[00:04] <JanC> mcphail: wait, I can run speccy games on my Ubuntu phone now?  :)
[00:06] <JanC> & antony_ ^^^
[00:09]  * JanC must have several tapes with speccy apps somewhere on his parent's attic...
[00:13] <mcphail> JanC: don't know if there is a speccy emulator yet, but the gameboy emulator runs well so it proves a speccy emulator would work
[00:14] <mcphail> JanC: I might look at porting one of the javascript emulators :)
[00:21] <JanC> mcphail: although I have fond memories about the Sinclair ZX Spectrum 48K, I don't _really_ intend to run Speccy apps  :)
[00:27] <mcphail> JanC: shame on you!
[02:24] <kivi> popey, pretty sure you qualify for this: https://support.lenovo.com/us/en/documents/hf004122
[02:24] <kivi> battery recall for the thinkpad x220, and other models
[02:56] <antony_> Fixed the dpad, disabled rotation and zooming
[06:29] <zbenjamin> mcphail: why don't you use cross building instead of emulation?
[07:07] <didrocks> zsombi: hey! so here is the workaround I needed to use due to the ListModel/ModelObject automated casting: https://github.com/didrocks/splitthebill/commit/9d1e72131f7a2842bea08b4e144895ef0a06f4d2#diff-63dd6639b235200e552147bf0a9003b8R59
[07:08] <didrocks> zsombi: quite ugly I know (as we can't save back in u1db ListModel/ModelObject as they are seen as null), but maybe you would see some room of improvments? I'm opened to any suggestions
[07:10] <dholbach> good morning
[07:46] <dholbach> salut davidcalle
[07:47] <popey> morning
[07:47] <dholbach> apart from the missing note for non-current snappy docs, was there anything else that was broken or missing?
[07:47] <dholbach> hey popey
[07:48] <davidcalle> dholbach, popey, morning
[07:49] <davidcalle> dholbach, well, I'm not sure about the "Snappy" name for landing pages. I haven't found anything else. Oh, maybe the script itself should be renamed to not have snappy in its name.
[07:50] <dholbach> ok, adding a note
[07:51] <dholbach> davidcalle, ok that looks like it won't be too much work: https://code.launchpad.net/~developer-ubuntu-com-dev/developer-ubuntu-com/snappy-docs-import/+merge/265004 :)
[07:52] <davidcalle> dholbach, I'd be glad to tackle both this morning
[07:52] <dholbach> as you like it :)
[07:53] <davidcalle> :)
[08:07] <mcphail> zbenjamin: I do, for most of it, but there isn't a cross-compiler for everything and dependency management for some things in click chroots is impossible if the armhf packages are incompatible with the amd64 ones
[08:21] <didrocks> zsombi: another question once you are around on the toolkit, do we expose in any way to QML the QStandardPaths? (like the .local for one app, and so on…)
[08:21] <zsombi> didrocks: nope, we don't
[08:22] <zsombi> didrocks: we shouldn't, developers should use content providers to store their data
[08:23] <didrocks> zsombi: importing an image from content hub, the contentitem.move() api wants a path though, is there any example I'm missing?
[08:23] <zsombi> didrocks: maybe ask kenvandine
[08:24] <didrocks> zsombi: I think there is a gap missing, if you expect them to do that and they expect you to provide the path, I'll see with him then :p
[08:32] <seb128> didrocks, isn't content-hub supposed to provide you the uri?
[08:33] <zsombi> kalikiana: t1mp: up for a review? https://code.launchpad.net/~zsombi/ubuntu-ui-toolkit/configurableColumns/+merge/267520
[08:34] <didrocks> seb128: yeah, but it's storing in the cache dir
[08:34] <didrocks> seb128: there is a way to say "store in the app scope", but I still don't have the prefix
[08:35] <seb128> hum, k
[08:35] <seb128> I though content hub was supposed to store in the app private dir
[08:36] <didrocks> seb128: by default, it's the cache, there is this app scope, but you don't have access for the .move() to the prefix
[08:36] <seb128> you want to move to a specific location?
[08:37] <seb128> I know that for e.g settings bg import we just put the file where content-hub does the import and set the gsettings key value to that url
[08:37] <t1mp> zsombi: I'll check it out
[08:37] <didrocks> seb128: yeah, part of the issue is if you relocate the database to another device for instance (with different user's name then)
[08:37] <zsombi> t1mp: thx!!! resizing comes after this
[08:37] <didrocks> seb128: also, if you don't rename the file, it works the first time, then the second, as the file already exists, it's aborted
[08:38] <seb128> that seems like a content-hub issue, it should build unique names/deal with existing filenames in the import location
[08:38] <didrocks> agreed
[08:38] <didrocks> (that's why I pinged ken)
[08:39] <seb128> reported_bugs+=1 ;-)
[08:39] <didrocks> seb128: yeah, only the 6th today :p
[08:39] <seb128> ;-)
[08:49] <zbenjamin> mcphail: yeah that should be fixed though...
[08:59] <zsombi> t1mp: kalikiana: one more, a pretty small one: https://code.launchpad.net/~zsombi/ubuntu-ui-toolkit/resizableColumns/+merge/267554
[08:59] <zsombi> t1mp: kalikiana: about resizing the columns
[09:31] <didrocks> seb128: FYI, here is the way I workarounded the issue: https://github.com/didrocks/splitthebill/commit/12d380b95a4bf2a1dc58226c814da551dfb2af13
[09:31] <didrocks> (doesn't save the "clean old files" though)
[09:31] <seb128> didrocks, thanks
[09:43] <davidcalle> dholbach, pushed some changes, I've done a lot of testing with a dummy branch around what happens when docs are changed, removed, added, renamed, etc. Works great for a first iteration! Good to go on my end.
[09:44] <dholbach> <3
[09:44] <dholbach> I'll take a look in a sec and do some more testing
[09:44] <dholbach> thanks a lot for helping with this!
[09:47]  * davidcalle hugs dholbach
[09:47] <davidcalle> I'm afk for a moment (errands with Tristan)
[09:48] <dholbach> see you in a bit
[09:52] <dpm> davidcalle, when you're back - someone mentioned to me on G+ that the webapp generator is not working. Are you aware of any known issues?
[10:02] <t1mp> zsombi: I have an MR for you as well, to give the MainView and header a white background by default, https://code.launchpad.net/~tpeeters/ubuntu-ui-toolkit/white/+merge/267270
[10:48] <davidcalle> dpm, yes, fix is in staging and in the mojo spec, waiting for the next deploy (https://bugs.launchpad.net/developer-ubuntu-com/+bug/1476362) The latest deploy was somehow missing click on the server to actually build packages.
[11:04] <dpm> ack, thanks davidcalle
[11:06] <t1mp> AdaptivePageLayout blog post on https://developer.ubuntu.com/en/blog/2015/08/10/adaptive-page-layouts/
[11:08] <dpm> \o/
[11:08] <dpm> t1mp, nice work
[11:09] <t1mp> thanks :)
[11:22] <mardy> t1mp: " The contents of virtual columns that are not visible, will be shown in the right-most visible column" -- t's unclear, are they visible or not? :-)
[11:23] <zsombi> mardy: they will get visible once you put a page into them
[11:24] <zsombi> mardy: so assume you have a layout fo 2 columns defined, once you add a page using addPageToNextColumn(), the column will be filled. Same with more columns
[11:24] <mardy> zsombi: ah, ok. Still, that sentence is quite puzzling
[11:25] <zsombi> mardy: also, multiple column configuration and resizing is coming, see th eMRs I requested review for
[11:26] <t1mp> mardy: the "virtual column" is not visible, but its contents is added to the right-most visible column. If it is on top, it will be visible.
[11:26] <t1mp> mardy: right.. I made a distinction between the virtual columns and their contents there. I'm not sure if I can write it down more clearly.
[11:28] <t1mp> mardy: I didn't want to discuss every possibility in detail, because that would get boring. Instead I say this in the end: The screenhots do not cover all the possible orders in which the pages left and right can be added, so I encourage you to run the code for yourself and discover its full behavior.
[11:28] <mardy> t1mp: here's an attempt: "When a page is added to a virtual column, it will be shown in the right-most visible column"
[11:29] <t1mp> mardy: yeah thanks, I'll rewrite it similar to that :)
[11:31] <t1mp> mardy: updated. thanks
[11:38] <zsombi> t1mp: happroved
[11:50] <dholbach> davidcalle, great work
[11:51] <dholbach> davidcalle, shall we both have a quick look over the code again or shall we just ask for a review from mhall119 once he's back again?
[11:53] <davidcalle> dholbach, when does he get back?
[11:53] <dholbach> in 1.5 weeks or something?
[11:54] <dholbach> yep, 1.5 weeks
[11:54] <davidcalle> dholbach, I talked to him briefly yesterday and he wanted us to test this on staging, so what about having both a last look at the code, merge and we'll attempt a deploy to staging?
[11:54] <dholbach> <3
[11:54] <davidcalle> \o/
[12:10] <JoeyChan> hi popey,  pls try this branch:  lp:~qqworini/ubuntu-rssreader-app/reboot-add-cmake
[12:10] <mcphail> zbenjamin: there are still some packages (python ones, mostly irrc) where the :armhf package dependecies clash with the :amd64 ones, making the chroot method impossible. But, I agree, the chroot works extremely well for most things and your blog post about it is vital reading
[12:10] <JoeyChan> to c if fat package works
[12:10] <popey> sure JoeyChan
[12:11] <JoeyChan> :)
[12:13] <popey> JoeyChan: it builds!
[12:13] <popey> correctly!
[12:13] <JoeyChan> lol
[12:13] <JoeyChan> popey:   can u run it on device ?
[12:13] <popey> Let me try....
[12:16] <mcphail> antony_: The UbuntuBoy update is great! Only problem I'm having now is that a long press on a button brings up a "copy" dialog box. Don't know if this can be javascripted out?
[12:23] <zbenjamin> mcphail: \o/ thanks :D
[12:24] <mcphail> zbenjamin: :)
[12:25] <popey> JoeyChan: http://people.canonical.com/~alan/screenshots/device-2015-08-11-132530.png
[12:25] <popey> JoeyChan: \o/
[12:25] <JoeyChan> lol
[12:26] <JoeyChan> nice
[12:26] <popey> good work JoeyChan !
[12:26] <JoeyChan> :D
[12:28] <JoeyChan> popey:  so, can we set a new trunk now ?  :)
[12:30] <popey> JoeyChan: sure!
[12:31] <JoeyChan> popey:  how ?  just register a new series ?
[12:32] <popey> JoeyChan: yes.
[12:32] <popey> :)
[12:32] <JoeyChan> ok, do it now
[12:34] <popey> JoeyChan: i think you can do that by just pushing to lp:shorts-app/reboot  ?
[12:34] <popey> unless I'm mistaken
[12:36] <JoeyChan> popey:  no idea ...  :P
[12:36] <dholbach> davidcalle, I read the code again and I think I'm fine with it
[12:37] <dholbach> parts of it could be a bit simpler I feel, but for now I think it's good enough
[12:37] <popey> JoeyChan: heh, okay, lets ask dholbach :)
[12:37] <popey> dholbach: how do you register a new series in lp?
[12:37] <dholbach> popey: let me check
[12:38] <JoeyChan> popey:  done register a new series:  https://launchpad.net/ubuntu-rssreader-app/reboot
[12:38] <popey> JoeyChan: oh, my bad, I was looking in the wrong place
[12:38] <popey> yeah, duh!
[12:38] <popey> shorts-app confused me
[12:38] <popey> dholbach: nvm
[12:38] <popey> JoeyChan: https://launchpad.net/ubuntu-rssreader-app click "Register a series "
[12:38] <dholbach> yep, or https://launchpad.net/ubuntu-rssreader-app/+series
[12:39] <popey> that's it, thanks
[12:39] <dholbach> I was just looking at an LP project where I wasn't the driver of the project
[12:39] <JoeyChan> popey:   dholbach:   done  :)
[12:39] <dholbach> so I wasn't offered the choice :)
[12:39] <dholbach> JoeyChan, popey: how's life?
[12:39] <popey> stressful whenever launchpad is involved :)
[12:39] <popey> nice one JoeyChan :)
[12:40]  * popey goes to eat some haggis for lunch
[12:40] <JoeyChan> popey: dholbach but no branch linked with the series
[12:41] <dholbach> https://launchpad.net/<project>/<series>/+setbranch
[12:42] <JoeyChan> dholbach:  ok,  try it now
[12:43] <dholbach> try what?
[12:43] <JoeyChan> dholbach:  I don't know which branch to set ?
[12:43] <dholbach> how would I know
[12:43] <dholbach> I never worked on the rss reader
[12:43] <dholbach> what are you trying to do? set a branch for 'reboot'?
[12:44] <JoeyChan> ok~
[12:44] <dholbach> popey: ^ maybe a new branch needs to be created for this?
[12:45] <dholbach> bzr push lp:~ubuntu-rssreader-dev/ubuntu-rssreader-app/reboot
[12:45] <dholbach> or something
[12:45] <dholbach> :)
[12:45] <davidcalle> dholbach, same for me. Let's merge it :) I'll try the staging deploy this afternoon.
[12:45] <dholbach> davidcalle, you're a hero
[12:46] <davidcalle> dholbach, barely your sidekick!
[12:46] <dholbach> bah..... I'm just toying around most of the time ;-)
[12:47] <davidcalle> dholbach, approving the merge and pushing to trunk, can you +1 on it?
[12:47] <dholbach> davidcalle, mhall119 said he was planning to get the mojo-ised version deployed and then do a second deployment with the other stuff
[12:47] <davidcalle> dholbach, hmm
[12:47] <dholbach> davidcalle, I think I remember that he mentioned that a 'production' branch might make sense
[12:47] <davidcalle> Ok, right
[12:47] <dholbach> so we could land stuff in trunk together
[12:48] <davidcalle> dholbach, ok, creating a "production" branch then, and merging this to trunk
[12:49] <JoeyChan> dholbach:  done ! thx for your tips :D       https://code.launchpad.net/ubuntu-rssreader-app
[12:49] <dholbach> JoeyChan, rock on! :)
[12:50] <JoeyChan> popey:   series setup done ~        https://code.launchpad.net/ubuntu-rssreader-app
[12:50] <dholbach> davidcalle, we probably need a cron entry too :)
[12:51] <dholbach> davidcalle, like on the first lines of https://code.launchpad.net/~dholbach/developer-ubuntu-com/store-data/+merge/262684
[12:51] <davidcalle> dholbach, probably, yes :) also adding python-something-markdown to the mojo spec
[12:51] <dholbach> davidcalle, I added it to the pip requirements.txt file
[12:51] <davidcalle> dholbach, oh right then :)
[12:51] <dholbach> or do we need to do more?
[12:51] <davidcalle> dholbach, I don't think so
[12:51]  * dholbach has no clue
[12:51] <davidcalle> no no, it's good
[12:52] <dholbach> ok good O:-)
[12:52] <davidcalle> "afaik" :)
[12:52] <dholbach> "for funny values of 'get it working'"
[12:53] <davidcalle> dholbach, I'll try a deploy today then, there is a list of steps involving juju and sacrificing a goat, will be fun :)
[12:53] <dholbach> davidcalle, sorry, I can't see blood
[12:54] <davidcalle> :-)
[12:56] <dholbach> davidcalle, can you merge again?
[12:56] <dholbach> (sorry)
[12:57] <dholbach> https://bazaar.launchpad.net/~developer-ubuntu-com-dev/developer-ubuntu-com/snappy-docs-import/revision/162
[13:11] <karni> jhodapp: would you know who'd be best to ping regarding support for webp images? (Telegream stickers)
[13:12] <davidcalle> dholbach, I'm not on top of my cron syntax: when does it happen?
[13:12] <jhodapp> karni, I'd ping dbarth, he should be able to point you to the right person
[13:12] <karni> jhodapp: cheers
[13:12] <jhodapp> np
[13:13] <karni> dbarth: hey, I'd like to understand what's required for us to support webp images, can you point me to the right person that I could talk to about this?
[13:15] <dholbach> davidcalle, 0:15, 2:15, 4:15, 6:15, ...
[13:15] <dholbach> not sure if that's too much or not enough?
[13:16] <davidcalle> dholbach, maybe too much since we have a manual importer, but I don't have any strong feelings about it.
[13:16] <dholbach> davidcalle, every 4 hours?
[13:16] <dholbach> every day?
[13:16] <davidcalle> dholbach, on second thought, it gives a clear message: don't edit these pages, it's reverted every 2h!
[13:17] <dholbach> hah, yes :)
[13:17] <davidcalle> dholbach, let's keep it this way for staging, at least we will see if it fails after a few runs
[13:17] <dholbach> that's right :)
[13:57] <dbarth> karni: hi
[13:58] <karni> dbarth: hi
[13:58] <dbarth> karni: webp image support, i would expect that to show up in blink, for the low level parts, and maybe oxide / client apps for mime-types
[13:59] <dbarth> karni: we have the oxide weekly later today, i will put that on the map
[13:59] <karni> dbarth: thank you. for the record, what I mean is if I give a QImage a path to a webp image, I'd like that to load. Not talking about oxide here, I believe.
[14:00] <karni> dbarth: Telegram stickers are using webp format, and currently they simply don't show up.
[14:03] <dbarth> oh
[14:03] <dbarth> then that's SDK land
[14:04] <dbarth> karni: and for the record, they just show up nicely in webbrowser-app
[14:05] <dbarth> karni: so maybe use a webview for that part of the interface where you need those stickers
[14:05] <karni> dbarth: oh! I should review code then, perhaps the paths where somehow incorrect o_O
[14:06] <dbarth> a bit heavy handed, or maybe not, let me know
[14:06] <karni> dbarth: um.. no ;) I mean, that would mean a web view for the whole conversation view, which basically makes it almost a webapp :D and I can't use webview for each sticker separately (too heavy). but good tip, maybe I missed something. Thanks David
[14:07] <dbarth> karni: i read recently that instagram was using (is using?) a webview for most of its, so-called, native app
[14:07] <karni> dbarth: I believe that's what gmail did on Android as well
[14:07] <dbarth> see
[14:48] <balloons> popey, is this ready to land? https://code.launchpad.net/~mzanetti/reminders-app/uitk-1.3/+merge/265886
[14:48] <balloons> if so, we'll disable utopic.
[14:50] <popey> yes, see my comment :)
[14:51] <popey> balloons: can it be run against vivid?
[14:51] <popey> balloons: because it's all "needs fixing" which doesn't inspire confidence
[15:51] <balloons> popey, sorry, forgot to respond here. I ran it on the device under click, but I can't get to the results atm
[15:52] <balloons> when the site is back, I'll share. I agree we can remove it, and rely on another run under vivid
[15:56] <popey> balloons: that would be super, thanks
[16:44] <nemo> mcphail: hey. I'm back kinda
[16:44] <nemo> mcphail: you around?
[16:45] <nemo> mcphail: WRT the link errors. koda's speculation was needing to update the libSDL names in SDLh.pas -  the link errors are actual missing functions, not failure to locate the libs, so that makes me think it is linking against wrong SDL
[16:45] <nemo> do you have 2 copies on there, and both in the link path picked up by cmake?
[16:45] <nemo> or alternatively, maybe cmake found the wrong copy
[17:36] <mcphail> nemo: yes - my feeling is it is trying to link against SDL1.2 rather than SDL2, but I'm feeling a bit out of my depth about how to correct it. I'm working until v. late tonight, but if you have any suggestions I'll have a look in my "wind down" time before bed
[17:37] <nemo> mcphail: well. would help to check the cmake cache to see what it is trying to use
[17:38] <nemo> mcphail: also, if you could post a build log in verbose mode, that might help see what it is linking against
[17:38] <nemo> VERBOSE=1 that is
[17:38] <antony_> mcphail the dpad should be working now
[17:39] <mcphail> nemo: great - I'll dig those out for you
[17:40] <mcphail> antony_: cheers! - downloading...
[17:40] <nemo> mcphail: thanks
[17:43] <mcphail> antony_: seems a bit better, but I'm still getting a "copy" dialog on long button presses
[18:01] <mcphail> What is the command to kill the unneeded click chroots. For some reason I appear to have > 5000
[18:02] <mcphail> I think bzoltan_ or someone posted something a couple of days ago
[18:12] <davmor2> ogra_: d-day would be awesome if it didn't go white-screening on me on the longer games :(
[18:22] <ogra_> davmor2, uh, i never had that here ... do you play on arale ?
[18:22] <davmor2> ogra_: no krillin
[18:22] <ogra_> oh
[18:23] <ogra_> thats brave ... do you use a looking glass ?
[18:23] <ogra_> even on arale the font is hard to read :)
[18:23] <davmor2> ogra_: no I can see it fine, try using an xps 13 with no scaling
[18:23] <ogra_> i tried ;)
[18:23] <davmor2> ogra_: you read things
[18:24] <ogra_> well, the announcement messages of new items between the levels
[18:24] <ogra_> they use an awful font .. and are nearly unreadable on arale already
[18:24] <ogra_> i fear i cant do much about the white screen issue until the browser gets fixed though
[18:24] <davmor2> ogra_: 1-7 onwards starts to white screen at the end of the level, 1-10 is neigh on impossible to complete due to white screen :(
[18:25] <ogra_> hmm, i should perhapps add a disclaimer to the description then
[18:25] <ogra_> "only to play on arale" or some such
[18:26] <ogra_> (it didnt strike me that anyone on krillin would actually want to play it since the screen is so tiny ... i actually played 5 levels or so to make sure it runs)
[18:26] <davmor2> ogra_: or virgin krillin maybe, I'll try that on my non-dogfood device, it maybe that I have too many apps on there so not enought room
[18:29] <ogra_> davmor2, unlikely ... the white screen issue is a ram issue
[18:29] <ogra_> only open apps should matter
[18:29] <davmor2> ogra_: ah okay
[19:15] <tshirtman> mcphail: hey, so i built a modified package, installed it with pkcon install-local --allow-unstrusted, i can run it from command line, but i'm denied running system python, i intend to build and ship python for the app later, for now, can i make apparmor give me access to it?
[19:15] <mcphail> tshirtman: change the security profile to unconfined and the world's yours
[19:16] <tshirtman> ah, perfect :)
[19:17] <mcphail> tshirtman: example - http://termbin.com/w3il
[19:19] <tshirtman> ah, template, i was trying in policy_groups
[19:19] <tshirtman> thanks
[19:21] <mcphail> tshirtman: you can't upload unconfined apps to the store, but they are fine for testing
[19:23] <tshirtman> yeah, that makes sense
[19:23] <tshirtman> but it doesn't work for me yet
[19:24] <mcphail> tshirtman: got to get back to work but good luck with this. Your project looks interesting and I hope you get it working
[19:24] <tshirtman> oh wait, stypid mistake :)
[19:24] <tshirtman> stupid
[19:24] <tshirtman> "" missing around template >_>
[19:25] <mcphail> tshirtman: good old json!
[19:44] <tshirtman> hm, my app seems to take forever to start, and i don't have a log for it in ~/.cache/upstart, weird
[21:15] <tshirtman> mcphail: any idea how to get logs from my app trying to start if there is nothing in .cache/upstart/*myappname*
[21:15] <tshirtman> ?
[21:16] <mcphail> tshirtman: I don't think that is quite the correct path. 1 sec...
[21:17] <mcphail> tshirtman: ~/.cache/upstart/application-click-blah.blah.blahdeblah
[21:17] <tshirtman> there was one file of the first package i did there, but its name was wrong (and it contained the errors i had previously), so i removed it and restarted the app
[21:18] <tshirtman> and now, no file
[21:18] <mcphail> tshirtman: yes, but the application-click- bit will be prepended to the log name
[21:18] <tshirtman> yes, but * will fill it :)
[21:19] <tshirtman> sorry, i realise on irc, it may have been changed to bold
[21:19] <mcphail> tshirtman: ha! thought you were irc bolding rather than shell globbing!
[21:19] <mcphail> tshirtman: does your app just write to stdout?
[21:20] <tshirtman> np, anyway, my app is starting, write "kivy-test" with a loading symbol, and nothing happens
[21:20] <tshirtman> hm, i didn't change anything about that, i guess i could redirect to a file in run.sh
[21:21] <mcphail> tshirtman: no, writing to stdout would be the correct thing to do...
[21:22] <mcphail> tshirtman: is this a compiled app or a script?
[21:23] <tshirtman> i took your template, changed the content of run.sh to run python ~/path/to/main.py
[21:23] <tshirtman> and used click to build, and pkcon to install-local
[21:23] <tshirtman> (changed the other files for name and other things)
[21:24] <mcphail> tshirtman: you didn't bundle main.py in the .click?
[21:24] <tshirtman> no, but i guess i could
[21:24] <tshirtman> i redirected output, and the file is not created, so i wonder if my command actually runs
[21:25] <mcphail> tshirtman: might be worth trying a simple python one-liner bundled in the click first, and build up from there until it breaks. Note that I think only python3 is installed by default
[21:26] <tshirtman> i apt install-ed python and a lot of other stuff
[21:27] <mcphail> tshirtman: also note that running a click changes the working directory (hence all relative paths) to the click install location
[21:27] <tshirtman> also, that's probably for other chan, but if there is a way to grow / that would be useful
[21:27] <tshirtman> i used ~/ as much as necessary
[21:27] <tshirtman> so it should be ok on that side
[21:27] <tshirtman> i could hardcode absolute path, to be sure, though
[21:28] <mcphail> tshirtman: try a simple python "hello world" first and build from there. So many things could be causing the breakage
[21:29] <tshirtman> yeah, will do
[21:29] <tshirtman> actually, i'll even test "echo hello world" first
[21:29] <tshirtman> that should go to the log in .cache, right?
[21:29] <mcphail> tshirtman: yes - if you have an unconfined app
[21:29] <tshirtman> i should
[21:30] <mcphail> tshirtman: as soon as you confine your app, you have to bundle shell commands in the click. That's why my syncthing click bundles sed, for example
[21:31] <tshirtman> hm, nothing, should i reboot the device? :D
[21:32] <mcphail> tshirtman: very odd
[21:32] <tshirtman> also, if i was doing something illegal (app confined), it would crash, right?
[21:33] <tshirtman> here it's just pretending it's loading the app, and nothing happens, the loading wheel just turns and turns
[21:33] <popey> if the app tries to do anything it shouldn't you'll see apparmor failures in dmesg
[21:33] <mcphail> tshirtman: yes - confinement should post error to the log
[21:33] <popey> so "sudo dmesg | grep DEN"
[21:33] <tshirtman> i was seeing errors in dmesg before
[21:34] <popey> would show "DENY" style messages with your app name and what it's trying to do
[21:34] <tshirtman> but not since i have this forever loading one
[21:34] <tshirtman> yeah, i had these before i put uncontained
[21:34] <mcphail> tshirtman: can you share your .click?
[21:34] <tshirtman> sure
[21:34] <tshirtman> when the device is booted again :)
[21:34] <mcphail> :)
[21:35] <tshirtman> ah, now it crashes again, fun
[21:35] <popey> yeah, lets have a play :)
[21:35] <tshirtman> (and the icon got fixed, it was invisible… fun)
[21:35] <mcphail> tshirtman: going to have to do a bit of work for 30 mins, so would be great if you could email/send link (my email is in the syncthing manifest file)
[21:36] <tshirtman> will do
[21:36] <mcphail> cheers - see you in a bit
[21:36] <tshirtman> but it may work now, hello world in file
[21:36] <tshirtman> i guess the system was confused somehow
[21:36] <tshirtman> by previous version of the package
[21:36] <mcphail> progress!
[21:36] <tshirtman> yes
[21:36] <tshirtman> baby steps, baby steps
[21:37] <mcphail> tshirtman: I always delete the old package before changing anything
[21:37] <tshirtman> the .click
[21:37] <tshirtman> ?
[21:37] <tshirtman> or uninstall?
[21:37] <mcphail> uninstall
[21:37] <tshirtman> ok
[21:37] <tshirtman> i'll do that now
[21:39] <tshirtman> that broken readline >_<
[21:41] <tshirtman> ok, new error, sdl2 - RuntimeError: Failed to created a window surface 0x1
[21:42] <tshirtman> i guess i can try with the libs pointed at earlier
[21:50] <tshirtman> core dump \o/
[21:51] <popey> \o/
[21:51] <popey> covering all the error condition based here tshirtman
[21:53] <tshirtman> i guess that must be because my kivy was built against a different version of sdl2
[21:54] <tshirtman> so i'll rebuild against the source of these ones
[21:56] <tshirtman> adb push is slow, i got the binaries from github faster than i'm pushing them through usb >_<
[21:56] <tshirtman> orders of magnitude faster
[22:07] <mcphail> tshirtman: the version of sdl2 in the repos will segfault in SDL_INIT. Stick my version in the lib/arm-linux-gnueabihf directory and you should be golden
[22:10] <tshirtman> mcphail: i put them in my app's dir /sdl2 and added LD_LIBRARY_PATH=sdl2/ before my python command, so it uses them, so now, instead of being unable to create a window, it segfaults, so i assume it's because of an api mismatch, since i built kivy against the libsdl2-dev from the repos
[22:10] <tshirtman> but i'll try moving to lib/arm-linux-gnueabihf and remove the LD_LIBRARY_PATH, it's easier to start with
[22:11] <mcphail> tshirtman: I don't htink that should happen. The headers are very similar
[22:12] <mcphail> tshirtman: (that is lib/arm-linux-gnueabihf within the .click package, btw)
[22:12] <tshirtman> yeah, relative path, so in the source dir that i packager, right?
[22:12] <tshirtman> package*
[22:13] <mcphail> yes
[22:13] <tshirtman> i don't patch the .click manually :D
[22:13] <mcphail> :)
[22:13] <mcphail> tshirtman: you _can_ play with the installed click under /opt/ if you need to
[22:13] <popey> yeah
[22:14] <popey> i frequently muck about with the contents in /opt/
[22:14] <tshirtman> hm, that could be a faster way to iterate
[22:14] <mcphail> tshirtman: provided you dn't change anything in the .desktop file or .apparmor file you are usually OK
[22:14] <tshirtman> hm, nope, still coredump
[22:15] <tshirtman> i guess i can't run it through gdb?
[22:15] <mcphail> tshirtman: you can
[22:15] <tshirtman> hm, though, if it coredump, i should be able to take that
[22:15] <mcphail> tshirtman: you can install gdbserver and debug remotely
[22:16] <tshirtman> ah yeah, forgot about that option
[22:16] <tshirtman> hm, maybe the coredump file is saved somewhere already, though?
[22:16] <mcphail> dunno
[22:18] <mcphail> popey: while you are here, can you remember the command bzoltan_ (or someone) gave to get rid of spurious click chroots?
[22:18] <tshirtman> darn, no ulimit command
[22:19] <mcphail> popey: for some reason I've got >5000 on this device and it is killing bash completion for "mount"
[22:20] <popey> yes
[22:21] <mcphail> tshirtman: which version of ubuntu are you running on your device?
[22:21] <popey> i have put it in my tidyup scripts
[22:21] <popey> sudo schroot -e --all-session
[22:22] <mcphail> popey: cheers! Maybe I will be able to type "mount <tab>" without my CPU melting now
[22:22] <popey> hah
[22:22] <tshirtman> mcphail: stable, i used "$ ubuntu-device-flash touch --channel=ubuntu-touch/stable/ubuntu --bootstrap" to flash
[22:22] <popey> tshirtman: i have ulimit on my device
[22:23] <mcphail> tshirtman: ok - so probably using mir0.12?
[22:23] <tshirtman> ah, that's some fun, ulimit as a user is found, but not as sudo
[22:23] <tshirtman> sudo -s to the rescue
[22:24] <popey> heh
[22:24] <tshirtman> hm, maybe i didn't need sudo actually
[22:25] <mcphail> tshirtman: one thing to check would be the mir version. I think my build of sdl only works for <=0.12. 0.13 and above need to be built from a special branch
[22:26] <mcphail> (I think)
[22:27] <tshirtman> hm, apt search shows 0.13 versions, for mir-client-platform-* so i must have that
[22:28] <mcphail> popey: what mir version is on the OTAs?
[22:28] <popey> uhhh
[22:28] <popey> 0.14 on rc-proposed
[22:29] <tshirtman> and i suppose downgreading is both hard and wrong, so the less hard option will be to rebuild sdl2 and kivy :) hopefuly not too much FUN
[22:29] <popey> up/downgrading isnt hard
[22:29] <popey> you can switch back and forth easily
[22:29] <tshirtman> for deb packages?
[22:29] <popey> oh, well, yeah
[22:29] <popey> if the older version is in the repo
[22:29] <tshirtman> :)
[22:29] <popey> apt-get install foo=version
[22:29] <popey> but I wouldn't use apt on the phone
[22:30] <tshirtman> too late :D
[22:30] <popey> heh
[22:30] <popey> its fine
[22:30] <popey> for hacking
[22:30] <tshirtman> i installed vim git and everything to build kivy already
[22:30] <tshirtman> well, / is 2GB, so i still have to be careful
[22:30] <tshirtman> managed to fill it once already
[22:31]  * mcphail notes it is 0.13 on the default OTA so that theory was wrong
[22:32] <tshirtman> so your build should work?
[22:32] <mcphail> tshirtman: yes, I think so. Works here...
[22:33] <tshirtman> ok, back to gdp then :)
[22:33] <tshirtman> b*
[22:33] <mcphail> tshirtman: can you build and run the sdl-template app from my/sturmflut's github repo?
[22:34] <tshirtman> probably :)
[22:35] <tshirtman> https://github.com/Sturmflut/ubuntu-touch-sdl2-gles2-template ?
[22:36] <tshirtman> can i just clone it in the device and run click build? or do i need to run cmake& make and stuff?
[22:36] <tshirtman> (is click magicall for this? :o)
[22:37] <mcphail> tshirtman: you'd need to create a click chroot to cross-compile it
[22:37] <mcphail> tshirtman: or open it in the sdk qtcreator
[22:37] <tshirtman> right, so stop hacking on the device and install a real dev env :)
[22:38] <mcphail> :)
[22:41] <tshirtman> http://www.theorangenotebook.com/2014/05/building-cross-platform-click-packages.html ?
[22:44] <mcphail> tshirtman: https://developer.ubuntu.com/en/blog/2015/04/13/inner-workings-sdk/
[22:46] <tshirtman> thanks
[22:46] <tshirtman> but actually, i installed cmake on the device, because why not, will probably do a real dev env sometime later :) that will help a lot
[22:47] <mcphail> tshirtman: no wonder you are running out of space!
[22:47] <tshirtman> still 57M!
[22:47] <tshirtman> cmake is actually not that big, i was surprised
[22:48] <mcphail> enough for anyone!
[22:48] <tshirtman> so the demo works
[22:48] <tshirtman> no background to the app, so it's on top of the gui, funny
[22:48] <mcphail> great - so SDL should work OK
[22:48] <mcphail> tshirtman: I think that is a mir bug or something
[22:49] <mcphail> tshirtman: transparency/corruption is a problem
[22:49] <tshirtman> ohh, the effect when using app switch (left swipe) is very nice, the cube seems to stand out
[22:50] <tshirtman> well, it's not hard to add a background to the app :)
[22:50] <tshirtman> but that's great, i guess i can try building my kivy against the source in this one, and see how it goes
[22:50]  * mcphail hopes tshirtman is getting exited about the possibilities of porting to ubuntu touch :)
[22:50] <tshirtman> i am :)
[22:51] <tshirtman> i hope the ubuntu devs are excited by the idea of getting a load of apps easily ported to ubuntu-touch :P
[22:51] <tshirtman> well, some kind of a load
[22:51] <mcphail> tshirtman: I've got another couple of hours of work to get through before bed so I'll leave you to it for now. Good luck!
[22:51] <tshirtman> thanks a *lot* for your guidance
[22:51] <tshirtman> much appreciated
[22:51] <mcphail> my pleasure