[04:58] <pitti> Good morning
[05:06] <didrocks> good morning
[05:06] <pitti> bonjour didrocks !
[05:08] <didrocks> bonjour pitti, comment ça va ?
[05:08] <pitti> didrocks: c,a va bien !
[05:09] <pitti> eh -- mon cle' de compose e'st casse' !
[05:09] <pitti> happens once in every blue moon
[05:09] <didrocks> c'est l'heure de passer en azerty !
[05:10] <pitti> jamais !
[05:10] <didrocks> ne jamais dire jamais :p
[05:11] <pitti> ah, changer la configuration du clavier et retour l'arrange
[05:12] <didrocks> bizarre, comme si la configuration changeait toute seule
[06:34] <larsu> good morning!
[06:37] <didrocks> hey larsu!
[06:37] <pitti> hey larsu
[06:38] <larsu> hi pitti and didrocks! How are you?
[06:38] <pitti> larsu: quite well, thanks!
[06:38]  * larsu convinced half the systemd team to join telegram yesterday :)
[06:38] <didrocks> larsu: Good thanks! Fighting with different behavior of pep8 through releases though :/
[06:38] <larsu> didrocks: uh oh. good luck!
[06:38] <didrocks> larsu: how are you?
[06:38] <larsu> pitti: awesome. Eating tasty breakfast at a new cafe
[06:38] <larsu> didrocks: ^
[06:38] <pitti> yummy
[06:39] <didrocks> heh ;)
[06:39] <larsu> includes a croissant
[06:39] <larsu> without chocolate ;)
[06:39] <pitti> larsu: avec du chocolat ? :-)
[06:39] <larsu> ha! Got there first l)
[06:39] <didrocks> phew!
[06:39] <didrocks> that was close
[06:39] <didrocks> avoided a catastrophe
[06:39]  * pitti takes off the asbestos pants again
[06:39] <pitti> err
[06:40]  * pitti perde le pantalon d'asbesto à nouveau
[06:40] <larsu> hum?
[06:40] <pitti> désolé pour la language fausse
[06:40] <didrocks> don't get it either :p asbesto?
[06:40] <pitti> for the chocolate flamewar
[06:40] <pitti> nevermind
[06:41] <pitti> "avoided a catastrophe"
[06:41] <didrocks> I get it's not that asbesto: https://en.wikipedia.org/wiki/Asbestos
[06:41] <didrocks> guess*
[06:41] <pitti> didrocks: that's the very thing you should have when entering fire/a flame war
[06:42] <pitti> didrocks: don't tell me you've never heard of asbestos before? Or the proverbial pants made of it?
[06:42] <didrocks> ah, ok, getting you now :) Never heard about the proverbial pants made of it though, no
[06:42] <pitti> wow
[06:52] <RAOF> You've never run around in your pretty little asbestos suit?
[06:59] <larsu> all the time@
[06:59]  * larsu has problems with punctuation recently
[06:59] <larsu> s/@/!
[07:31] <Laney> HAI
[07:31] <larsu> morning Laney!
[07:32] <Laney> what's up
[07:32] <larsu> not mch
[07:32] <larsu> you?
[07:32] <pitti> hey Laney
[07:32]  * Sweet5hark1 see the asbestos chocolate mix in backlog and is suspicous ...
[07:36] <didrocks> good morning early Laney :)
[07:36] <didrocks> hey Sweet5hark1
[07:36] <Laney> also not much
[07:36] <Laney> yeah so I can go early to climb ^_^
[07:39] <didrocks> heh
[07:59] <willcooke> hey ho
[07:59] <TheMuso> Hey willcooke.
[08:00] <willcooke> hey TheMuso
[08:03] <didrocks> hey willcooke, TheMuso
[08:46] <didrocks> pitti: since a few weeks, everytime I update systemd, I do have the unity lockscreen taking over, is that known?
[08:47] <didrocks> I guess the issue is either unity or systemd side (like logind…)
[08:47] <pitti> didrocks: yes, bug 1473800
[08:48] <pitti> didrocks: I haven't found time yet to dig into details, just found a smaller reproducer
[08:48] <didrocks> pitti: as long as it's widespread and in your capable hands, I trust it will get fixed then ;)
[08:48]  * didrocks hugs pitti
[08:48]  * pitti hugs didrocks back
[08:49] <pitti> didrocks: matter of time mostly
[08:49] <didrocks> yeah, I justed wanted to confirm that wasn't setup-specific
[08:49] <didrocks> just*
[08:52] <pitti> didrocks: no, seems to happen everywhere
[08:52] <darkxst> hey pitti didrocks willcooke
[08:52] <pitti> hey darkxst, how are you?
[08:53] <didrocks> good evening darkxst!
[08:53] <darkxst> pitti, good, finally kicked the flu this week!
[08:54] <pitti> didrocks: good to hear! did it stick around for that long?
[08:54] <willcooke> hey darkxst
[08:54] <darkxst> pitti, it was a solid 2 weeks of bad
[08:58] <darkxst> pitti, plymouth themes are not in initramfs right?
[08:58] <willcooke> Trevinho, thanks for you comments on the doc.  Do you need me for your Kylin meeting?
[08:59] <pitti> darkxst: they will be if you install e. g. cryptsetup
[08:59] <darkxst> yet I uploaded a fix Mon or so and it only kicked in today (apparently, according to a few users)
[08:59] <didrocks> darkxst: mirrors delay perhaps?
[08:59] <darkxst> pitti, ah, fix was for password dialof
[08:59] <Trevinho> willcooke: well, no, but if you want to give your welcome (as 1st time), feel free ;)
[08:59] <darkxst> dialog
[08:59] <didrocks> ah ok then :)
[09:04] <darkxst> pitti, so I suppose it will never happen again, but how does a package force an initramfs rebuild
[09:05] <pitti> darkxst: the postinst can call update-initramfs -u, if the update-initramfs command is available
[09:07] <darkxst> pitti, ok
[09:15] <Trevinho> Laney: ehy, can you give a look at https://code.launchpad.net/~3v1n0/ubuntu/wily/gnome-terminal/scrollbar-background-fix/+merge/270605 ?
[09:16] <Laney> Trevinho: busy atm, hopefully will do later
[09:16] <Laney> what did upstream say?
[09:17] <Trevinho> Laney: well, they want to wait for a proper solution for now... Which is scrolled window, but it needs gtk patching
[09:17] <Trevinho> Laney: also they want disable OSB by default in terminal... But I think that's because of a misunderstanding. Since OSB doesn't cover content. This happens in the terminal BTW (when using the experimental and very broken first patch)
[09:18] <Laney> "this"?
[09:19] <larsu> .c
[09:19] <Laney> .h
[09:19] <didrocks> .go
[09:19] <larsu> I should disable my touchpad
[09:19] <larsu> accidentally hit it while typing a filename
[09:20]  * larsu is happy this didn't happen while typing a password :)
[09:20] <Laney> I wonder why disable while typing doesn't really work
[09:20] <Trevinho> shouldn't touchpad be disabled temporary while writing?
[09:20] <Trevinho> ah, ok.. :)
[09:20] <larsu> Laney, didrocks: good reply, though ;)
[09:20] <Trevinho> In mine it works
[09:20] <Laney> think it's a driver thing
[09:20] <larsu> "should"
[09:20] <darkxst> larsu, disable while type should always be on now, unless you set it otherwise in the past
[09:21] <larsu> darkxst: never did
[09:21] <larsu> I think it's just broken on this system?!
[09:22] <larsu> what's the settings key for that?
[09:23] <darkxst> larsu, oh no settings key, it was completely removed actually
[09:23] <larsu> makes sense
[09:23] <darkxst> it should always be on!
[09:23] <larsu> ya, clearly
[09:23] <larsu> nobody would ever want to turn it off, unless it's buggy
[09:24] <darkxst> larsu, don't underestimate 'everybody;
[09:25] <Laney> Trevinho: maybe larsu could review that patch code-wise in the meantime ... :)
[09:25] <Trevinho> I think so :) (:-* larsu)
[09:29] <Laney> a suspicious silence begins
[09:30] <Laney> :D
[09:30] <larsu> I'm looking at it already!
[09:32] <Laney> ♥
[09:32] <larsu> Trevinho: this is the bug about porting vte to gtkscrolledwindow https://bugzilla.gnome.org/show_bug.cgi?id=733210
[09:33] <larsu> Trevinho: and apparently overlay scrollbars won't be enabled in terminal by default before then
[09:33] <larsu> but your patch is only to get the colors right, even on non-os?
[09:33] <Trevinho> larsu: yeah, I know... I linked that even on LP and on my buzilla bug...
[09:33] <Trevinho> larsu: however that patch (only for terminal) is broken
[09:34] <Trevinho> larsu: it needs some gtk patching
[09:34] <larsu> Trevinho: ugh, I guess this is why there's no activity on it
[09:34] <Trevinho> larsu: here's more info https://bugzilla.gnome.org/show_bug.cgi?id=754796#c8
[09:34] <larsu> Trevinho: I'm fine with your patch (if it works) until they have a proper solution
[09:34] <larsu> yeah I saw that one
[09:34] <larsu> thanks
[09:35] <Laney> does it go with some theme fix?
[09:35] <Trevinho> larsu: he says that they won't be enabled by default in any case... But as I commented I think that's because they implemented it badly and so they see an issue when interacting with it
[09:35] <larsu> Trevinho: want to fix it?>
[09:35] <larsu> :P
[09:35] <Trevinho> larsu: about my patch... yeah, it only clears the colors under the scorllbar
[09:35] <larsu> Laney: no I think it should work as-as
[09:35] <larsu> *is
[09:35] <Laney> ok cool
[09:35] <Trevinho> larsu: I wish, but... Well no time for that right now :)
[09:35] <Laney> will test this later
[09:35] <larsu> Trevinho: ;)
[09:36] <larsu> Laney: thanks
[09:36] <Trevinho> larsu: yeah, it goes without any theme fix. Theme can be improved once that is landed, though
[09:36] <Laney> man
[09:36] <Trevinho> Laney: how does it work for landing it? Does it use the train? As in that case we can just send alltogether...
[09:36] <Laney> no
[09:38] <Laney> I just upload this
[09:38] <Laney> oh MAN
[09:38] <Laney> I just remembered which way round you use "--" between gbp and bzr bd
[09:38] <Laney> maybe I am finally learning this
[09:39] <larsu> :)
[09:39] <Laney> there should be an international treaty
[09:39] <Laney> about consistency of command line interfaces
[09:40] <Laney> transgressions prosecuted at the hague
[09:40] <Trevinho> git guys would be at that hague for years now
[09:40] <larsu> Laney: hahahhaha welcome to posix
[09:41] <larsu> wtf
[09:41] <larsu> G_FILE_CREATE_REPLACE_DESTINATION exists but g_file_create() doesn't look at it
[09:43] <larsu> ah, g_file_replace() does...
[09:43] <Laney> doesn't g_file_create() say that the fie has to not exist?
[09:44] <larsu> it does, but it takes that flag and silently ignores it
[09:49] <davmor2> Trevinho: well the guys aren't gonna help if you start by calling them gits ;)
[09:57] <Laney> Oh crap
[09:57] <Laney> update-rc.d: error: Read-only file system
[09:58] <Laney> This bug is the biggest ARGH
[10:49]  * Laney commits crimes
[10:50] <pitti> Laney: mounting stuff r/w? :-)
[10:54] <Laney> I am running as root :)
[10:57] <davmor2> Laney: Gggggroot
[11:00] <Laney> I am kloot
[11:00]  * Laney puts them on
[11:00] <didrocks> phew, just thought about renewing my bike rental subsciption before going :)
[11:03]  * didrocks goes for a run
[11:09]  * Trevinho feels like a newbie... But.... when you're in a bug, isn't really any way to switch focus from upstream project to downstream without changing the uri manually (so that I can target/nominate for a serie/version)?
[11:13] <ogra_> Laney, anything wrong with "echo "manual" >/etc/init/whoopsie.override ? (why do you want update-rc.d ? )
[11:13] <Laney> ogra_: it already uses that
[11:13] <Laney> and, init system agnostic
[11:13] <ogra_> then i dont understand the last bug comment
[11:14] <Laney> which part of it?
[11:14] <ogra_> "update-rc.d does renames"
[11:14] <Laney> they don't work
[11:14] <ogra_> if you already use the cho
[11:14] <Laney> this is well known
[11:14] <ogra_> *echo
[11:14] <Laney> i don't use the echo
[11:14] <Laney> the code already uses update-rc.d
[11:14] <ogra_> oh, i understood you the other way round :)
[11:15] <Laney> it makes sense to use it
[11:15] <ogra_> "<Laney> ogra_: it already uses that"
[11:15] <Laney> we just have to work around brokenness
[11:20] <Laney> ogra_: how on earth do you test livecd-rootfs/lxc-android-config changes?
[11:20] <Laney> just upload and make an image?
[11:24] <ogra_> Laney, install the deb from a chrootewd recovery works ...
[11:25] <ogra_> for livecd rootfs i usually test the resulting changes only .... (you can do local builds but thats a lot of effort)
[11:25] <Laney> I did the changes manually
[11:25] <Laney> maybe I just upload this if whoopsie-preferences works
[11:40] <desrt> morning
[11:40] <desrt> anyone want to go to a concert in the hammer this evening?
[11:41] <larsu> hi desrt
[11:41] <larsu> no thanks, going to the movies tonight
[11:41] <desrt> shucks
[11:41] <desrt> everyone is bailing on me
[12:59] <willcooke> desrt, what time?
[13:07] <desrt> willcooke: this evening, starting around 7
[13:09] <willcooke> Doesn't seem to be a direct bus route from here. :(
[13:09] <larsu> willcooke: you're on an island....
[13:09] <desrt> hamilton is a pretty small airport
[13:09] <willcooke> They should have thought about this
[13:09] <larsu> heh
[13:09] <desrt> which is annoying
[13:10] <desrt> they're talking about building a second large airport in toronto
[13:10] <Laney> you could get quite close
[13:10] <Laney> just cross the bering strait
[13:10] <desrt> imho they should just add an ICE from toronto union to hamilton airport and expand it there
[13:10] <desrt> (lol.. imagine that, a real train network in canada)
[13:44]  * larsu wonders if we already have translated country names somwehere
[14:03] <willcooke> am I here?
[14:04] <davmor2> willcooke: no you are there I am here
[14:04] <willcooke> ah ha!
[14:04] <larsu> is that a philosphical question?
[14:04] <willcooke> larsu, more a technological one
[14:04] <willcooke> my internet went out in the middle of a hangout with didrocks
[14:04] <larsu> in that case, the answer is probably yes
[14:05] <larsu> french connections aren't very good
[14:05] <willcooke> ha, I bet didrocks disagrees
[14:06]  * didrocks downloads some gigabytes to proove larsu wrong
[14:06] <davmor2> willcooke: you know the Film the French Connection?  That's faster paced than their French connection ;)
[14:06] <didrocks> but I need to find some GB of something to download :p
[14:07] <larsu> hehe
[14:08] <larsu> didrocks: I should shut up .. .only on lte :/
[14:09] <didrocks> oh right!
[14:21] <willcooke> Dell has sold out of XPS13 Ubuntu editions in Europe!
[14:21] <larsu> desrt: like so?
[14:22] <desrt> yes
[14:22] <desrt> telegram for work is weird
[14:22] <desrt> it's too mixed in with my personal life
[14:22] <larsu> indeed. You started there :)
[14:22] <desrt> and it's difficult to manage priority of messages this way
[14:22] <larsu> oh wait I started with work stuff
[14:22] <desrt> also: all of the beeping is getting annoying
[14:22] <larsu> indeed :)
[14:22] <larsu> so... I think I'm writing a db. stop me
[14:23] <desrt> you don't have enough features for a DB :)
[14:23] <larsu> I mean the api
[14:23] <desrt> GVariant is a pretty great API for what you're doing
[14:23] <desrt> maybe with some very very thin layering of convenience API on top
[14:23] <desrt> ie: lie
[14:23] <larsu> lie about what?
[14:23] <desrt> make it look like you're returning "City" objects or some BS
[14:23] <desrt> but really it's just a GVariant (ssss)
[14:23] <desrt> have nice get_name() etc.
[14:23] <desrt> and city_free() -> g_variant_unref()
[14:24] <desrt> as for how you manage which items you return, well, i don't know
[14:24] <larsu> yeah that makes the most sense I guess
[14:24] <larsu> and is extensible
[14:24] <desrt> right
[14:25] <desrt> if some day you decide not to use GVariant, then you don't break API
[14:46] <Sweet5hark1> cyphermox: any help neeeded for sponsoring still?
[14:47] <cyphermox> no, I'm getting to it now
[14:47] <Sweet5hark1> cyphermox: great, thanks
[15:06] <larsu> desrt: g_variant_new_from_data() gives me an empty array ... how do I start debuggin this?
[15:07] <larsu> g_variant_get_size() is correct
[15:07] <larsu> g_variant_n_children() returns 0
[15:21] <mzanetti> anyone here knows how to change the keyboard layout in a unity8-desktop-session-mir?
[15:27] <desrt> larsu: probably you didn't put an array into the memory you're trying to load
[15:27] <desrt> your code was appending a bunch of (ssss) in sequence, wasn't it?
[15:27] <desrt> you can't do that... you need to use a builder
[15:27] <larsu> desrt: I did, but it was still compressed (got the data from the wrong stream)
[15:27] <larsu> sorry :)
[15:27] <desrt> ah
[15:27] <desrt> that'll do it too :)
[15:27] <larsu> ya, writing it as a a(ssss) now
[15:27] <larsu> Laney would be proud
[15:28] <Laney> bahahaha
[15:28] <larsu> 83ms to rip through all of it
[15:28] <desrt> i welcome a patch to add streaming support to GVariantBuilder
[15:28] <desrt> this was one of the design goals of GVariant
[15:28] <larsu> tokenizing and folding every iteration
[15:28] <desrt> but nobody implemented it yet
[15:28]  * larsu wonders if he should even bother making that faster
[15:28] <larsu> desrt: didn't you say it's impossible?
[15:28] <desrt> no.  it's easily very possible.
[15:29] <desrt> GVariant is designed in such a way that the natural way of using a builder (with nesting) results in all of the data being written immediately with no back-seeking required
[15:29] <Laney> larsu: try on a phone
[15:29] <Laney> or the armhf porter :)
[15:29] <desrt> only the offsets are required to be held until the end of each nesting level
[15:29] <desrt> ie: a few bytes per item, regardless of the size of the item
[15:29] <desrt> so it's quite efficient
[15:30] <larsu> Laney: same porter but going into an armhf schroot?
[15:30] <Laney> porter-armhf.c.c
[15:30] <larsu> ok will try later
[15:30] <Laney> actually those might be fast
[15:30] <Laney> if it's armhf on arm64
[15:30]  * Laney doesn't remember
[15:30] <larsu> I can always build an index as an optimization
[15:31] <larsu> probably we're already much faster than before because we're not creating all those objects
[15:31] <desrt> at which point seeking will be very useful :)
[15:31] <desrt> you're sort of exploring a bad side of GVariant here
[15:31] <larsu> Laney: but we still need to fix bug #1439475
[15:31] <desrt> since you _will_ be creating lots of objects (GVariants)
[15:31] <Laney> larsu: sure
[15:32] <larsu> Laney: I wonder where we should put country translations...
[15:32] <Laney> larsu: orthogonal though, don't get tangled up trying to fix all the things at once
[15:32] <larsu> inded
[15:32] <larsu> *indeed
[15:32] <Laney> probably geonames.org (where this database comes from) has something
[15:32] <larsu> at least now we can search for Nëw York and get New York ;) (thanks desrt ;) )
[15:32] <desrt> that's not true.
[15:32] <desrt> you can search for New York and get Nëw York, but not the other way ;)
[15:32] <larsu> Laney: it does, it's even in our dataset
[15:33] <Laney> cool
[15:33] <larsu> desrt: right sorry
[15:33] <Laney> what's the problem then?
[15:33] <larsu> we probaly want to not ship all of them
[15:33] <larsu> I'd expect the original name and the german name on a system with de_DE
[15:34] <desrt> it has nothing to do with the system locale
[15:34] <desrt> and everything to do with the locale of the name itself
[15:34] <desrt> which is actually really interesting
[15:34] <larsu> hm? That's what I'm just saying
[15:34] <desrt> you have a chance to handle that in an intelligent way here, since the city names come from particular countries
[15:34] <larsu> I want the original name AND the name in the system locale
[15:34] <desrt> oh.  weird.
[15:35] <desrt> good luck :)
[15:35] <larsu> why?
[15:35] <desrt> because this is m*n for m city names and n locales
[15:35] <desrt> and your database doesn't contain this info
[15:35] <desrt> you mean like köln vs. cologne, right?
[15:35] <larsu> yes
[15:35] <desrt> and brussels vs bruxelles vs brussel
[15:36] <larsu> it does... but I think it doesn't tell me which languages alternate names come from
[15:36] <pitti> brüssel *cough*
[15:36] <desrt> you mean bruessel :)
[15:36]  * larsu pats pitti on the back to help with the cough
[15:36] <Sweet5hark1> pitti: bless you
[15:36] <larsu> desrt: g_str_match_string() doesn't do that one yet ;)
[15:37]  * Laney is off for some errands (skipped lunch) - ttyl
[15:37] <desrt> it does, if you tell it your locale is "de"
[15:37] <desrt> that's the purpose of the "translit_locale" argument to g_str_tokenize_and_fold()
[15:37] <larsu> does it use my system locale if I pass NULL
[15:37] <larsu> NEAT!
[15:37] <desrt> no
[15:37] <desrt> it uses C
[15:38] <desrt> anyway.. my point is that you need to figure out which language each of these names is from so that the transliteration can handle it intelligently
[15:38] <larsu> yeah I think I don't know that
[15:38] <larsu> so ... include all of them?!
[15:38] <desrt> that would be weird
[15:38] <desrt> ö for example appears in german and swedish, but you don't agree on what the ascii-ification rules are
[15:39] <larsu> in u-c-c, I can find all of brüssel, brussels, and bruxelles right now
[15:39] <larsu> I think someone decided to index ALL THE THINGS
[15:39] <desrt> how about bruessel? :)
[15:40] <larsu> nope
[15:40] <desrt> time to step up your game
[15:40] <larsu> not running de_DE
[15:40] <desrt> here's a hint: use wikipedia
[15:41] <desrt> "this page in other languages" is your friend :)
[15:41] <larsu> uh oh
[15:41] <desrt> Bruselský region
[15:41] <desrt> :)
[15:42] <desrt> ბრიუსელი_(რეგიონი)
[15:42] <desrt> 布鲁塞尔-首都大区
[15:42] <desrt> could be that this way lies madness...
[15:42] <larsu> indeed
[15:43]  * larsu gets feature parity on the phone first
[15:43] <larsu> but this will need to work
[15:43] <desrt> (madness and large DBs)
[15:43] <larsu> not *that* large
[15:43] <desrt> all the cities of the world translated into all of the languages of the world?
[15:44] <larsu> most of them are 1:!
[15:44] <larsu> 1:1
[15:44] <desrt> i wouldn't bet on that...
[15:58] <desrt> so that was weird
[15:58] <larsu> oh you're back
[15:58] <desrt> lost internet in my house, and my phone wouldn't work either
[15:58] <desrt> traceroute showed that both connections were dying at the main internet exchange in toronto
[15:58] <desrt> seems to be back on my phone now...
[15:59] <desrt> home connection is also working again
[15:59] <desrt> scary feeling... losing internet and not even being able to use your phone to get around it :p
[15:59] <larsu> desrt: is returning GList still cool?
[15:59] <desrt> no
[15:59] <desrt> never was
[15:59] <larsu> zero-terminated array?
[16:00] <desrt> slightly nicer....
[16:00] <larsu> iterator?
[16:00] <desrt> are you returning an array of GVariants?
[16:00] <larsu> yes
[16:00] <desrt> did you consider returning a GVariant array?
[16:00] <larsu> it's not in the API
[16:00] <desrt> huh?
[16:00] <larsu> GVariant isn't
[16:00] <larsu> but yeah, could do I guess
[16:01] <desrt> oh.  you want people to use your array natively?
[16:01] <larsu> I want people to not use g_variant_
[16:01] <desrt> right
[16:01] <larsu> in case I change my mind
[16:01] <desrt> i'd use an array with a length field
[16:01] <desrt> and null termination for convenience
[16:01] <desrt> like:
[16:01] <larsu> and then require calling free() in a loop?
[16:01] <desrt> gpointer* get_the_thing(thing, int *n);
[16:01] <larsu> and again for the container?
[16:01] <desrt> ya
[16:01] <desrt> this is where it starts to suck
[16:02] <larsu> right
[16:02] <desrt> how about GListModel? :D
[16:02] <larsu> ..
[16:02] <larsu> more and more I dislike the way we do c
[16:02] <desrt> the way we do C is nice
[16:02]  * larsu shouldn't have to think about this shit
[16:02] <desrt> you're just trying to do it without fully committing to doing it
[16:03] <desrt> so you're in this weird in-between way
[16:03] <desrt> and of course that's going to suck
[16:03] <larsu> fair enouhg
[16:03] <larsu> it's all because of qt then!
[16:04] <desrt> also: don't think that using another programming language would free you from having to decide about datatypes...
[16:04] <larsu> returning a list in python is totally accepted
[16:04] <desrt> except that probably you'd actually want to do it as a generator, right?
[16:04] <larsu> or most other languages for that matter
[16:04] <larsu> heh
[16:04] <desrt> :)
[16:04] <desrt> same shit everywhere...
[16:05] <desrt> besides... the most python-list-like thing that we have is GPtrArray
[16:06] <larsu> returning that was never cool, either
[16:06] <desrt> you still haven't decided if you want to reuse the same thing when submitting a new query or not
[16:06] <desrt> sounds like you're half way between designs on that one
[16:07] <larsu> I have
[16:07] <larsu> I won't for now
[16:07] <larsu> but might add a GListModel on top
[16:07] <desrt> if i was doing this i'd do something a bit crazy:
[16:08]  * larsu is all ear
[16:08] <desrt> admit, on your API, that you internally have an indexed array of items from 0 to [whatever]
[16:08] <desrt> and have a global API for "get me item number 7"
[16:08] <desrt> then have your search return a sorted list of indexes
[16:09] <larsu> hm, intriguing
[16:09] <desrt> you can write a convenience API on top of that if you'd like it to return the things themselves, but whatever
[16:09] <desrt> the nice part about this is that it makes filtering the list when you type more-and-more specific search terms very easy
[16:09] <desrt> you just scan through the sorted array of index numbers and very quickly notice which ones are no longer there
[16:09] <desrt> doing a GListModel on top of this, for example, would be trivial
[16:09] <larsu> but I have to redo all of it when backspace is hit
[16:09] <desrt> nah
[16:10] <desrt> you just do the opposite
[16:10] <desrt> notice which ones got added
[16:10] <desrt> really, your merging algorithm would handle both cases, of course
[16:10] <desrt> but in practice it would only ever be doing one or the other during a given iteration
[16:10] <larsu> I mean I'd have to go through the whole list again
[16:10] <larsu> to find the ones that were addded
[16:10] <desrt> yes.  of course.
[16:11] <desrt> in my mind you're scanning the full list every time you do a search anyway
[16:11] <larsu> how do you suggest this works? Pass in an array of previous matches?
[16:11] <larsu> (indexes, of course)
[16:11] <desrt> i'm only talking about the model level.. and maybe widget proxies
[16:11] <desrt> and maybe animations....
[16:11] <desrt> depending on how crackful you want to get
[16:11] <larsu> we can always do these one level up...
[16:11] <desrt> yes.  precisely
[16:12] <desrt> the nice thing about ints is that the ints are _identity_
[16:12] <desrt> in a very natural way
[16:12] <larsu> an "id" string is as well
[16:12] <desrt> if you try to do that with pointers then you have to play the caching game
[16:12] <desrt> and then you can't sort them
[16:12] <larsu> yeah...
[16:12] <larsu> true
[16:13] <desrt> anyway... just an idea
[16:13] <desrt> and not sure how well it works for cities with multiple names
[16:13] <larsu> I'm getting around to it :)
[16:13] <larsu> they have the same int, no?
[16:13] <larsu> so guint * query (string)
[16:13] <desrt> but then what do we sort by?
[16:13] <larsu> merge with the previous match
[16:13] <desrt> the name of the city in which language?
[16:13] <larsu> sort
[16:14] <larsu> desrt: thats the other problem we discussed earlier, no?
[16:14] <larsu> hm, geonames also has an id for each city
[16:14] <desrt> ya... of course.. just mentioning that there is potentially an interesting intersection here
[16:14] <larsu> but not consecutive
[16:15] <larsu> desrt: I like the part where I don't have to return a weird array
[16:15] <desrt> is it something reasonable to (user-visibly) sort on?
[16:15] <larsu> just a guint* and a length
[16:15] <larsu> no, not really
[16:15] <larsu> and having proper indices will make it easier on the backend for me
[16:16] <desrt> and the g_free(array) thing is nice
[16:16] <larsu> (looking into the giant a(ssss) directly)
[16:16] <larsu> indeed
[16:16] <desrt> so i think you have your API then
[16:16] <larsu> desrt: thanks :)
[16:16] <desrt> then you can (easily) do some fancy model stuff on top of it later, and efficiently so
[16:16] <desrt> no prob
[16:16] <larsu> an qt can do the same
[16:16] <larsu> if Laney's up for it
[16:49] <tjaalton> does wily have the gio bits for google drive integration?
[17:20] <willcooke> tjaalton, don't think so.  They were introduced in 3.18 and we're on 3.16.
[17:28] <willcooke> g'night
[22:21] <TheMuso> tkamppeter: Just a heads up, you accidentally deleted some entries from the desktop 16.04 planning sprint blueprint when you added your entry. I've fixed it. I'm guessing you already had the page open, and edits were made since you had that page open.
[22:30] <tkamppeter> TheMuso, sorry, it was open for some days and I simply clicked the button for editing the Whiteboard.
[22:31] <tkamppeter> TheMuso, Bugzilla does better in such a case, if yoiu submit based on an outdated state of the page, you get a warning  message.
[22:39] <TheMuso> tkamppeter: Yeah as I said, I figured something like that was the case. Probably worth filing a bug against launchpad about that.