=== ChanServ changed the topic of #ubuntu-uds-client-1 to: Currently no events are active in this room - http://summit.ubuntu.com/uds-1403/client-1/ - http://irclogs.ubuntu.com/2014/03/13/%23ubuntu-uds-client-1.html === jasoncwarner___ is now known as jasoncwarner === ChanServ changed the topic of #ubuntu-uds-client-1 to: Track: Client | Unity 7 Defaults and Settings for 14.04 LTS | Url: http://summit.ubuntu.com/uds-1403/meeting/22206/client-1403-unity7-defaults-and-settings/ [13:59] hello! [13:59] heya all. [13:59] hey [13:59] bregma, hello [14:00] seb128, good day [14:00] bregma, do you host or should I start this one? [14:00] seb128, I have no idea how to host, so maybe you should [14:00] seb128, all: I have no working hangouts setup right now, but will follow the session cast and try my best to type real fast on IRC ;) [14:01] ok, I'm asking because I started sessions as usual to find out that in some cases other already did that for themself [14:01] Sweetshark, ok [14:01] bregma, https://plus.google.com/hangouts/_/hoaevent/AP36tYcTs29nNDDMeCm2QozMy3TTDdokJyEQQnjxepvHPj7ZN8EAyw?authuser=0 [14:02] those who want to be in the hangout ^ [14:04] who wants to come? [14:04] ;-) [14:04] ara, ? [14:04] you said you were interested [14:04] sure, joining [14:04] thanks [14:06] http://youtu.be/Or2bipn5A3o [14:06] stream [14:07] seb128: works (but has echos) [14:07] seb128: (doh, only has echos when you have two tabs with it) [14:07] Sweetshark, i had the same issue :) [14:10] If you have a laptop with touchscreen and add an external screen without touch the touch area is stretched over both screens. [14:11] "the touch area"? [14:11] OveRisberg, that's a bug [14:12] Ok... is it a known bug? [14:12] nothing to do with that session though [14:12] and my mike doesn't work. [14:12] I think we should do LIMs.. [14:14] OveRisberg, bug #1287341 [14:14] Ubuntu bug 1287341 in gnome-settings-daemon (Ubuntu) "Touchscreen controls both screens in dual-monitor setup" [Undecided,Incomplete] https://launchpad.net/bugs/1287341 [14:15] seb128: Thanks [14:16] Do we think we can really make a list of hardware that doesn't lie? [14:18] I would *think*/*hope* a monitor that doesn't give correct info would give you at lease a default 96 DPI (which is to small for a HiDPI monitor) [14:18] I don't think we need data to know that LIM are better, and letting OEMs set non-default settings for something like this seems confusing [14:18] qengho: only sir mix-a-lot devices dont lie [14:18] bschaefer: Too many say "I'm 6 feet wide" or something. [14:19] best case it gives correct info, and we scale, worst case no info and scale 1 [14:19] qengho, eww, well if its outside the range of 300/400? DPI we'll just default to 1.0 [14:20] I think we should limit range to between 1 and 2 by default, and ask the hardware for a suggestion, but only treat it as a suggestion for between 1x to 2x. [14:21] More than 2x might make it hard to correct bad settings. [14:21] That's the worst possibility, which Jason was getting to. That the user can't see to change something that is wrong. [14:22] yassas! [14:22] i think that sounds reasonable, if the DPI default setting is outside scaling 1.0 -> 2.0 (96 -> 1962 DPI), default 1.0 [14:24] scaling-factor (int), text-scaling-factor(float) [14:24] Can we get the EDID scaling reported back to us somehow, so we can construct a blacklist and make better default scaling decisions in 14.10 ? [14:26] jasoncwarner: yep, value is stored in percent: 100 = scale factor 1 [14:26] (in LibreOffice) [14:26] jasoncwarner: Chromium has resources for whole number, but we do math to get exact, right now. [14:27] really, a rounding up from 1.5 -> 2.0 is pretty noticeable. Marcos idea was if we are equal to or larger then 2.0, then switch to 2.0 [14:28] bschaefer: what does that look like? 1-1.5 = 1.0? 1.5-1.999 = 2.0? 2.0 - x.0 = 2.0? [14:28] why only scale Unity to whole numbers? [14:28] *not [14:28] jasoncwarner: there's n<1, too [14:28] by default anyway [14:29] qengho: true, was trying to type quickly ;) [14:29] jasoncwarner, well from bregmas testing, 2.0 is to large for his setting [14:29] n > 1.5 goes to scaling factor 2. [14:29] rounding could work, i need a hi dpi screen :) [14:30] I think it requires some nice testing [14:31] bregma, correct, (once half the window is over to the new monitor) [14:31] qengho, Sweetshark: is libreoffice/chromium scaling by windows? [14:31] like what if I've 2 screens, with different dpi settings and one lo/chromium on each [14:31] seb128: nope [14:32] the biggest problem with the setting scaling-factor is its global [14:32] can they have different scaling to match the respective screens? [14:32] if say you have a HiDPI monitor attached (needing 2.0 scaling) and you have a laptop screen that only uses 1.0 scaling [14:32] seb128: and if we update the value in the config you would need restart LibreOffice for it to notice. [14:32] if we set scaling-factor to 2.0, your laptop screen will be to large [14:33] seb128: though LibreOffice obviously can change scaling on the fly -- but it doesnt watch the config. [14:33] bschaefer, right [14:33] seb128: Chromium only knows of a single display, which I think is the display owning the top left corner. [14:33] so only set it to 1 or 2 [14:33] of the Cr window. [14:33] seb128, its a very tricky situation global vs per screen support [14:34] bschaefer, right [14:34] the issue is that unity has that nice "value by screen" [14:34] but nothing else does :) [14:34] but we can't even do heuristics for GTK [14:34] like if both screens configs don't match [14:34] which one do we pick?. [14:34] we screw up one screen either way [14:35] conservative could work, if and only if all monitors scaling is 1.5 (>) then we can switch [14:35] also what about when the user changes the setting in LibreOffice itself and overwrits the value? We _could_ consider hiding that in the LibreOffice options UI, but that would kinda suck for non-unity DEs ... [14:35] seb128, yeeah :( [14:35] Sweetshark: non-unity DEs? [14:35] Sweetshark: ;) [14:35] jasoncwarner: trololol [14:35] Sweetshark: :) [14:37] that was the overall idea, it would be nice if we could control all the settings possible, but I think its up to the user to change the settings how they want [14:37] (without fighting the DE) [14:39] seb128: so for LibreOffice to handle changing the scaling factor dynamically, writing the config isnt really good as it only will be noticed on startup. Instead exporting that setting via IPC (dbus?) would likely be better. [14:40] Yeah, the problem is unity does Per-Monitor, Gtk does a global scaling [14:41] I think qml does the right thing [14:41] so things are going to work for new apps [14:41] but that's it [14:42] unity/qml are right and the futur for us [14:42] (though its somewhat pointless as LibreOffice has one scale factor for all windows. So if you have a Calc window on a HIDPI screen and an Writer windows on a non-HiDPI one, they have to have the same scaling). [14:42] it's a bit unfortunate most of our apps world isn't there though [14:42] If we play it conservative, we use the lowest scaling value to auto set the global settings [14:43] seb128, yeah, but we also dont want to restrict unity based on toolkits :( [14:43] For scaling defaults, I'd like to make sure we design something where single apps don't need to know about our decided policy. I don't want to carry diffs in packages, "if gsettings is 0, if release is trusty, set to 96dpi". [14:44] How about having a opt-in checkbox ('set font scaling ~everywhere') somewhere? [14:44] Sweetshark: I think that is what we just said [14:45] jasoncwarner: yeah, typing bandwidth and latency ;) [14:45] text-scaling-factor is nice, since its a float (but it only scales text!) [14:45] text-scaling-factor does not mix very well with scaling-factor [14:45] Sweetshark: it's like being in the future....I look so smart! [14:47] and if they check both, use the lowest setting? [14:48] or, only 1 box can be checked? [14:48] that works [14:51] bregma, what about detecting a dpi default? [14:52] defaulting to one is easier for us :) === qengho_ is now known as CardinalFang === ChanServ changed the topic of #ubuntu-uds-client-1 to: Track: Client | Unity8 UI Touch/Desktop shell | Url: http://summit.ubuntu.com/uds-1403/meeting/22222/client-1404-unity-ui-shell/ [14:56] jasoncwarner: we should play poker like that. === CardinalFang is now known as qengho [14:56] Sweetshark: that would be nice, eh? [14:59] bregma, Saviq, others interested by the coming unity8 hangout, https://plus.google.com/hangouts/_/hoaevent/AP36tYegL3bfdWtjR2B7mDJ-xUWL-ZQF1AnSj1DBGGAGKXQpgkEuLQ?authuser=0 [15:01] nobody for that session? [15:01] isn't this Saviq's session? [15:02] bregma, sounds like it's yours :) [15:02] bregma, he said it's yours [15:03] not mine, blueprint is created by kgunn [15:04] ok, seems that session has already workitems and stuff and nobody has an agenda [15:05] is anyone having an agenda/things to discuss or should we just dismiss it? [15:05] sounds like dismissal time [15:05] done [15:05] thanks everyone [15:16] is this session not happening or something? [15:17] also wondering that ^ [15:18] ok, seems that session has already workitems and stuff and nobody has an agenda [15:18] is anyone having an agenda/things to discuss or should we just dismiss it? [15:18] sounds like dismissal time [15:18] done [15:18] let me delete it from the schedule [15:18] seb128: ah thanks [15:45] __lucio__, hey, could you put a summary of the push notification session/outcome on http://pad.ubuntu.com/nRKeBvia0H? (that's for the vUDS wrapup summary) [15:45] bfiller, ^ same for touch apps on desktop would be nice, if you can ;-) [15:46] <__lucio__> seb128, whats the dealine for that? (got some meetings now) [15:46] __lucio__, wrapup is at 7pm utc, so you have like 3 hours [15:46] <__lucio__> ack [15:46] thanks === ChanServ changed the topic of #ubuntu-uds-client-1 to: Currently no events are active in this room - http://summit.ubuntu.com/uds-1403/client-1/ - http://irclogs.ubuntu.com/2014/03/13/%23ubuntu-uds-client-1.html === test is now known as Guest25034 === iBelieve_ is now known as iBelieve === iBelieve_ is now known as iBelieve