=== tmpRAOF is now known as RAOF [07:13] landing \o/ [07:50] hiya [07:50] can sombody please take a look at https://code.launchpad.net/~ubuntu-mate-dev/compiz/mate-consistent-keybindings/+merge/258251? [07:51] it's in the sponsoring queue, so I thought I'd prod you guys about it :) [08:07] mzanetti: so should we start another landing? [08:22] pstolowski: it seems that at the end my landing took predence over the concierge one? [08:26] tsdgeos, hmm, indeed [08:26] tsdgeos, i'll rebuild === davidcalle_ is now known as davidcalle [09:42] hey bregma, do you know somebody who could take a look at https://code.launchpad.net/~ubuntu-mate-dev/compiz/mate-consistent-keybindings/+merge/258251? [09:42] dandrader: something is borked at https://code.launchpad.net/~dandrader/unity8/app-state-handling/+merge/258653 [09:42] there's a few ================ [09:43] tsdgeos, ok, will check it more closely. maybe I didn't catch all conflicts [09:45] tsdgeos, I think it's a bug in launchpad's web diff view [09:46] tsdgeos, ah, it must be because its prerequisite has conflicts with trunk [09:46] ah, can be [09:47] i also marked those [10:11] dholbach, I'll be working my way through all the outstanding Compiz MPs over the next couple of days: I just got back from a couple weeks of vacation and things piled up [10:11] <3 [10:11] thanks a lot bregma [10:16] tsdgeos, hey, can you please install the nichtlustig scope and tell me if that's a bug in unity8 or in the scope? [10:17] mzanetti: E_NO_CONTEXT [10:17] erm... huh? [10:18] tsdgeos, it doesn't show anything? [10:18] mzanetti: what's the nichtlusting scope? [10:18] tsdgeos, no the diff looks correct. had to redo the stack of commits. [10:18] is it on the shop? [10:18] tsdgeos, s/no/now [10:19] tsdgeos, yes [10:19] tsdgeos, this is the card definition it uses: https://developer.ubuntu.com/en/scopes/guides/scopes-customization-branding/ [10:19] wrong link [10:19] E_TOO_MANY_SIMULTANEUS_CHATS [10:19] this one: http://bazaar.launchpad.net/~mzanetti/+junk/nichtlustig-scope/view/head:/src/scope/query.cpp#L26 [10:20] now the question is why the spacings between the tiles are broken [10:26] mzanetti: ok, give me a sec [10:29] mzanetti: so what's wrong exactly? [10:29] the empty spacing? [10:29] or? [10:34] tsdgeos, yeah [10:34] tsdgeos, it's more... first I'd say the spacing is too big. [10:34] tsdgeos, sometimes though, it's missing completely [10:35] refresh the scope and it'll change [10:35] not here, it's always the same [10:36] tsdgeos, scroll down the scope [10:36] tsdgeos, also, you'll see cachebuffer operating too early and tiles disappear before the are leaving the scene [10:36] not here [10:36] mzanetti: are you on rtm or vivid? [10:36] vivid [10:37] ok, going back to scopes and using the manage scopes got me the bad spacing [10:37] can't get the tiles to disappear [10:37] but that may happen at some point if they're wrongly layouted [10:37] http://i.imgur.com/d7tmwlz.png [10:38] this is when I scroll to the bottom [10:38] 100% reproduceable by doing: [10:38] * favorite the scope [10:38] * go to it using horizontal swipes [10:38] * pull down to refresh it [10:39] then slowly scroll to the bottom, watch topmost ones disappear too early, see spacing broken for bottom cards [10:41] mzanetti: can i get the scope for the desktop please? [10:41] :/ [10:44] mzanetti: can i get the scope for the desktop please? [10:44] i guess we're at the point already that some stuff is much easier to get on the phone than on the desktop [10:53] ok compiled manually [10:53] i'll have a look at least as to why there's the empty space === MacSlow is now known as MacSlow|lunch [11:05] not sure why my IRC highlight has gone today [11:05] tsdgeos, you still need something? [11:10] tsdgeos, also I'd have another one of those questions. If you install the "südwestpresse" scope (search for "swp" in the store). then open a preview, you'll see that the image in the header is too large [11:10] again I think even if the scope developer provides an image that doesn't fit properly, we should probably set the size ourselves somehow [11:14] mzanetti: no, it's ok [11:15] i've a fix at least for the empty space [11:15] tsdgeos, cool :) [11:16] not sure if that'll fix the clipping [11:16] i'm hoping it will === dandrader is now known as dandrader|afk [11:24] mzanetti: https://code.launchpad.net/~aacid/unity8/card_height_overlay_header/+merge/258756 [11:25] tsdgeos, will this still work for cards that have a headeroverlay AND and a summary? [11:25] given that the summary is still outside the art in that case [11:26] mzanetti: the if for the summary is higher in the chain [11:26] ack === dandrader|afk is now known as dandrader === alan_g is now known as alan_g|lunch === MacSlow|lunch is now known as MacSlow [12:52] mzanetti: have you tried the patch? === alan_g|lunch is now known as alan_g [13:03] mzanetti: also answered https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1453546 [13:03] Ubuntu bug 1453546 in unity8 (Ubuntu) "cannot use color:/// uri for scope header foreground and divider color" [Undecided,New] [13:46] tsdgeos, well, regardless if it's intentional or not... it's bad [13:46] IMHO, that is [13:46] why ? [13:47] how would you use a gradient in the foreground color or in the divider line? [13:47] because for example the swp scope. the blue in the logo can't be the same blue as the one of the back arrow [13:47] tsdgeos, not necessarily a gradient... but a color:///#aarrggbb [13:47] well that works of course [13:47] just don't use color:/// [13:47] tsdgeos, not for me [13:48] ah... ok... then it's the doc that sucks [13:48] mhall119, think we can improve that? ^^ [13:49] mzanetti: i don't know why it sucks [13:49] the docs say one should use color:///#aarrggbb, but apparently only "#aarrggbb" works [13:50] mzanetti: no, one should use color://// in the fields that support background uris [13:50] read the part i quoted on the bug [13:50] and which fields are that? [13:50] "some" is really not precise enough imo [13:50] mzanetti: where is that? [13:50] agreed [13:51] mhall119, https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1453546 [13:51] Ubuntu bug 1453546 in unity8 (Ubuntu) "cannot use color:/// uri for scope header foreground and divider color" [Undecided,New] [13:51] mzanetti: actually it's pretty defined i think [13:51] the ones that say "background scheme" it's a "background scheme" (i.e. the color:///) stuff [13:51] the rest just say "color" [13:51] and color is "red" or "#aarrggbb" [13:52] but i guess it could be improved [13:52] it took me a while to understand it and i kind of know what it says === elopio_ is now known as elopio [13:52] also it should say that for !some one should use either an svg color name or a "#aarrggbb" value [13:52] but I'm so sure I tried that... it only painted black all the time for me [13:52] let me try again [13:53] well then that's a bug [13:55] it works... seems I didn't try that [13:57] so i guess it needs a rewording [13:57] saying "in the above list color means bla and background scheme means blo" [13:58] yes === marcusto_ is now known as marcustomlinson [13:59] tsdgeos, why making it different though... [13:59] it should always be the same, except background one additionally supporting gradient:/// [13:59] imo [14:00] mzanetti: and images in urls [14:01] i don't know i didn't do that spec [14:01] i'm just telling you what it says and how we implemented it [14:01] honestly i have no opinion [14:01] sure. [14:07] greyback, what optimization does a const int parameter bring? [14:14] dandrader: it's more for humans here tbh. Without the compiler will usually have to copy the parameter, so for objects that can be a cost. For int, it's pretty cheap, so not much of a gain. But I like using const whenever possible [14:14] http://www.gotw.ca/gotw/081.htm is nice summary [14:18] tsdgeos, hey, i guess you want somebody from unity8 team to top-approve https://code.launchpad.net/~aacid/unity8/edit_reviews/+merge/258623 ? [14:18] pstolowski: i'd like to yeah [14:18] greyback: const won't stop the compiler doing a copy of the object, const & will [14:19] tsdgeos, i will prepare a silo with our stuff [14:19] also it's actaully slower to pass POD as const & than just by copy [14:19] greyback, I don't see the point of a "const int". But I do see it for "const FooStruct&". In the latter you avoid copying over a struct. but and int? nothing is smaller than an int [14:20] tsdgeos: you're right [14:20] greyback: http://www.macieira.org/blog/2012/02/the-value-of-passing-by-value/ is a very nice article about it [14:20] when you have days to digest it all [14:20] tsdgeos: what say you? "const int" vs "int" [14:20] i went straight into the conclusions :D [14:21] greyback: no i mean const int & vs int [14:21] const int vs int i'd say it's the same [14:22] dandrader: then it's merely a hint for humans [14:22] but one I like [14:22] and the compiler in case you try to change it :D [14:23] cimi: can you do https://code.launchpad.net/~aacid/unity8/edit_reviews/+merge/258623 ? [14:23] I would say that "const int" param is a copy of an int that the function code can't modify. no optimization [14:23] tsdgeos, sure [14:23] cimi: cheers [14:24] greyback, I don't know. I've never seen a function taking a "const int"..... [14:24] dandrader: yep, we've already established it's not a performance optimization, it's a hint for human & compiler [14:25] greyback, I think this hint does more harm than good, as it bloats the signature for no real benefit [14:26] I mean, who care is the function implementation changes the int it gets as a parameter? [14:26] dandrader: 6 chars != bloat in my book [14:26] *cares if [14:26] mzanetti, hey, ok if i request a silo with unity8 branch and related stuff for review editing functionality? [14:27] pstolowski, sure. can we add some little more stuff into that? [14:27] pstolowski, I would help with the testing abviously [14:29] * mterry upgrades to wily [14:29] mzanetti, sure [14:32] tsdgeos, approved your branch. fixes the nichtlustig scope spacings [14:32] tsdgeos, however, I can still see cards disaappear too early [14:32] damn [14:32] ok, i'll try to reproduce harder [14:35] tsdgeos, I can only repro it with the nichtlustig scope... not for example with the xkcd one that also uses header-overlay with no summary [14:39] tsdgeos, can you bump the version with your edit_reviews branch? i need to make click scope depend on that, otherwise the scope will break === dandrader is now known as dandrader|lunch [14:49] mzanetti: it's probably not a journal? [14:49] pstolowski: bump the version of what? [14:49] tsdgeos, unity8 [14:50] doesn't every release do that? [14:50] or you mean bump to 8.03 ? [14:50] tsdgeos, sure, but i need the number upfront to update my control file :) [14:51] tsdgeos, right... [14:51] pstolowski: not sure what i have to do then :D [14:51] mzanetti: ↑↑↑ ? [14:52] tsdgeos, actually.. i wonder if it's a good idea to introduce such dependency [14:52] tsdgeos, i've never done that before [14:52] tsdgeos, meh, scratch that [14:52] pstolowski: ok :D [14:53] mzanetti, feel free to add more stuff to line #70 [14:53] pstolowski, ack [14:54] pstolowski, I see this targetting wily. how are you handling that? [14:54] does that mean editing reviews will only work in november? [14:55] MacSlow, hey... so, about that shellrotation AP tests. what's the status? [14:55] which raises the question of [14:56] shall we jump to wily or what? [14:56] yeah... current situation is not really clear at all [14:56] mzanetti, ready for more review eyes [14:57] MacSlow, has that long standing issue been fixed now? [14:58] mzanetti, qml-cache issue can be avoided by unsetting QV4_ENABLE_JIT_CACHE... this isn't really an issue that branch is meant to solve. [14:58] ah ok [14:58] mzanetti, apart from that I've 100% success-rate [14:58] ok. sounds good [14:58] thanks [14:59] mzanetti, I'll just hope to dig myself 100% into unity8-launcher work now... as that _much_ more fun than the ap-test and related parts :) [15:00] ok [15:01] mzanetti, btw... QV4_ENABLE_JIT_CACHE really has to be unset... just assigning 0 can still keep the QML-cache bug happen again [15:03] mzanetti, just wiping ~/.cache/QML before the ap-test-run isn't enough, because the shell-rotation AP-test is using scenarios now - upon QA-team request - thus on ap-test-run causes multiple unity8 start/stop cycles which in turn recreate the qml-cache if $QV4_ENABLE_JIT_CACHE is defined [15:04] MacSlow, so does that mean if we land it, our jenkins runs will fail? [15:04] mzanetti, no... it can [15:05] mzanetti, the failure-rate due to the qml-cache bug is... about 20% [15:05] so yes, our jenkins runs will fail then [15:05] mzanetti, iirc we have a bug files for this too... [15:05] filed [15:05] mzanetti, i'm not sure about willy; i'm definately planning to land in vivid as wel [15:05] l [15:06] mzanetti, i assume willy is kind of required, then we need another silo for vivid overlay [15:06] willy :D [15:07] gosh ;) [15:07] hahaha [15:07] pstolowski, if that's supposed to to go into vivid, then I think you should just target vivid in the spreadsheet [15:07] mzanetti, sure, but we need to target both, no? [15:08] pstolowski, mzanetti: what a difference a 'l' can make :) [15:08] pstolowski, and that's where I'm a bit lost too tbh [15:09] pstolowski, so I guess the only way out would be to branch vivd+overlay in the repos, land to wily and backport to the other... [15:09] which is a pain [15:09] but anything else is just pointless in one way or the other === dholbach_ is now known as dholbach [15:10] mzanetti: ok, i can actually reproduce it, it's just more subtle that i expected [15:10] tsdgeos, I have a feeling after your first fix it's less visible, yes [15:12] mzanetti, ok, i got some clarity about click scope part of it at least; it has 15.04 branch which I need to target for "vivid"; trunk is for W [15:12] aha! [15:12] yeah, that's what I meant with the backporting thing [15:12] it's really the only thing that works [15:13] we don't have a 15.04 branch in unity though... I suppose I should just create it then [15:14] mzanetti, yes... i can small a lot of fun [15:14] pstolowski: "small" "willy" - what fun typos are you making ;) [15:14] mzanetti, i've changed the spreadhseet to target vivid [15:14] ok, cool [15:14] greyback, :) === dandrader|lunch is now known as dandrader [15:58] mzanetti: ok, found out the problem [15:58] i'm not taking into account the various margins around the journal when calculating what is really seen [15:58] so if we were clipping it'd be right, btu since we're not need to take them into account [15:59] aha [15:59] will fix tomorrow [16:00] * tsdgeos waves [16:40] my laptop battery only charges up to 76% any more :'( === alan_g is now known as alan_g|EOD === greyback__ is now known as greyback