Merge lp:~andreas-pokorny/mir/load-all-supported-input-platforms into lp:mir
Status: | Merged |
---|---|
Approved by: | Andreas Pokorny on 2015-11-25 |
Approved revision: | 3016 |
Merged at revision: | 3126 |
Proposed branch: | lp:~andreas-pokorny/mir/load-all-supported-input-platforms |
Merge into: | lp:mir |
Prerequisite: | lp:~andreas-pokorny/mir/combine-modifier-state-accross-devices |
Diff against target: |
774 lines (+366/-77) 22 files modified
benchmarks/frame-uniformity/touch_producing_server.cpp (+0/-1) include/platform/mir/input/platform.h (+2/-2) include/test/mir_test_framework/executable_path.h (+1/-0) src/include/server/mir/default_server_configuration.h (+1/-11) src/include/server/mir/input/input_probe.h (+48/-0) src/platform/graphics/platform_probe.cpp (+5/-1) src/platforms/evdev/platform_factory.cpp (+1/-1) src/platforms/mesa/server/x11/input/input.cpp (+1/-1) src/server/input/CMakeLists.txt (+1/-0) src/server/input/default_configuration.cpp (+16/-39) src/server/input/input_probe.cpp (+104/-0) src/server/symbols.map (+1/-1) tests/mir_test_framework/CMakeLists.txt (+1/-0) tests/mir_test_framework/executable_path.cpp (+10/-0) tests/mir_test_framework/fake_input_device_impl.cpp (+1/-1) tests/mir_test_framework/fake_input_server_configuration.cpp (+0/-1) tests/mir_test_framework/input_testing_server_options.cpp (+0/-1) tests/mir_test_framework/stub_input.cpp (+2/-2) tests/mir_test_framework/stub_input_platform.cpp (+1/-1) tests/umock-acceptance-tests/test_libinput.cpp (+8/-14) tests/unit-tests/input/CMakeLists.txt (+1/-0) tests/unit-tests/input/test_input_platform_probing.cpp (+161/-0) |
To merge this branch: | bzr merge lp:~andreas-pokorny/mir/load-all-supported-input-platforms |
Related bugs: |
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
PS Jenkins bot (community) | continuous-integration | 2015-10-14 | Approve on 2015-11-25 |
Cemil Azizoglu (community) | Approve on 2015-11-20 | ||
Alan Griffiths | 2015-10-14 | Abstain on 2015-11-20 | |
Daniel van Vugt | 2015-10-14 | Needs Fixing on 2015-10-28 | |
Chris Halse Rogers | 2015-10-14 | Approve on 2015-10-19 | |
Review via email:
|
This proposal supersedes a proposal from 2015-10-05.
Commit message
Add automatic searching and probing for input platforms
With this specifying the input platform via --platform-
Description of the change
The most notable change of this MP will make the libinput and x11 platform work without additional command line parameter.
This MP adds probing for input platforms. mirserver will now search inside the default platform-path for loadable modules that fulfil the input platform ABI. Then only those modules that claim to be better than 'dummy' will be used. Hence input-stub.so will not be loaded by default, but requires loading through the platform-input-lib parameter.
Next step: remove input reader and event hub..
WRT to testing this MP - or the evdev input platform in general: pick silo-56 for phone images. This will still have problems with the clock (which is fixed by lp:~andreas-pokorny/mir/use-realtime-clock-for-input-events). Xenial already has the right libinput package.
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:2993
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
FAILURE: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:2993
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
FAILURE: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
Chris Halse Rogers (raof) wrote : | # |
+ mir::log_
Could you also log driver version please?
+ return std::move(
Pessimisation: this breaks return value optimisation. When the Wily clang builds work again this will be an error.
+ EXPECT_
+ EXPECT_
+ EXPECT_
Seems like you could replace OfType with OfPointerType and then use
EXPECT_
Which I think is both more natural and also is a slightly stronger test.
=======
Non-review comment:
+ mtf::UdevEnviro
Heads up: I think that doing this will provide negative value. Also, und→and :)
Andreas Pokorny (andreas-pokorny) wrote : | # |
Fixed all of the above.
> =======
> Non-review comment:
> + mtf::UdevEnviro
> mock_libudev when the new input reading stack is the only in use.
>
> Heads up: I think that doing this will provide negative value. Also, und→and
> :)
I want to simulate the cases no devices found, devices found and some or all can be opened, devices found and none can be opened. I cannot do that with umockdev...
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:2995
http://
Executed test runs:
FAILURE: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
Chris Halse Rogers (raof) wrote : | # |
On Thu, Oct 8, 2015 at 5:56 PM, Andreas Pokorny
<email address hidden> wrote:
> Fixed all of the above.
>
>> =======
>> Non-review comment:
>> + mtf::UdevEnviro
>> und
>> mock_libudev when the new input reading stack is the only in use.
>>
>> Heads up: I think that doing this will provide negative value.
>> Also, und→and
>> :)
>
> I want to simulate the cases no devices found, devices found and some
> or all can be opened, devices found and none can be opened. I cannot
> do that with umockdev...
You can do all except the last, and the last would be a simple patch to
umockdev.
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:2997
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
FAILURE: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
Alan Griffiths (alan-griffiths) wrote : | # |
Looks like infrastructure glitch:
inflating: package_
error: closed
[timestamp] Start PPAs and local Archive : 2015-10-
rm: cannot remove 'ppa_sources': No such file or directory
~~~~
(Non blocking) looks like we need to bump the server ABI
Alan Griffiths (alan-griffiths) wrote : | # |
"The most notable change of this MP will make the libinput and x11 platform work without additional command line parameter."
Am I expecting too much?
$ sudo DISPLAY= bin/mir_demo_server --vt 1 --arw-file&
$ bin/mir_demo_server --file mir-socket --host-socket /tmp/mir_socket
[Switch to vt 1]
[Switch to vt 7]
ERROR: /home/alan/
Dynamic exception type: boost::
std::exception:
13, "Permission denied"
It looks as though the nested server is selecting the wrong drivers. (It should "agree" with the host.)
NB I can workaround like this:
$ DISPLAY= bin/mir_demo_server --file mir-socket --host-socket /tmp/mir_socket --launch bin/mir_
Alan Griffiths (alan-griffiths) wrote : | # |
OK, I don't think the auto-configuration is quite there yet. But this is a step forwards.
I guess there will be an ABI break needed for 0.18.
Andreas Pokorny (andreas-pokorny) wrote : | # |
Oh .. I wasnt aware there is a mismatch between the x11 graphics and input probing.
Soooo.. a host-socket setting should win over a DISPLAY env, shouidnt it?
That should be a simple fix in graphics.cpp.
PS Jenkins bot (ps-jenkins) wrote : | # |
PASSED: Continuous integration, rev:2997
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
Andreas Pokorny (andreas-pokorny) wrote : | # |
> Oh .. I wasnt aware there is a mismatch between the x11 graphics and input
> probing.
> Soooo.. a host-socket setting should win over a DISPLAY env, shouidnt it?
>
> That should be a simple fix in graphics.cpp.
Nested.. not as simple as I thought..
Daniel van Vugt (vanvugt) wrote : | # |
Landed and then reverted:
* Modifiers don't work any more (can't move/resize windows using Alt)
* Default mouse speed too high
* A little too many log messages come from this.
Only the first issue is a blocker.
Andreas Pokorny (andreas-pokorny) wrote : | # |
> Landed and then reverted:
> * Modifiers don't work any more (can't move/resize windows using Alt)
Yes modifiers are per device right now .. which isnt what we want within the dispatcher. Working on a proper abstraction/fix for that.
> * Default mouse speed too high
Yay! So mouse acceleration is working. Configuration branches are on the way..
> * A little too many log messages come from this.
>
> Only the first issue is a blocker.
Andreas Pokorny (andreas-pokorny) wrote : | # |
> Landed and then reverted:
> * Modifiers don't work any more (can't move/resize windows using Alt)
> * Default mouse speed too high
> * A little too many log messages come from this.
>
> Only the first issue is a blocker.
This should fix the issue..
https:/
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:3000
http://
Executed test runs:
SUCCESS: http://
FAILURE: http://
FAILURE: http://
FAILURE: http://
FAILURE: http://
FAILURE: http://
FAILURE: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
Alan Griffiths (alan-griffiths) wrote : | # |
Nits:
+#include "mir/module_
#include <memory>
#include <string>
+#include <vector>
The other changes to src/include/
~~~~
- virtual std::shared_
virtual std::shared_
virtual std::shared_
virtual std::shared_
@@ -384,8 +385,7 @@
CachedPtr<
CachedPtr<
CachedPtr<
- CachedPtr<
- CachedPtr<
+ CachedPtr<
You can also lose the declaration of Platform on line 135.
~~~~~
Probably not something to address in this MP but...
+#ifdef MIR_BUILD_
+ ON_CALL(mock_x11, XOpenDisplay(
+#endif
Should the test setup really depend on the *build* platform?
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:3001
http://
Executed test runs:
SUCCESS: http://
FAILURE: http://
FAILURE: http://
None: http://
FAILURE: http://
FAILURE: http://
FAILURE: http://
None: http://
Click here to trigger a rebuild:
http://
Andreas Pokorny (andreas-pokorny) wrote : | # |
This is currently blocked by:
https:/
Andreas Pokorny (andreas-pokorny) wrote : | # |
^ No longer blocking.. brandon made a similar MP and that landed today.
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:3002
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
None: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
FAILURE: http://
None: http://
Click here to trigger a rebuild:
http://
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:3003
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
FAILURE: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:3005
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
Andreas Pokorny (andreas-pokorny) wrote : | # |
^
demo server and acceptances tests crash...
Daniel van Vugt (vanvugt) wrote : | # |
OK, I've invested time in testing and lots needs fixing unfortunately. At least three are show-stoppers...
(1) Mouse now moves too fast. Even my slowest mouse.
(2) "evdev-input" log messages on startup report nothing useful. Can we remove or change them to report actual device descriptions?
(3) Input resampling has been lost (can only verify on desktop due to #5). It was introduced on 2014-08-29 but we only just fixed Unity8 this week to utilize it and achieve halved latency. It's taken 14 months to come to fruition this week, so please try to avoid regressing on it.
Test: env MIR_CLIENT_
Verify only one square painted per frame.
(4) Input events are strangely erratic and no longer smooth (even when not resampled), compared to lp:mir. Compare 'mir_demo_
(5) Servers crash on startup on mako (vivid rc-proposed):
[1446022093.989552] <WARNING> mircommon: Failed to load module: bin/../
[1446022093.989948] mircommon: Loading module: bin/../
[1446022094.000600] mircommon: Loading module: bin/../
ERROR: /home/dan/
Dynamic exception type: N5boost16except
std::exception:
Daniel van Vugt (vanvugt) wrote : | # |
(5) Fixed by installing package 'libinput10' but ...
(6) mir_demo_
Daniel van Vugt (vanvugt) wrote : | # |
(7) Multi-touch gestures no longer work (on mako). Try moving a window in mir_proving_server using three fingers and it mostly does nothing, or has a mind of its own. If you run 'mir_demo_
Daniel van Vugt (vanvugt) wrote : | # |
Although (1) sounds unimportant and not like a blocker, if you fix (1) and remove the default acceleration that libinput introduces, then that might actually fix (4) as well.
Alan Griffiths (alan-griffiths) wrote : | # |
My issue above fixed. Looks like enough issues remain to be dealt with for me to wait before looking again.
Andreas Pokorny (andreas-pokorny) wrote : | # |
(1) This is the default setting of libinput - 1.1.0 will change that i.e use differnt profiles depending on the mouse or device type. I found the default setting just good..
(2)
(3) There is a known problem with libinput-0.21 in the overlay ppas. Patches are w
(4)
(5) better than hanging the system
(6) they are reported like that from the kernel.. in general to small.. but d
Andreas Pokorny (andreas-pokorny) wrote : | # |
oops hit reply to soon..
(3) (4) require a libinput patch that has not yet landed in archive or those ppas. This might be the reason for (7) too.
(6) proper touch screen drivers should report resolution values to allow scaling values to actual units. All touch screen drives that we use only report a resolution of zero. That invalid setting is handled inside libinput and it returns the plain value reported by the kernel. Those values as you have noticed are too small. To come up with proper values we need to have a device dependent scaling. I dont see
Andreas Pokorny (andreas-pokorny) wrote : | # |
.. any other client than finger paint use those values at the moment. I do hope that since they are now exposed via libinput that the other display servers and toolkits start to use those. Within the Qt event feeder they are mapped to a rectangular region.. which isnt really accurate either.
Andreas Pokorny (andreas-pokorny) wrote : | # |
just checked mako again: (3) (4) and (7) is the same cause - and needs the current libinput patch-set as in:
https:/
Daniel van Vugt (vanvugt) wrote : | # |
(3) and (4) were discovered with a mouse on desktop. So any touchscreen fixes will not help (3) and (4). It appears likely that fixing (1), while not a major issue itself, will probably fix (3) and (4). Because (1) is the only reason I can think of for libinput to tinker with the mouse event stream.
Andreas Pokorny (andreas-pokorny) wrote : | # |
(2) I just improved logs with lp:~andreas-pokorny/mir/better-input-device-related-logs
We should really consider having something like a severity filter..
Ok I misunderstood (3) and (4) so what I just tried on lp:mir was:
A: Resampling with input-evdev:
export MIR_CLIENT_
mir_demo_server --platform-
B: Resampling with android input
export MIR_CLIENT_
mir_demo_server --launch-
C: No resampling with input-evdev
export MIR_CLIENT_
mir_demo_server --platform-
D: No resampling with android input
export MIR_CLIENT_
mir_demo_server --launch-
using a cheap USB MS mouse on top of libinput-1.1.0 with my patches
A and B look nearly identical - mouse moves slightly faster than with android input but there are ugly gaps/jumps .. in both cases every 5 or 10 seconds in the trace drawn by fingerpaint.
C and D look nearly identical - mouse moves slightly faster than with android input but there in both cases the trace displayed is smooth and sticks close to the cursor. Additionally there is no odd jump in both the data sets. As opposed to the resampled cases the plotted squares have a similar distance and give a good indication of mouse movement.
I repeated A and B a few times and I was able to reproduce the odd jumps for A and B sometimes. It seems that the resampling is very sensitive to scheduling. My conclusion from that:
we should make sure that we only use resampling when we have a horrible input device (driver) or when the client is unable to process the user input fast enough. Additionally we should spend the next 14 month with verifying that the algorithm yields proper results.
PS: I did mess with android input in the past and found the bug that disabled the pointer acceleration:
https:/
Not sure if it still works..
Daniel van Vugt (vanvugt) wrote : | # |
AFAIK, resampling uses the original event timestamps which in android-input are accurate because they come directly from the kernel. So if those timestamps are being faked in libinput then that might explain the irregularity and sensitivity to scheduling. Better check that libinput (and Mir itself) is not inserting new timestamps. Our input reports also rely on those timestamps being the original event time from the kernel.
It's also worth noting my mouse setting is 1000Hz (kernel param usbhid.
Andreas Pokorny (andreas-pokorny) wrote : | # |
> AFAIK, resampling uses the original event timestamps which in android-input
> are accurate because they come directly from the kernel. So if those
> timestamps are being faked in libinput then that might explain the
> irregularity and sensitivity to scheduling. Better check that libinput (and
> Mir itself) is not inserting new timestamps. Our input reports also rely on
> those timestamps being the original event time from the kernel.
>
> It's also worth noting my mouse setting is 1000Hz (kernel param
> usbhid.
Well whatever you observe I cannot observe that. I see negative effects of resampling with and without libinput involved.
PS Jenkins bot (ps-jenkins) wrote : | # |
PASSED: Continuous integration, rev:3007
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
Andreas Pokorny (andreas-pokorny) wrote : | # |
There is now a silo for vivid overlay and this newer libinput package is about to land in xenial:
https:/
With that the known touchscreen troubles are gone - mouse is still fast though (as actually requested by many people)
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:3008
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
PS Jenkins bot (ps-jenkins) wrote : | # |
PASSED: Continuous integration, rev:3009
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
Alan Griffiths (alan-griffiths) wrote : | # |
-if(MIR_
+if(MIR_
...
+#ifdef MIR_BUILD_
+ ON_CALL(mock_x11, XOpenDisplay(
+#endif
Should the test setup really depend on the *build* platform? (Or is this poor naming - should MIR_BUILD_
Andreas Pokorny (andreas-pokorny) wrote : | # |
> -if(MIR_
> "mesa-x11")
> +if(MIR_
> ...
> +#ifdef MIR_BUILD_
> + ON_CALL(mock_x11, XOpenDisplay(
> +#endif
>
> Should the test setup really depend on the *build* platform? (Or is this poor
> naming - should MIR_BUILD_
Yeah rather poor naming. MIR_BUILD_PLATFORM indicates that we build for platform X11 .. to be honest the way autoconf names target, host and build still confuses me regularly.
Good that you mention this. This is one of the things I wanted to have another look.
Alan Griffiths (alan-griffiths) wrote : | # |
Nit:
+struct StubEmergencyCl
+{
+ void add(mir:
+};
This needs to implement the other overload.
Alan Griffiths (alan-griffiths) wrote : | # |
Also, in the following scenario:
$ sudo bin/mir_demo_server --window-manager system-compositor --display-config sidebyside --vt 1 --arw-file
$ bin/mir_demo_server --host /tmp/mir_socket --display-config clone --window-manager tiling
$ bin/mir_
I find that Alt+middle-
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:3010
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:3011
http://
Executed test runs:
FAILURE: http://
SUCCESS: http://
FAILURE: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
Alan Griffiths (alan-griffiths) wrote : | # |
A bit more detail on the problems I see:
I'm using a lenovo with a trackpoint.
The trackpoint controls the cursor unless the middle "mouse" button is pressed. E.g. Alt+left button drags the triangle, With Alt on its own or with Alt+right it just move the cursor.
Just pressing the middle button (with or without Alt) and the cursor no longer responds to the trackpoint.
It might be diagnostic that the left and right buttons do work, but the middle button doesn't - is something misidentifying it as a two button mouse?
Plugging in a spare 3-button mouse does work.
Cemil Azizoglu (cemil-azizoglu) wrote : | # |
613 -if(MIR_
614 +if(MIR_
Why is this needed? If the test is being built for Android only why should mesa related stuff be linked in?
Andreas Pokorny (andreas-pokorny) wrote : | # |
> A bit more detail on the problems I see:
>
> I'm using a lenovo with a trackpoint.
>
> The trackpoint controls the cursor unless the middle "mouse" button is
> pressed. E.g. Alt+left button drags the triangle, With Alt on its own or with
> Alt+right it just move the cursor.
>
> Just pressing the middle button (with or without Alt) and the cursor no longer
> responds to the trackpoint.
>
> It might be diagnostic that the left and right buttons do work, but the middle
> button doesn't - is something misidentifying it as a two button mouse?
>
> Plugging in a spare 3-button mouse does work.
Not quite. I diagnosed that with your evemu record and libinput-
So by default - it configures your device to use the middle button as scroll-button - so trackpoint motion is converted to scroll axis events while middle button is pressed. You should be able to deactivate that behaviour with the lp:~andreas-pokorny/mir/example-configure-input-devices MP - since it overwrite those settings.
Alan Griffiths (alan-griffiths) wrote : | # |
> So by default - it configures your device to use the middle button as scroll-
> button - so trackpoint motion is converted to scroll axis events while middle
> button is pressed. You should be able to deactivate that behaviour with the lp
> :~andreas-
> those settings.
I tried adding that, but (after a few merge conflicts - I'm a little way off trunk because of the landings backlog) the server exited because if couldn't find an input platform.
I'll try again later. Meanwhile I won't block on a problem that's being worked on elsewhere.
Andreas Pokorny (andreas-pokorny) wrote : | # |
I thought we do not build the unit tests for
> 613 -if(MIR_
> STREQUAL "mesa-x11")
> 614 +if(MIR_
>
> Why is this needed? If the test is being built for Android only why should
> mesa related stuff be linked in?
Back then I got linker errors whithout that depending on the ordering of the MIR_PLATFORM variable. This is no longer the case..
In general we should find a way to avoid the machinery of MIR_TEST_PLATFORM as far as possible.
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:3013
http://
Executed test runs:
FAILURE: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:3015
http://
Executed test runs:
FAILURE: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
PS Jenkins bot (ps-jenkins) wrote : | # |
PASSED: Continuous integration, rev:3016
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Continuous integration, rev:3016
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
FAILURE: http://
Click here to trigger a rebuild:
http://
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Autolanding.
More details in the following jenkins job:
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
FAILURE: http://
PS Jenkins bot (ps-jenkins) wrote : | # |
FAILED: Autolanding.
More details in the following jenkins job:
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
FAILURE: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
SUCCESS: http://
deb: http://
FAILURE: http://
SUCCESS: http://
FAILURE: http://
FAILED: Continuous integration, rev:2992 jenkins. qa.ubuntu. com/job/ mir-ci/ 5149/ jenkins. qa.ubuntu. com/job/ mir-android- vivid-i386- build/4287 jenkins. qa.ubuntu. com/job/ mir-clang- vivid-amd64- build/3194 jenkins. qa.ubuntu. com/job/ mir-mediumtests -vivid- touch/4229/ console jenkins. qa.ubuntu. com/job/ mir-mediumtests -wily-touch/ 187/console jenkins. qa.ubuntu. com/job/ mir-wily- amd64-ci/ 1303 jenkins. qa.ubuntu. com/job/ mir-wily- amd64-ci/ 1303/artifact/ work/output/ *zip*/output. zip jenkins. qa.ubuntu. com/job/ mir-wily- i386-ci/ 187/console jenkins. qa.ubuntu. com/job/ mir-mediumtests -builder- vivid-armhf/ 4230/console jenkins. qa.ubuntu. com/job/ mir-mediumtests -builder- wily-armhf/ 188/console
http://
Executed test runs:
SUCCESS: http://
SUCCESS: http://
FAILURE: http://
FAILURE: http://
SUCCESS: http://
deb: http://
FAILURE: http://
FAILURE: http://
FAILURE: http://
Click here to trigger a rebuild: s-jenkins. ubuntu- ci:8080/ job/mir- ci/5149/ rebuild
http://