=== matt_c is now known as screaming_kid === screaming_kid is now known as matt_c [09:46] persia, good $TIME [09:50] Just settling in. You have a pastebin? [09:53] persia, a pastebin? [09:54] * NCommander is regoing through every package, and the only ones left so far are cario-clock (in progress), and fbreader which we talked about earlier [09:54] evience and bluez-gnome have hildon modifications, but when they were merged in Intrepid, people dropped their patches :-/ [09:57] Not precisely. I'm not sure about evince, but the patch for bluez-gnome was cleaned up, made non-arch-specific, tested for MID, UMPC, and Desktop (on a 30" screen even), and sent upstream for review. [09:58] Right I saw that [09:59] * NCommander was expecting a lot more packages in mid with lpia tweaks ... [09:59] So how can I help you? [09:59] I thought there were bunches, but you think there aren't? [09:59] I opened the mid seed [09:59] And I simply downloaded everything that wasn't a library [10:00] heh. That's the first part of my script :) [10:00] Ah, so there is more [10:00] But at least this is enough to get ubuntu-mid usable as a baseline on non-lpia [10:00] * NCommander sees if his packages have left NEW yet [10:01] nope [10:01] Well, my script is to pull the non-libraries from the gutsy, hardy, and intrepid seeds (not yet updated for jaunty), unpack, and grep lpia in all occurrances of debian/rules. [10:02] oooh handy [10:02] That's a great way to find lpia [10:02] They will probably be flushed on Monday. You still want me to download and process the script tonight? [10:02] But sometimes the LPIA modifications are just in control [10:02] Well, my script isn't perfect, but also running find against debian/control with the relevant exec isn't terribly complicated :) [10:03] Its up to you, I know which ones I have already touched, so its just a matter of doing it [10:03] Probably get some false positives for arch-specific packages, but those are quick to review. [10:03] I don't think lpia has any arch specific packages [10:03] aside from its kernel [10:03] OK. I'll run it and we'll see. Since it's source-based, and you've already fixed a bunch of sources, those oughtn't show up. [10:03] Oh, very cool [10:04] xserver-xorg.input-synaptics is an example of an arch-specific package built on lpia. [10:04] So that one we just need to change to be not arch-all, right? [10:04] er arch lpia [10:04] No. We want it on lpia. [10:04] That's one of the false positives I referenced when talking about adding debian/control checking to the script. [10:05] Otherwise my touchpad doesn't work, which would annoy me. [10:05] Remember that arch-specific packages can cover multiple architectures. [10:06] For instance, wine is i386, amd64, and lpia. It doesn't work on other arches (and probably can't easily work) [10:06] Well, presumably, it's possible to branch wine to work on alpha, ia64, and powerpc, if people have windows binaries for those architectures, but it's probably not worth it. [10:07] ACtually, I use wine to cross-compile Windows binaries to PowerPC [10:07] :-) [10:07] Well, yes, but you're clearly mad, so it's not relevant for the general use case :) [10:09] Me and my multiple personalities disagree withy ou! [10:09] All of them? No internal dissent on this issue? Everyone agree's you're perfectly balanced? [10:12] Yes :-) [10:12] Well [10:12] Well, I'm glad some good could come of this then :) [10:12] configure thinks my build environment isn't sane :-P [10:12] I dunno if I should laugh or cry about that [10:13] laughing is preferred. Pavlovian reinforcement and all that. [10:17] persia, cario-clock split out [10:18] * NCommander finds it funny/sad every hildon package is CDBS based [10:18] Actually, most of them aren't. [10:18] They use CDBS for simple-patchsys, but don't include debhelper.mk. [10:19] It's a convention adopted for historical reasons within MID, now mostly irrelevant (but not worth adjusting) [10:22] persia, I mean most the applications that use hildon are CDBS [10:22] Every once I've split out was CDBS based package [10:23] Odd. Most of the ones I remember looking at used debhelper 5 and CDBS's simple-patchsys. [10:23] Then again, I like CDBS, so maybe I just didn't notice much. [10:25] persia, strange. cario-clock has a hildon patch, but the resulting binaries don't end up linked against hildon O__o; [10:26] I suspect it's not actually a hildon patch. [10:26] The term "hildon patch" and the use of "--enable-hildon" ended up getting used for a fair number of non-hildon-specific things. [10:26] "--enable-hildon" if lpia [10:27] But the resulting binaries were identical .... [10:27] Yeah. That doesn't mean what you think it means. [10:27] For the same historical reasons previously mentioned "--enable-hildon" was generally used for making things MID friendly, even when it wasn't actually using hildon. [10:28] fair enough [10:28] mcasadevall@blacksteel:~/src/lpia/cairo-clock-0.3.3/debian$ diff cairo-clock cairo-clock-hildon [10:28] Common subdirectories: cairo-clock/DEBIAN and cairo-clock-hildon/DEBIAN [10:28] Common subdirectories: cairo-clock/usr and cairo-clock-hildon/usr [10:28] O_o; [10:28] (I know the cairo-clock-hildon folder is properly getting build w/ --enable-hildon) [10:29] any ideas? [10:32] I probably broke it when I tried to merge in the collection of outstanding changes I found laying about. [10:33] Have you checked what --enable-hildon actually does? [10:36] It turns on -DHILDON [10:37] scratch that [10:37] And what does that do? [10:37] It wasn't getting turned on [10:37] Hrm [10:37] * NCommander is looking at the autotools magic [10:38] Ok [10:38] The automagic is broken [10:38] :-P [10:38] (You are in a maze of twisty passages, all alike) [10:40] xyzzy [10:42] * NCommander works out the necessary voodoo to fix [10:56] persia, woo, fixed [10:57] Thanks. That's my mistake, and as you can see from the changelog, not exactly recent. [10:57] Its fine [10:57] It was easily to overlook, if I didn't actually use ldd to check the dependencies, I would have missed it myself [10:58] All I did was run the result in MID. Since it looked OK, I went with it. [11:00] * NCommander kicks his MID image [11:00] We need jaunty based ones so I don't need to recompile on intrepid [11:02] Well, we're getting close to Alpha 1, which is when the first test images start, usually. [11:02] If you want you could try debian-cd, but I'm going to wait to see what happens, personally. [11:02] We need to get an updated ubuntu-mid metapackage [11:02] Or the image will be miserably broken [11:03] persia, I can't test cairo clock, its complaining I don't have an accelerated envionment :- [11:04] Yeah, well. I used to have one of those. [11:04] I'm just going to confirm everything links together and upload and hope for the best [11:04] The drivers died early in the intrepid cycle, and my hardware requires binary blobs, for which the vendor has yet to produce a working driver for Intrepid. [11:05] Could not load "/usr/share/cairo-clock/glade/cairo-clock-ume.glade"! [11:05] Maybe cairo-clock-hildon should depend on whatever in MID provides that [11:05] That's not a video driver issue. [11:05] No, regular cairo clock doesn't work [11:06] "This clock will not function properly without compiz" [11:06] Anything under /usr/share/cairo-clock/ should be provided by cairo-clock. You might have to change a pathname? [11:07] * NCommander takes a closer look [11:07] Oh, I see [11:07] cairo-clock-ume.glade is provided, but isn't installed :-P [11:07] (no one caught that because the hildon patch didn't actually apply [11:13] NCommander: I've got a (painful) patch for contacts, so no touching. [11:13] StevenK, too late [11:13] Sigh [11:13] StevenK, what does your patch do? [11:14] Adds a contacts-hildon package [11:14] it wasn't that painful ... [11:14] Sure it was, it involved CDBS [11:14] No ... [11:14] Let me see the rules you came up w/ and what I came up with [11:15] StevenK, if you want to can clear the package in NEW, and then accept the merge for the seed changes :-) [11:16] http://paste.ubuntu.com/72820/ [11:16] http://pastebin.ca/1257568 [11:16] I think our patches are nearly identical [11:17] Do you add -DMAEMO_CHANGES into the autoconfage? [11:17] * persia is suddenly glad that perl is not the default language for debian/rules [11:17] If you don't, then it isn't hildonized [11:17] StevenK, I added it via make [11:17] $(MAKE) -C $(CURDIR)/debian/build-hildon CFLAGS='-g -O2 -g -Wall -O2 -DMAEMO_CHANGES' [11:17] * NCommander saved himself some trouble [11:18] I tested it on a MID image to confirm my changes work and checked w/ ldd [11:19] Personally, I prefer doing it via autoconf [11:20] It was a two-line patch to something we're patching anyway [11:20] StevenK, if you want to clobber my upload, go ahead [11:22] That involves caring. And I finished this work like a week ago, but just never had time to grab persia or someone else and say "Is this crack?" [11:23] It's pre-DIF. We like crack. [11:24] Well, decent quality crack at least. A bad cut is bad no matter how you look at it. [11:25] StevenK, sorry about that. I didn't know :-/ [11:26] Anyway, which is the better patch? [11:26] If the better patch is StevenK's, it ought be pushed anyway. [11:26] Mine didn't require the recreation of an autotools fudge [11:26] History is just a changelog entry. [11:26] And autotools patches are a headache either way [11:27] Isn't there already an autoconf patch? [11:27] persia, it patches configure directly [11:27] Not extactly the best patch ever created [11:27] Oh [11:27] scratch that [11:28] Its just one giant patch with hildon, autotools, and autoreconf mixed together [11:28] * NCommander blecks [11:30] * persia restarts the source pull script, realising that the apt-cache is kinda out-of-date [11:30] persia, as soon as I finish testing cairo clock, I'll upload [11:30] * NCommander is just waiting on pbuilder ATM [11:35] Is it by design we include OOo on mid images? [11:35] * persia enjoys the power of &. Changes download speed from the archive by a factor of about 10. [11:36] NCommander, It's a side effect of how treb works. If you know a way to provide an OOo reader without OOo, it would be interesting. [11:36] Ah [11:36] We need something the equivalent of of Word Reader for Windows [11:37] Well, sorta, but yes. [11:37] treb has roughly the right interface, but it's not exactly implemented in a space-efficient way. [11:37] It would require going into treb's source and figuring out what fat to remove [11:38] Ewww :-P [11:38] It's just debian/control :) [11:38] OOo 2.x is a giant blob of fat [11:38] It would actually require modifications to the beast [11:38] Yep. [11:38] OK. Local mirror of seeded packages updated. Time for find. [11:38] we might be able to just steal mojo's right away [11:39] * NCommander thinks we should relax treb to a recommendation :-P [11:39] Just so I can install it on something that doesn't have a gigabyte free [11:39] oh, echan :) [11:39] persia, ok, cairo clock works on a MID image [11:39] (woo!) [11:39] define "works" [11:39] does the munite hand actually move ? [11:39] The preferences panel is properly hildonized [11:40] * ogra has never seen cairo clock actually work [11:40] YOu need acceleration for it to actually work it seems [11:40] well, i have accelleration on the Q1 [11:40] but that doesnt make the hands move :) [11:40] And it still doesn't work? Hrm. [11:40] It works on my desktop [11:41] right, only talking abut mobile [11:41] ogra, the hildon patch wasn't applied [11:41] Due to a bug in the autotools script [11:41] hmm, bt it starts fine on the hildon desktop [11:41] i would expect it to not even start [11:41] Well [11:41] It appears [11:42] and the hands move! [11:42] what it does for me is that it drwas the clock fine, and you see actual movement of the hand [11:42] (I turned on the second and and its moving, so it the minute) [11:42] Ok, it works! [11:42] woo [11:42] but it never manages to actually move really ... it just jumps back and forth by a pixel [11:43] please leave it running for some minutes [11:43] * NCommander watches the second and minute hands move [11:43] for me it perfectly emulates an alarmclock with low battery :) [11:44] oh, you have a hand for the seconds ? ok, i never got that [11:44] so far so good [11:44] ogra, oh, I turned that on in the preferences [11:44] ah [11:44] (the hildon patch fixes the preferences) [11:44] right, i never got preferences [11:44] so it was actually the missing patch [11:44] Right [11:44] Well, no [11:44] The last uploader broke the autotools rules [11:45] I only caught that because I was checking with ldd to make sure the split package was linking against hildon [11:45] so the -DUSE_HILDON rule never kicked in [11:46] ogra, can I upload now :-) [11:46] sure [11:46] So I've run the script. http://paste.ubuntu.com/72833/ shows packages that match "lpia" in debian/rules, and are likely candidates for investigation from my current sources. [11:46] dont bothe me :) [11:46] *bother [11:46] its the time of cycle where everyone should just upload ... we can fix stuff later :) [11:46] I'll investigate the control ones, and try to weed out false positives. [11:46] I think your script is dated, I fixed galculator [11:47] unless its checking in the changelog [11:47] cheese and contacts also ahve been fixed [11:47] My apt-cache downloaded galculator-1.3.1 [11:47] * ogra finds it really unfortunate that galculator needs that .desktop patch [11:47] And I just updated it. Maybe something's out of date? [11:47] such a waste of bytes [11:47] I just pulled it from the jaunty source [11:47] its up to date [11:48] evince doesn't have a working hildon patch ATM [11:48] I'd like someone to review fbreader with me [11:48] evince should get even more than a hidlon patch [11:48] I'm hestiant at touching either grub or usplash :-P [11:48] there is a grab and drag patch *somewhere* that should be added [11:49] so it behaves like midbrowser [11:49] ogra, evience doesn't have a hildon patch since Intrepid Alpha 3 [11:49] * NCommander knows that first hand since I disabled it :-P [11:49] yeah, because we stopped using it i think [11:49] ogra, no, because GTK+hildon+evience broke stuff ;-) [11:50] *if* we use it, the g&d one should be added as well [11:50] * NCommander remembers when he did that merge with seb128 [11:50] I personally would like to see Mozilla be removed with something less fat [11:50] * NCommander nitpicks :-P [11:51] webkit isnt so much smaller [11:51] what would you use ? dillo ? [11:51] Midori is a fairly small webbrowser (its webkit based, but its fairly cruft free) [11:51] Not sure if its fully production ready yet, but 1.0 has been quite stable [11:52] but doesnt get as much developer attention in ubuntu as xul does [11:52] But its not XUL :-) [11:52] I'm not sure why it's out of date. Here's the rough and dirty script if you've the bandwidth to replicate: http://paste.ubuntu.com/72835/ [11:52] we have a huge xulrunner tea we can poke for fixes, that somewhat justifies using xulrunner [11:52] *team [11:53] if we'd use midori we were on our own [11:53] persia, your not checking against jaunty? [11:53] ogra, I'm used to embedded environments where 16MB of space was considered a luxery :-P [11:54] heh [11:54] NCommander, RIght. See this is why this is a bad time to run the script :) [11:54] i think that time is over nowadays [11:54] Just ignore anything you know to be fixed. [11:54] though i still have two old ipaqs in a box soemwhere :) [11:54] i know what you mean [11:54] I'm disturbed grub has an lpia specific tweak [11:55] Actually, that's just a comment. [11:55] \o/Q [11:55] *! [11:55] It's not an especially smart script :) [11:55] Just something to check stuff, and build quick lists for processing. [11:56] ogra, can you merge my seed changes into the actual seed? [11:56] NCommander, Isn't it worth building the complete list first? [11:56] If alpha 1 is around the corner, we'll have a really broken image if the seed is left the way it is [11:57] Alpha 1 is the 20th, I think. Still a couple days. [11:57] yeah [11:57] Fair enough [11:57] i'll do seed work tomorrow anyway [11:57] ANyway, who would like to look over fbreader [11:57] * persia wonders if StevenK is still awake and might be willing [11:58] * ogra doesnt want to do "normal" work on a sunday :) [11:58] persia, I'm just not sure how to make the seed changes so the end result si desirable [11:58] * NCommander doesn't want to blow something up [11:58] NCommander, Well, if you're changing something, swap the new package name for the old package name. [11:58] No, I need to add a new dependency [11:58] You oughtn't need a deep understanding of seeds for this stuff. [11:59] But I'm not quite sure the dependency tree will work out [11:59] Oh, I remember now. [11:59] fbreader depends on libzgui which depends on a backend [11:59] The LPIA hack forces the dependency ont he backend on mameo [11:59] and germinate wil take care for you :) [11:59] as long as the package has the dep right [12:00] I'd recommend downloading mobile-meta, modifying update.cfg to point at your branch, and seeing what results you get. [12:00] Fair enough [12:00] * NCommander needs to get his day started [12:00] Obviously that change doesn't belong in the repo, but it lets you test sample results of seeds. [12:04] persia, right [12:04] persia, we'll probably have everything de-lpiaified by today or tommorow \o/! [12:06] Right. All the control.candidates that are not also rules.candidates are false-positives. [12:06] cool [12:06] (p.s. Don't look at gtreamer plugins as an example of how to handle architecture easily) [14:17] NCommander, I reran the script in an updated Jaunty environment. The rules list is http://paste.ubuntu.com/72885/ [16:34] hello can i install ubuntu-mid-edition on a samsung omnia? [16:43] can anything running windows mobile run ubuntu mid edition? === playya_ is now known as playya [17:06] no one able to answer the question? could ubuntu mid edition be installed on a samsung omnia? [18:00] s0u][ight, currently the MID edition is targeted to the Intel Atom CPU with the poulsbo video subsystem. It likely will be lacking drivers for many devices as it was directly targeted to the Intel Crown Beach development platform. [18:01] davidm, it probably will never work on the samsung omnia right? [18:01] what is the CPU on that device? [18:01] i don't know [18:01] lemme look it up [18:03] :s lag in my network [18:03] There is an image of the MID that runs on the samsung Q1 Ultra but it lacks some drivers since the Q1 did not use the Atom chip, but it was the only platform we had that was portable for testing somethings on. [18:04] davidm, i think the q1 and omnia are totally different [18:04] since the price of the one is more expensive [18:04] can't even open google ;| [18:05] -_-' (love that emoticon) [18:08] davidm, irc works fine but i can't surf :| lol [19:28] Please bear with me on this question. Are you guys using Hildon? I assume so. What IM client does Ubuntu mobile ship by default? I'm an upstream developer on Pidgin and there's a port of Pidgin for the N810 tablet that uses Hildon. Is that something Ubuntu Mobile would be interested in? If so, would it help to merge the pidgin-maemo code upstream?