=== maclin1 is now known as maclin | ||
=== maclin1 is now known as maclin | ||
* guest42315 sudo snappy install coffee | 07:05 | |
Saviq | tsdgeos, hey, got a sec? need a sounding board | 08:23 |
---|---|---|
tsdgeos | Saviq: sure | 08:24 |
Saviq | tsdgeos, last issue with autopkgtests | 08:24 |
Saviq | tsdgeos, paths.h is generated on build, but when testing archive/ppa-built packages, the non-installed paths there are not the same as what you get in the testing environment | 08:25 |
Saviq | e.g. when the packages were built | 08:25 |
Saviq | it was file:///build/unity8-aDQvXH/ | 08:26 |
Saviq | but when the tests are ran | 08:26 |
Saviq | it's file:///tmp/adt-run.jl4VOK/build.64t/ for example | 08:26 |
Saviq | basically, our binaries get some random paths baked in | 08:26 |
tsdgeos | right | 08:27 |
tsdgeos | hmmm | 08:28 |
tsdgeos | no, not right | 08:28 |
tsdgeos | why does this happen? | 08:28 |
tsdgeos | does adt copy the things around? | 08:28 |
Saviq | this actually makes ApplicationWindow test fail under adt-run *if* you don't build the packages at the same time (e.g. what you should do - take the packages as built in the PPA, not build them locally) | 08:29 |
Saviq | tsdgeos, PPA builds the packages, the build path on the builders is /build/foo | 08:29 |
Saviq | tsdgeos, then, adt takes those packages as they are, builds just the missing mocks and runs the tests, but the build path is different - /tmp/adt-foo | 08:30 |
tsdgeos | ah so there's two builds | 08:30 |
Saviq | in effect, yes | 08:30 |
Saviq | but even if there were not | 08:31 |
Saviq | the path baked into paths.h on the builders | 08:31 |
Saviq | are invalid as soon as the build is done | 08:31 |
Saviq | at that point paths.h is only useful when running installed | 08:32 |
Saviq | == running unity8 itself | 08:32 |
Saviq | so, short term I could put the graphics from qml/graphics/applicationIcons and qml/Dash/graphics/{phone,tablet} in qrc | 08:34 |
Saviq | they don't belong where they are any more anyway | 08:34 |
Saviq | but that still means paths.h is useless non-installed | 08:35 |
Saviq | (btw this worked for adt-run when you did build packages within adt-run because the build dir was reused) | 08:35 |
Saviq | the real fix would be to runtime-detect our build root | 08:36 |
tsdgeos | yep | 08:38 |
tsdgeos | make it be XDG_SOMETHING | 08:38 |
Saviq | wdym | 08:41 |
Saviq | ? | 08:41 |
tsdgeos | so there's all these XDG_ variables | 08:41 |
tsdgeos | we can just assume our stuff is asumed in XDG_something/path | 08:42 |
tsdgeos | and then set XDG_ accordingly for the tests? | 08:42 |
tsdgeos | not sure how viable that would be | 08:42 |
tsdgeos | how did you plan to detect the root? | 08:42 |
tsdgeos | pwd or something? | 08:43 |
Saviq | we'd need to up from pwd until reached build root | 08:43 |
Saviq | +go | 08:43 |
Saviq | but that sounds nasty | 08:44 |
tsdgeos | would it even work? | 08:44 |
tsdgeos | i mean in paths.h we have | 08:44 |
tsdgeos | return QString("@CMAKE_BINARY_DIR@/po/locale"); | 08:44 |
tsdgeos | and | 08:44 |
tsdgeos | return QString("@CMAKE_SOURCE_DIR@/qml/"); | 08:44 |
Saviq | right | 08:45 |
tsdgeos | how do we find those if we have two build roots? | 08:45 |
tsdgeos | is the source even around? | 08:45 |
Saviq | yeah, source is always around | 08:45 |
Saviq | but if we start requiring some env vars to be set, even more meh | 08:46 |
Saviq | we're not really interested about "previous" CMAKE_BINARY_DIR | 08:47 |
Saviq | because it's invalid | 08:47 |
Saviq | we only care about current paths | 08:47 |
Saviq | I can't shake the feeling that we're either doing something wrong or that someone must've solved this already... | 08:48 |
Saviq | obviously the fact that QML makes relative paths nice and easy doesn't help | 08:48 |
Saviq | because that only works until you start passing paths to C++ | 08:49 |
* Saviq tempted to start moving all that stuff to qrc, only reliable way | 08:49 | |
Saviq | not that it really solves the issue | 08:49 |
Saviq | on that note, can you use qrc in qml-only files? how do you tell Qt where the resources are? | 08:50 |
Saviq | mzanetti, ideas? (read up to ~10:23 ↑) | 08:56 |
* mzanetti reading | 08:57 | |
mzanetti | +1 for runtime detection... /me never liked the paths.h | 09:01 |
mzanetti | now... I don't have the solution obviously | 09:01 |
* mzanetti thinks | 09:01 | |
Saviq | the only thing I could think of is putting a special file in project root and go up from $CWD to find it | 09:03 |
Saviq | but that seems meh | 09:03 |
Saviq | we could have paths.h as a library and LD_PRELOAD it... that one seems overkill | 09:06 |
mzanetti | Saviq, well, putting a special file is meh indeed, but why not trying to find the actual files? | 09:06 |
Saviq | mzanetti, as meh as a special file I'd say | 09:07 |
mzanetti | slightly less meh | 09:07 |
mzanetti | not super pretty still, yes | 09:07 |
Saviq | with the added problem that you might want to move or rename the non-special file | 09:07 |
mzanetti | yeah well... the test would catch that pretty fast ;) | 09:07 |
Saviq | sure, but I'd rather be explicit about it | 09:08 |
Saviq | I just feel like this must be a solved problem | 09:09 |
Saviq | can't believe we're the first ones to ancounter it | 09:09 |
Saviq | *encounter | 09:09 |
mzanetti | Saviq, doesn't dpkg export some build_root var or something? | 09:11 |
Saviq | mzanetti, even if, only useful if building a package | 09:12 |
mzanetti | isn't that what we're talking about? | 09:12 |
Saviq | mzanetti, the problem we're facing is running tests | 09:12 |
mzanetti | yes, afaiu it's still dpkg who is building the last mocks and runs the tests, no? | 09:13 |
Saviq | no | 09:13 |
mzanetti | hmm, ok | 09:13 |
Saviq | adt-run | 09:13 |
Saviq | but even so | 09:13 |
mzanetti | I thought that's dpkg-magic | 09:13 |
mzanetti | so, doesn't adt-run export some information about the env? | 09:13 |
Saviq | don't think we should rely on that | 09:14 |
Saviq | because you don't want to have to export this manually when running outside of adt-run | 09:14 |
tsdgeos | Saviq: well, default to what we have unless the env var is exported and then use that? | 09:16 |
Saviq | tsdgeos, yeah, I was thinking that, but still feels like a band-aid | 09:16 |
tsdgeos | seems good to me tbh | 09:16 |
tsdgeos | all these "let's detect where i am" are like this | 09:17 |
Saviq | trueth | 09:17 |
tsdgeos | use some env vars, if not use some paths i got while compiling, if not just try /usr and if not just bail out and go crazy | 09:17 |
Saviq | aargh why does QtC not open CMakeLists.txt files any more but tries to build the project ;( | 09:25 |
Saviq | oh ok, need to open from project pane not filesystem | 09:25 |
Saviq | mzanetti, tsdgeos, opinion: should we rely on paths coming from paths.h to be /-terminated, or assume they're not? | 09:43 |
tsdgeos | Saviq: if we're going to accept envvar, assume not | 09:43 |
tsdgeos | there's nothing more annoying that stuff failing because you have to add / to an envvar | 09:43 |
Saviq | +1 | 09:44 |
mzanetti | wfm | 09:44 |
mzanetti | I usually add trailing slashes, but mostly not rely on them being there | 09:44 |
mzanetti | which causes my code to often have // in it... but at least it works | 09:44 |
Saviq | yeah, so assume not it is | 09:45 |
Saviq | oh wow, QtC remembered changes after crashing, kudos ;D | 09:48 |
=== alan_g is now known as alan_g|lunch | ||
=== alan_g|lunch is now known as alan_g | ||
mzanetti | mterry, can't repro this: https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1498486 | 13:45 |
ubot5 | Ubuntu bug 1498486 in unity8 (Ubuntu) "Clicking on a 'folded' launcher icon should always result in a launchable icon" [Undecided,New] | 13:45 |
mzanetti | mterry, for me the clicked icon ends up being completely unfolded | 13:45 |
mzanetti | on krillin that is | 13:45 |
mterry | mzanetti, huh. I only tested on mako, assumed they had same size in landscape | 13:46 |
mterry | mzanetti, do you have 3 icons not shown in the direction you are testing? | 13:46 |
mzanetti | however, I have seen issues that ListView.flick() doesn't always flick with the same strength | 13:46 |
mzanetti | mterry, yes, I have like 20 icons in the launcher | 13:46 |
mterry | humph | 13:46 |
mzanetti | and trying somewhere in the middle | 13:46 |
mterry | mzanetti, maybe it's mako-specific (like it matters on a pixel level, instead of a grid unit level) | 13:47 |
mzanetti | probably | 13:47 |
* mzanetti searches for a mako device | 13:47 | |
mterry | will edit bug to not mention krillin | 13:47 |
mzanetti | mterry, hah... indeed. it flicks like twice as much on mako | 13:48 |
mzanetti | cycles through the whole list :D | 13:49 |
mterry | oh interesting it's so different | 13:49 |
mterry | not the whole list. just like 3 icons...? | 13:49 |
mterry | For me, the nearest-folded icon will bounce between top and bottom when I click on it | 13:50 |
mterry | But always be folded | 13:50 |
=== pat_ is now known as Guest18418 | ||
=== dandrader_ is now known as dandrader|afk | ||
=== dandrader|afk is now known as dandrader | ||
=== Guest18418 is now known as pmgowan | ||
=== alan_g is now known as alan_g|EOD | ||
dandrader | Saviq, do you know where things like Qt5Gui_PRIVATE_INCLUDE_DIRS come from? | 17:02 |
dandrader | Saviq, ie, what file defines it | 17:03 |
Saviq | dandrader, dpkg -L qtbase5-dev | grep cmake | 17:03 |
dandrader | Saviq, thanks | 17:04 |
Saviq | grep PRIVATE -r /usr/lib/x86_64-linux-gnu/cmake/Qt5Gui | 17:04 |
=== 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!