now known as davidm === cprov-out is now known as cprov [11:12] asac: /usr/bin/ld: warning: libsqlite3.so, needed by /usr/lib/xulrunner-devel-1.9b1/lib/libxul.so, not found (try using -rpath or -rpath-link) [11:12] /usr/bin/ld: warning: libmozjs.so, needed by /usr/lib/xulrunner-devel-1.9b1/lib/libxul.so, not found (try using -rpath or -rpath-link) [11:12] asac: This is a bit weird [11:12] asac: And this seems worse: /usr/lib/xulrunner-devel-1.9b1/lib/libxul.so: undefined reference to `JS_SetContextPrivate' [11:13] But then this might be tinymail's fault [11:13] But I don't see those [11:17] I think I saw something related as the second error in some Debian package a while ago [11:17] Somehow, upstream doesn't require js stuff when you use xul, I think it was in totem [11:18] Ah, xulrunner has its own libsqlite3 [11:19] asac: Hmm why does /usr/lib/xulrunner-devel-1.9b1/lib/libxul.so link to libsqlite3.so but doesn't have any rpath to the sqlite3.so in the package? [11:22] Is it possible to run ubuntu mobile on Arm (ie. is there images compiled for arm?) [11:22] No [11:23] lool: what are you doing? [11:24] lool: what are you trying to port? [11:24] asac: tinymail [11:24] he? ... StevenK already built that :) [11:24] asac: His version doesn't build here; didn't build for Mithrandir either, but I don't know why [11:24] lool: lets see if i still have the patch [11:25] asac: One thing wrong with the patch is to test for xulrunner last in the list; this prevents building on e.g. my environment where I have firefox [11:25] -dev [11:25] asac: I have an updated patch whic fails at this place if you like [11:25] lool: cant you say: --with-gecko=xulrunner or something? [11:25] please show me [11:26] asac: There's no such arg; but I planned adding one [11:26] asac: In fact I made the macros checking the module parametrizable, but didn't implement ac_arg_with yet [11:26] Ok, cool. is it possible to compile it for Arm11 ? [11:26] jussi01: It should be [11:27] But we don't build Ubuntu for arm at all right now [11:27] asac: http://people.ubuntu.com/~lool/xul-fixes.dpatch is what I use ATM [11:27] asac: But the "libxul-1.9 mozilla-js-1.9" in configure.ac I plan to drop as it didn't help [11:27] mjg59: great. Is it something ubuntu mobile would like to pursue in the future? [11:28] lool: http://people.ubuntu.com/~asac/tinymail.xul19.patch [11:28] s/libxul-embedding/libxul-embedding-1.9/ [11:28] that should work [11:28] jussi01: I don't think we're looking at it right now [11:28] jussi01: It'll depend on demand [11:29] mjg59: Ah, I understand. Thanks for your time. :) [11:29] lool: for me it builds against the trunnk tinymail [11:30] asac: Perhaps trunk builds [11:30] StevenK told me that you need trunk [11:30] ... but 0.0.5 builds for me.... [11:30] because of some features that modest needs [11:30] I took one patch from SVN [11:33] asac: Do you have libsqlite3-dev installed? [11:33] ii libsqlite3-dev 3.4.2-1build1 SQLite 3 development files [11:33] Might explain why you don't get the ld warnings as above [11:34] asac: Do you agree there should be a rpath? [11:34] no [11:34] thats the idea [11:34] just use libxul-embedding and the GLUE [11:34] asac: So should we drop the libsqlite3.so from xul then? [11:34] thats what you see in code [11:34] no not either [11:34] or wait [11:34] sorry [11:34] I mean a rpath on sqlite, not on xul [11:37] lool: please try just the patch i gave you and give me the error you get then [11:37] asac: I'll have to change it to pull xul first [11:37] you can also uninstall firefox-dev for a minute ;) === dholbach_ is now known as dholbach [11:43] asac: aha [11:44] asac: It built at first, but it stopped doing so when configured like the deb package --with-platform=maemo --enable-demoui --enable-gnome --enable-tp --with-html-component=mozembed [11:44] asac: configure: error: "You need Mozilla's XPCOM" [11:44] interesting [11:44] config.log? [11:44] Well it's broekn configure:26275: checking Mozilla XPCOM pkg-config module name [11:45] for me it builds with: sh autogen.sh --with-html-component=mozembed [11:45] (no module name printed) [11:45] lool: yes you have to use libxul-embedding-1.9 [11:45] the patch has just libxul-embedding [11:45] (as i told you in the beginning ... sorry should have repeated) [11:45] asac: Ah my mistake, sorry [11:46] * asac building without libsqlite3-dev [11:46] Right, configure passes with -1.9 [11:47] works as well [11:47] asac: Indeed, it built fine [11:47] cool ... now lets go for modest [11:48] whats the point of having the zoom method in there? [11:48] i think that logic should be moved to tinymail [11:48] can you do that? [11:48] that would simplify things a lot [11:53] asac: Aha, for some reason I used "mozilla-nspr-1.9" instead of "nspr"; why shouldn't I use that? [11:54] lool: in basic embedding cases you don't need nspr anymore [11:54] try to drop that [11:54] asac: Upstream is using nspr [11:54] yeah ... its for 1.8 [11:55] if you link against glue only and it builds without -DXPCOM_GLUE_USE_NSPR ... then you won't need it anymore [11:55] asac: I'm not sure I want to patch upstream to stop pulling nspr; the fact is they are right now, and I wonder why we have multiple versions and why one should be preferred and not the other (I hope I'm not too annoying in asking you to teach me about this :) [11:57] asac: Also, why did you use libxul-embedding-1.9 for mozilla_xpcom_pcs instead of libxul-1.9 ?! [11:59] lool: libxul-embedding is for standalone things ... libxul-1.9 is for dependent things (e.g. things that get loaded in an already running xpcom/xul environment (like plugins/native extensions) [11:59] lool: http://developer.mozilla.org/en/docs/XPCOM_Glue [11:59] asac: Ok, I got my version to build now by changing to the libnss nss the libnspr nspr and using -embedding everywhere [12:00] asac: The upstream m4 macros used mozilla_xpcom_pcs to lookup xpcom (I think) and mozilla_gtkmozembed_pcs to lookup gtkmozembed; why not use libxul for xpcom? Isn't xpcom the same in both? [12:01] asac: Ok, I think I got it [12:02] asac: What was previously only firefox-xpcom now is libxul-embedding-1.9 or libxul-1.9 depending on whether you're starting xul or whether you're started from xul; correct? [12:02] right [12:03] asac: Still, shouldn't it build would I use the libxul-1.9's xpcom? [12:03] lool: point is that they enforce frozen linkage now ... so directly linking doesn't give you any benefit over glue [12:03] lool: no ... because you lack symbols ... like you saw above [12:04] Ok [12:04] asac: And what's the story with nss? [12:04] and nspr [12:04] after all the libs are shipped in pkglibdir ... which means ... don't like against us :) [12:04] Are we moving to using these separate libs? [12:04] s/like/link/ [12:05] It would sound "logical" to me that if you try to link to xulrunner-1.9's gtkmozembed stuff, you would do good to try to link to ist nss and nspr at the same time [12:05] lool: apps that don't use xpcom can stay the same ... they use the system nspr/nss as usual [12:05] the other apps should drop the linkage as its not needed and more portable. for those few cases where you need to use the glue with -DXPCOM_GLUE_USE_NSPR we need to link against system-nspr/nss [12:06] lool: well ... if you don't use it you shouldn't do that because the glue loads the right nspr for you at runtime [12:06] asac: Concretely, I wonder whether I should be adding b-deps on libnss-dev and libnspr-dev as to explicitely reference these which are used in m4/moz.m4 [12:07] asac: Ah ok, nss and nspr are pulled over xpcom [12:07] lool: they shouldn't be used in xul 1.9 anymore ... just for 1.8 :) ... otherwise xulrunner-1.9-dev should give you the dependencies i guess? [12:07] asac: Yeah, but as I said hte macros are explicitely pulling these deps [12:08] yes ... the macro needs to be fixed then. [12:08] and my patch already touches it anyways [12:09] asac: Did you discuss the xulrunner 1.9 issues with upstream? [12:09] which issues? [12:09] I mean merging the changes you did and not checking for nss and nspr when xulrunner 1.9 is used? [12:09] mozilla upstream? [12:10] No, tinymail [12:10] Sorry, I wasn't clear :) [12:11] why would i? I first heard of tinymail yesterday :) ... my idea is to develop the patch here and send it upstream. They will need it if they want to support xulrunner 1.9 anyways. [12:12] asac: Oh ok; I have no idea of the date the patch was written at; I was just wondering what they though of this, whether they would take it etc. [12:13] yeah ... please send it upstream :) [12:13] if they have a bts you can cc me [12:13] asac: The reason I ask is because I'd like to know whether $mozilla_xpcom_includedir/unstable makes sense for older xulrunner versions as well [12:13] no it doesn't ... the diff should test if its there and otherwise don't use it [12:13] Since how long do they have /unstable and /stable? [12:14] since 1.9 [12:14] before it was different [12:14] as everything :) [12:14] haha [12:15] asac: The numbering in the .pc files sucks I think, is this temporary or is upstream also shipping with this -$major.$minor.pc naming scheme? [12:16] thats we only atm ... we will drop it with next upload. [12:18] asac: So the libxul-embedding-1.9 stuff shouldn't be sent upstream, rather it should be libxul-embedding I guess or the AC_ARG_WITH [12:18] yes [12:19] but please try to remove the nspr/nss linkage if you have 1.9 [12:20] asac: Is it ok to do that if MOZILLA_VERSION_U >= 1.9? [12:21] lool: for current xul its: #define MOZILLA_VERSION_U 1.9b1 [12:21] no idea if moz.m4 cuts of after 9 [12:24] asac: I'll tell you: #define MOZEMBED_MOZILLA_VERSION "$mozilla_version" [12:24] :-P [12:25] 1.9b1 gives "1 9 1" in m4/moz.m4's logic, bad === jussio1 is now known as jussi01 [12:33] lool: why we want >= 1.9 ... isn't that true then? [12:33] asac: Yeah, it's true, but it's true for a broken reason [12:33] ;) [12:33] seems like two wrongs do make a right :) [12:34] asac: So thanks for all your help and explanations [12:34] lool: please do modest as well :) [12:35] ... would be great if we could get that working asap [12:35] asac: Do you think we could have libxul*.pc set includedir to /usr/include/xulrunner-1.9b1/{includetype}? This would allow querying the proper dir [12:36] asac: tinymail + modest are StevenK's babies; you wouldn't be to take away a baby, would you? I just chimed in because it wouldn't build for me (due to firefox-dev) which made me go through the whole patch again [12:37] Now I guess I'll pursue until I have something to send upstream, since I learnt about it all :) [12:38] hmm ... i will think about the includedir variable thing. thanks for the suggestion [12:39] lool: Does that mean modest builds too? [12:39] StevenK: we didn't try that [12:39] * StevenK has been distracted [12:40] i thought you ran into issues yesterday [12:40] I did, I fixed those and ran into more :-) [12:41] hehe [12:43] StevenK: I guess it will build better soon, but then you poked Mithrandir on that already; I'm happy yo help if Mithrandir doesn't find out (unlikely) or ends up being too busy [12:46] lool: So, does tinymail require patching? [13:00] StevenK: At the very least to prevent usage of firefox-dev, yes [13:01] lool: Right, okay. Mind sending me the diff? [13:02] lool: And it seems Mithrandir is too busy, would you mind terribly poking at modest? [13:02] StevenK: Of course I will; if it builds for you that perhaps can wait until I cleanup some stuff as to render the patch acceptabel upstream? [13:02] StevenK: Ok [13:03] lool: Yeah, I was pondering sending Philip the patch, but only when I verified it could build against both === jussi01 is now known as jussio1 === dholbach_ is now known as dholbach === asac_ is now known as asac === cprov is now known as cprov-lunch === cprov-lunch is now known as cprov [15:57] asac, you there? [15:57] j [15:57] y [15:58] asac, 'Create new network' does not work on moblin environment. Driver times out , [15:59] you see in log that it calls the supplicant? [15:59] asac, yes [16:01] raji: ok same game then ... reproduce with supplicant alone and see which error you get [16:03] asac, do I need to give any different command to wpasupplicant, for telling it to create a new network? or is it just like testing connectivity? [16:05] davidm, bfiller: will mobile support backup? upgrade spec says backup occurs before upgrade. finalizing prd now [16:11] asac, Is it same as testing connectivity [16:34] raji: you can also hack network-manager to start wpasupplicant with -dd [17:14] Mithrandir: is there a problem with the nightly UME builds? Looks like none have updated on cdimage.ubuntu.com/moblin since 11-26 [17:25] kyleN: 3.4.17 there is a new sound control panel, I'll email you the screen shot (it's in yesterdays image) [17:26] bfiller: wrong channel - this one's public [17:26] ok [17:26] smagoun: right [17:54] asac, How do I hack? === cprov is now known as cprov-out === doko_ is now known as doko