[10:57] http://bazaar.launchpad.net/~sidi/notify-osd/xfconf-experimental/diff/375/373?remember=373 MacSlow this should be the patch for gconf/xfconf & for color/text size/opacity settings. You can have a look if you want but i still need to do some testing and to improve the autotools part before proposing it for a merge [10:58] SiDi, atm I'm working on backlog support [10:58] SiDi, current plan is to address gconf/xconf in two weeks (depending how fast I'll get other stuff sorted) [10:59] SiDi, gconf/xconf are low-impact (side-effect wise) imo... so I'll rather do them at the end (close to feature-freeze) [11:00] gee... lp is slow today [11:01] MacSlow: no problem as long as it goes in karmic :) [11:01] we need it for xubuntu [11:01] SiDi, argl... please no renaming of #defines [11:02] i can revert that if you want. They were not actually needed anymore as they were defined, but some fallback ones were needed. I renamed them to make it more clear what they were used for [11:05] SiDi, unit-tests should be done also [11:06] Can they be shell scripts ? [11:06] they could but that would be the easy/lazy way :) [11:06] they need to be properly integrated into "make check", "make test" [11:06] Using gconf/xfconf is much much much nicer via CLI than with C (especially gconf :d) [11:07] ok. it'll take me some time to figure how to do this :/ [11:07] yeah [11:07] MacSlow: would you like me to write a template of configuration framework, too ? [11:08] unit-tests are unpleasant I know [11:08] what do you mean by "template" thereß [11:08] ? [11:09] a defaults-template.c/h file, defining the conf_* functions that are needed (which would return some random values), so that if someone wanna write configuration with a file parser or another conf storage utility, they have a template to get started with [11:09] SiDi, nah don't bother [11:10] Okey [11:10] SiDi, that would best be done the GObject-way ... for consistency reasons [11:10] I don't think i understand more than 1% of gobject, MacSlow :) [11:11] and regarding the tweaking-capability we would not over-commit [11:12] SiDi, don't worry... your in good company... at least I can understand any possible "pain" from trying to wrap your brain around it if you tried before :) [11:12] especially if one is used to OO coming from C++ or Python [11:14] Yeh... if i really want efficient OO i prefer writing some Vala or some C++ (even if im not a fan of C++ at all). Gobject is way too tortuous [11:17] is './configure' || './configure --with-xfconf' acceptable or should i try to get something like './configure --conf-storage=gconf||xfconf' MacSlow ? [11:20] --with-gconf or --with-xconf (while being mutually exclusive) [11:20] ok [11:25] * MacSlow -> lunch === MacSlow is now known as MacSlow|lunch === beuno_ is now known as beuno === MacSlow|lunch is now known as MacSlow [12:50] SiDi: ping -- do you have a bug to track your progress with the gconf work? Some bugs against notify-osd are easily solved with these options, so I'd like something to mark them duplicates of :) [12:51] njpatel: i've got 3 bugs [12:51] one for color/opacity, one for text size, one for xfce [12:52] 394432 335383 331949 [12:52] and there is a branch, lp:~sidi/notify-osd/xfconf-experimental [12:52] i'm clearing the autotools and writing unit tests now, then it'll be ready for review, njpatel [12:52] SiDi: awesome, I'll start marking some of these bugs as dups, then [12:53] as you wish :) [13:17] * SiDi waves at dashua [13:22] Hey SiDi :) [13:22] You are a coding animal lately. [13:24] https://updown.ubuntuone.com/d286d58c-6db6-4d06-b34f-956c3f069af4 [13:24] Polished theme if anyone cares to test :) [13:25] dashua: i'm a dev after all :P [13:25] blergh, ubuntuone is broken with xfce Q_Q [13:25] Ah bah [13:27] * SiDi hates makefiles and autotools [13:34] SiDi, http://www.speedyshare.com/616120880.html [13:37] I know you're not a fan of the titlebar gradient, but I think it works in this case subtly [13:37] blargh, remember i run xfwm :P [13:38] which reminds me i have to refresh several themes for karmic @(^_^)@ [13:39] (btw, im still using Alvaro, dashua ;) ) [13:41] Oh nice [13:44] http://www.ubuntu-pics.de/bild/22305/screenshot_005_e6sgtG.png [13:44] I should polish up Alvaro's metacity [13:44] I haven't worked on it in some time [13:44] Your screenies look very nice on the Exaile page [13:54] This may compliment the new sexxy GDM somewhat. Iteration is nice. === kenvandif is now known as kenvandine [14:10] thanks dashua :) [14:10] the new progress bar owns in this screenshot dashua [14:11] and the metacity is actually quite cool. Did you try with a cream color such as tooltips' instead of white for metacity ? [14:12] hehe... dashua has disk errors , palimpsest disk fail? [14:16] mac_v, Ha yeah, that gnome-disk-utility is sensitive [14:17] SiDi, cream for the text? [14:17] dashua: you can turn it off , once you realize its not fatal [14:18] I don't where this utility appeared from. Never had it before. [14:18] Ah cool. [14:20] dashua: its a new karmic "feature" , you can turn it off from the startup apps , when Bug #412152 is properly fixed you can turn it back ON ;) [14:20] Launchpad bug 412152 in gnome-disk-utility "gnome-disk-utility nags me too much that my disk is failing" [Medium,Triaged] https://launchpad.net/bugs/412152 [14:20] mac_v, Thx :) [14:25] mac_v, I'm happy about this. -> http://www.ubuntu-pics.de/bild/22309/screenshot_006_51bYiB.png [14:26] (the bottom gradient looks weird ^.^) [14:26] (how do the ooo menus look like ?) [14:28] Bah not good [14:28] OOo is my theming nemesis [14:29] Oh the bottom is part of the metacity [14:30] http://imagebin.ca/img/JDL1f6Kl.png huh [14:37] dashua: looks great , but the bottom could be thinner ;) no pun intended [14:38] Ha ok. Thx for the feedback. [14:39] Remove the bottom gradient, you think altogether? [14:41] dashua: i'm not sure about removing , but generally a thinner border would be better [14:41] remove :P [14:41] Alright ;) [14:43] dashua: i really like the new scrollbars :) , awesome , link to the updated theme pls [14:44] MacSlow: any idea why gconf_client_get_float returns 0.0 and doesnt set error if the gconf key doesnt exist ? :x [14:45] mac_v, bzr branch lp:hanso [14:45] SiDi, check the GError [14:46] its NULL [14:46] ah... nice [14:46] SiDi, not the returned value... 0.0f can be returned if the requested gconf-key didn't exists [14:46] The sexxy scale and progressbar through were removed as it uses nodoka [14:47] SiDi, use the GError... don't irgnore it by just passing NULL to it in gconf_client_get_float() [14:47] MacSlow: i mean the error var is still null after the call to gconf_client_get_string [14:48] SiDi, from the GConf ref.-manual: "the value of key, or 0.0 if no value is obtained" [14:48] oh [14:48] so 0.0 is another error case :/ [14:48] SiDi, are you talking about gconf_client_get_float() or gcon_client_get_string()? [14:49] get_float [14:49] http://paste.ubuntu.com/255134/ MacSlow [14:50] returns : "before 0.8 | after 0.0 | returns 0.0", and never goes inside the "error set" part [14:54] SiDi, see what you get as error http://paste.ubuntu.com/255139/ [14:57] MacSlow: my problem is that, while there is an error (the key isnt set), the "error" var is still NULL after the gconf call [14:57] it doesnt initialize it as it should. and gconf-editor does not show the gconf key, im sure it doesnt exist [14:58] SiDi, do you have your branch somewhere public I can pull from? [14:59] lp:~sidi/notify-osd/xfconf-experimental [15:03] mac_v, SiDi, http://www.ubuntu-pics.de/bild/22310/screenshot_007_dX39ZY.png [15:05] dashua: hehe , i can see the disabled options having a problem , notice the forward , i had the problem while hacking Dust and dropped it since ti was too much work [15:06] SiDi, I'm on a conf-call atm [15:06] MacSlow: okey [15:07] SiDi, conf as in conference... not x/gconf :) [15:07] MacSlow: crap, was gonna use xfconf-query on you [15:08] mac_v, The nodoka elements? [15:08] It was too hard to support the theme when the engine is not packaged [15:10] dashua: yeah , its tough to get the disabled option the right shade for OOo and the rest , else it would look more like a highlighted item , rather than a dimmed one [15:10] Ah yeah. I see what you mean. [15:16] MacSlow: dont bother with that bug, i workarounded it by using fallback value if the opacity is set to 0.0.. [16:29] DanRabbit: ping [17:02] MacSlow: what the hell should the unit tests actually do ? :/ === beuno is now known as beuno-lunch [17:02] should i change the keys via gconf/xfconf and check that bubbles' properties have the good value ? [17:03] SiDi, test every introduced call with wrong keys and verifiing that they return default values (thus notify-osd continues to run and display stuff properly) [17:03] SiDi, and then test each gconf/xconf value properly (depending on what system your running on) [17:04] okey [17:04] im completely failing to find out how to build my test file, not to mention checking bubbles' properties :D [17:04] SiDi, as an example/template look at... notify-osd/tests/test-notification.c how to do it [17:04] okey [17:05] SiDi, the initial problem is to design the code in a way so it can easily be unit tested [17:05] SiDi, sometimes it's simple... sometimes it's less than obvious [17:05] e.g. stuff like DBus or gconf/xconf make unit-testing less obvious [17:05] MacSlow: i dont see how i could properly know that my bubble has finished redrawing after i used gconf or xfconf to change a value [17:06] SiDi, no... no need fo rthat [17:06] SiDi, just make sure (test) for the expected values (or fallback-values) [17:06] you'll probably only have to setup and query class Defaults I think [17:06] in defaults, then ? [17:06] yes [17:06] okey [17:07] automatic testing/verification of rendered output is very hard to test against [17:08] SiDi, you'll probably have to #ifdef some of the tests depeding of wether you compiled with xconf or gconf [17:08] SiDi, that's actually bad... but I don't currently see how to avoid that [17:09] SiDi, apart from forcing anybody wanting to run "make test" to install gconf and xconf (+ dependecies) [17:09] which would/will annoy a certain kind of people ;) [17:10] SiDi, unit-tests are dull and boring... but a very good way to thoroughly ensure all your written API-code is solid and bullet-proof [17:11] MacSlow: i was thinking of test-xfconf & test-gconf.c, and choosing which to use with Makefile.am [17:11] SiDi, that's why unit-test-cases should also pass "crap" to calls to verify it does crash/blowup or so [17:12] SiDi, don't you think it's simpler to just add to test-defaults.c? [17:12] SiDi, that'll save you some setup and figuring out [17:12] now that i know this file exists, yes, MacSlow :D [17:12] SiDi, if you want to run just your (the unit-test for class Defaults) [17:12] do [17:12] cd notify-osd/tests [17:13] ./test-modules -p /defaults [17:13] okey [17:13] that'll speed up your turn-around cycles while working on your unit-tests [17:14] SiDi, I'll be going offline in a few minutes... so should you have further questions... either email them to me ot try to catch me later (in 3 hours) here [17:16] MacSlow: sure. thanks for your patience :) [17:17] btw, can i put the DEFAULT_ defines in default.h instead of default.c ? [17:17] so i can use them from test-default.c instead of copy/pasting values [17:17] * MacSlow scratches head... [17:18] at the moment you just repeat the values in test-defaults, which will require not forgetting about updating them when you change the defaults [17:18] SiDi, that would introduce no harm... yes... that can be savely done [17:20] okey [17:20] SiDi, you're patch against notify-osd will become pretty large i can imagine... so don't be surprised that I'll will reshuffle things a bit and split it up in smaller (~800 loc) chunks for review later [17:20] MacSlow: i setup a branch and separated commits a little [17:21] SiDi, that's just common practice within our team [17:21] No problem :) [17:21] SiDi, just so that you know how I tick :) === beuno-lunch is now known as beuno [17:56] mac_v: I have a paper cuts task if you want it [17:56] djsiegel1: hmm... what is it? [17:57] move all paper cuts assigned to ayatana to invalid for ayatana, and tag them "ayatana" instead [17:57] mac_v it's only 17 bugs [17:58] djsiegel1: huh? move them to where? [17:58] which project? [17:58] no moving [17:58] mark invalid in ayatana [17:58] and tag them with "ayatana" [17:59] djsiegel1: ok... but what do you want me to state as the reason for that? [18:00] no reason needed [18:00] ayatana is being deleted and re-created [18:00] ok. [18:00] djsiegel1: by when do you want this done ? [18:01] now if you can, otherwise I can do it [18:01] djsiegel1: ATM , i'm busy ... [18:01] ok, no prob [18:17] djsiegel1: why is ayatana being deleted ? :| [18:17] SiDi: it's just being renamed [18:17] it's called ayatana-project, and someone wants it renamed to ayatana [18:18] which means it has to be deleted first... [18:18] why dont you just ping the LP admins about it ? :P [18:27] SiDi: LP admins are the ones who told me to do this [18:27] mac_v: r7 is also short 1 paper cut [18:28] djsiegel1: r10 will has a few more than needed , you could pick one you like ;p [18:28] ok, cool === rickspencer3 is now known as rickspencer3-afk === vorian is now known as rofl [19:47] djsiegel1: is synaptic being replaced in karmic itself? [20:14] mac_v: Oo? [20:15] SiDi: o.0? [20:15] is synaptic gonna be replaced ? [20:17] SiDi: eventually by appcenter ,either karmic or karmic+1 [20:19] i really need to subscribe to ubuntu-devel... [20:19] how many mails / day does that represent ? [20:20] SiDi: thats not in devel , i just gave you some top secret info [20:21] bleh [20:21] how do you know top secret info ? [20:21] SiDi: well i have my sources... [20:22] * SiDi gently tortures mac_v [20:22] I'm listening. [20:23] SiDi: only if you promise not to tell any one else? [20:23] Sure. [20:23] (btw, google appcenter :p) [20:23] SiDi: https://wiki.ubuntu.com/AppCenter [20:24] ;p === rickspencer3-afk is now known as rickspencer3 === rofl is now known as marvelous [22:50] * SiDi waves @ MacSlow [22:50] Was gonna send you an email :P [22:50] SiDi, just do it anyway :) [22:50] hi btw [22:51] MacSlow: mirco.mueller@canonical.com ? [23:01] SiDi, yes