Laney | hi, wondering if anyone is looking at mir blocked in xenial-proposed? | 09:57 |
---|---|---|
Laney | (liburcu hangs on ppc64el) | 10:01 |
alan_g | Laney: I'm not aware of it. But I don't see how Mir can impact liburcu? | 10:13 |
Laney | it's one of Mir's dependencies | 10:14 |
seb128 | alan_g, it doesn't impact it directly, it's the other way around | 10:14 |
=== chihchun_afk is now known as chihchun | ||
seb128 | MIR is blocked because one of the libs it's using needs to be fixed | 10:15 |
alan_g | seb128: ah, at least that makes sense. | 10:15 |
alan_g | So landing wouldn't break anything that isn't already broken? | 10:19 |
seb128 | that's not how it works | 10:21 |
seb128 | that library got an update | 10:21 |
seb128 | and the new mir built/got linked to that new version | 10:21 |
seb128 | so landing the mir update means landing the library update with it | 10:21 |
seb128 | so if the lib has a regression mir might have a regression as well | 10:21 |
seb128 | even if it's not mir's fault | 10:22 |
seb128 | or said differently the library you are using needs to be cleared out or it's going to block mir | 10:22 |
Laney | And blocking mir means it blocks anything that depends on mir (in this case GTK) | 10:22 |
alan_g | alf_: is it AlbertA trying to release Mir to xenial? | 10:26 |
alan_g | seb128: I've no idea how to "clear out liburcu" - is this urgent? Or can it wait for the USA to wake up? | 10:27 |
seb128 | alan_g, it can wait for them to wake up | 10:27 |
seb128 | alan_g, the issue with liburcu is that the build/tests hang on ppc64el | 10:28 |
seb128 | so somebody needs to debug why | 10:28 |
seb128 | (likely an upstream issue, it's also happening in Debian) | 10:28 |
alf_ | alan_g: Yes, 0.17.1 | 10:28 |
alan_g | seb128: can't we just use the old version? | 10:29 |
alf_ | seb128: problem with ppc64el is that we have no reasonable way to replicate the problems locally | 10:30 |
alan_g | Who can debug ppc64el? | 10:30 |
seb128 | alan_g, we would need to delete the update and rebuild mir | 10:32 |
seb128 | but that's not really constructive, we are going to need to update that lib one day | 10:32 |
seb128 | so better to fix the issue and move forward | 10:32 |
alan_g | seb128: the first problem is the lack of any ppc64el kit on which to investigate. | 10:35 |
seb128 | Laney guessed that would be an issue :-/ | 10:35 |
Laney | there's the bos01 cloud | 10:35 |
Laney | I would hope that it's possible to get access there | 10:36 |
Laney | although... I just tested on an instance there (abusing my access through autopkgtest) and it actually built :( | 10:37 |
Laney | one second | 10:37 |
Laney | https://bugs.launchpad.net/ubuntu/+source/liburcu/+bug/1511458 | 10:38 |
ubot5 | Ubuntu bug 1511458 in linux (Ubuntu Vivid) "liburcu fails to build on ppc64el" [High,Confirmed] | 10:38 |
Laney | seems it might be a cloud/kernel bug | 10:40 |
* Laney asked | 10:42 | |
=== tvoss is now known as tvoss|lunch | ||
=== dandrader is now known as dandrader|bbl | ||
=== tvoss|lunch is now known as tvoss | ||
=== chihchun is now known as chihchun_afk | ||
=== chihchun_afk is now known as chihchun | ||
pandatrone | what happened to mir?last update is 2015-11-13 | 14:09 |
pandatrone | 5 days ago | 14:09 |
alan_g | pandatrone: MP landings have been backing up because of trouble in CI. Hoping that is now resolved. | 14:14 |
pandatrone | alan_g, phew :D thanks! | 14:14 |
alan_g | were you waiting on something? | 14:17 |
pandatrone | alan_g, nope. i just like to read the new mir updates :D | 14:20 |
* alan_g resolves to write interesting commit comments. | 14:23 | |
pandatrone | :)) | 14:23 |
anpok_ | alan_g: evemu-record worked for me when I dont send it to background | 14:54 |
anpok_ | but it just terminates otherwise | 14:54 |
=== dandrader|bbl is now known as dandrader | ||
alan_g | OK, you need both devices at the same time? | 14:55 |
anpok_ | alan_g: hm I think so yes.. | 14:56 |
anpok_ | but there is also a slight chance that this will be fixed by: https://code.launchpad.net/~andreas-pokorny/mir/use-one-cursor-position-per-seat/+merge/277843 | 14:57 |
anpok_ | but I cannot remember the details | 14:57 |
tsdgeos | guys, i need this https://code.launchpad.net/~aacid/mir/eintr_dispatch_loop/+merge/277847 otherwise my unity8 aborts, can anyone with more knowledge of the code have a look and decide if it makes any sense? | 15:00 |
alan_g | anpok_: sent. I'll try your branch later | 15:06 |
alan_g | tsdgeos: looks weird to me too. I think RAOF wrote most of that code | 15:07 |
alan_g | Laney: is the problem with liburcu still needing help from Mir? | 15:10 |
anpok_ | alan_g: hm ok but you are pressing alt while moving the trackpoint? | 15:26 |
alan_g | yes | 15:26 |
Laney | alan_g: not right now, hopefully lp will fix it | 15:58 |
Laney | https://portal.admin.canonical.com/85976 for those who can see that | 15:58 |
alan_g | anpok_: which set of your branches should I merge to test laptop? | 16:25 |
alan_g | anpok_: reapplied lp:~andreas-pokorny/mir/load-all-supported-input-platforms and lp:~andreas-pokorny/mir/use-one-cursor-position-per-seat - I still have a problem. The trackpoint controls the cursor unless the middle button is pressed. Alt+left button drags the triangle, Alt on its own or Alt+right just move the cursor. Just pressing the middle button (with or without Alt) and the cursor only responds to the touchpad ( | 16:46 |
alan_g | but that has never worked with the "mouse" buttons). | 16:46 |
anpok_ | oh | 16:58 |
anpok_ | we do not unify the mouse buttons | 16:58 |
anpok_ | for the input stack the trackpoint and the mouse buttons belong to one device and the touchpad is separate.. | 16:59 |
anpok_ | that means touchpad movement + trackpoint buttons are two separate streams of events hence.. if someone checks for button state on motion events all will be released.. | 17:00 |
anpok_ | but thats simple to combine .. just like we did with modifier.. | 17:00 |
alan_g | I just don't want to break something that is already working (trackpoint+middle button) | 17:01 |
=== dandrader is now known as dandrader|afk | ||
alan_g | I thought it might be diagnostic that the left and right buttons do work | 17:02 |
anpok_ | hmm yes that is strange | 17:08 |
alan_g | is something misidentifying it as a *two* button mouse? | 17:13 |
anpok_ | oh I have to make sure the server sees the first events too otherwise it will miss the middle button down press | 17:17 |
anpok_ | yes I can reproduce .. it has nothing to do with any of my theories. | 17:18 |
alan_g | anpok_: reproduction is the first step towards diagnosis | 17:21 |
anpok_ | hum.. and I believe we dont want to have a separate button state between pointing devices | 17:29 |
=== chihchun is now known as chihchun_afk | ||
=== dandrader|afk is now known as dandrader | ||
=== chihchun_afk is now known as chihchun | ||
=== chihchun is now known as chihchun_afk | ||
=== dandrader is now known as dandrader|afk | ||
=== dandrader|afk is now known as dandrader |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!