Stskeeps | good morning | 07:25 |
---|---|---|
situ | Morning | 08:48 |
sledges | hellow! | 09:14 |
sledges | how have you been fellows? | 09:15 |
situ | sledges: Pretty good. I hope you enjoyed holidays. | 09:17 |
sledges | thanks, was very refreshing :) | 09:18 |
sledges | i see you're still talking about droid-hal-init ;)) | 09:19 |
*** vgrade has joined #sailfishos-porters | 09:19 | |
vgrade | morning | 09:20 |
situ | vgrade: Morning | 09:20 |
Stskeeps | morn vgrade | 09:20 |
Stskeeps | vgrade: aren't you like.. on a cruise? | 09:20 |
sledges | vgrade: ahoj | 09:20 |
Stskeeps | hrmm | 09:20 |
Stskeeps | or was that somebody else.. | 09:20 |
situ | sledges: We have fixed lipstick crash. Just need to get systemd services work in right order. | 09:21 |
Stskeeps | ah, 2009 | 09:21 |
Stskeeps | :P | 09:21 |
Stskeeps | vgrade: i have a n5 now too, bw | 09:21 |
Stskeeps | btw | 09:21 |
situ | Stskeeps: Please update topic of the channel :P | 09:21 |
Stskeeps | hehe | 09:22 |
sledges | situ: well done, what was the crash problem? | 09:23 |
* vgrade wonders why Mr Brook is posting old photos | 09:23 | |
vgrade | Mrs | 09:23 |
Stskeeps | vgrade: it said 'edited' | 09:23 |
Stskeeps | so might be why | 09:23 |
vgrade | wish I was on a criuse :) | 09:23 |
situ | sledges: https://github.com/siteshwar/android_hardware_qcom_display/commit/95f7ed38297e505f59d45c820d109bb01797ba4e | 09:24 |
sledges | situ: nice catch | 09:25 |
situ | sledges: Thanks to morphis, he pointed it to me :) | 09:26 |
sledges | sweeet! | 09:26 |
Stskeeps | we shouldn't need to mod any HAL bits, fwiw | 09:27 |
Stskeeps | it usually means we're using the HAL wrong if it crashes | 09:27 |
sledges | maybe thp could pin that better in hwc? | 09:28 |
Stskeeps | morphis sent a patch that looks like a reason why, to the hwcomposer qpa | 09:28 |
* sledges nods | 09:29 | |
thp | we had to patch samsung hwcomposer hal once because they dereferenced a null function pointer that was set from another thread with no synchronization | 09:38 |
thp | which works in surfaceflinger (luck?) but not from hwc hal | 09:39 |
Stskeeps | nod | 09:40 |
Stskeeps | should be exception, not rule though | 09:40 |
thp | indeed | 09:40 |
* thp drops out again | 09:40 | |
sledges | o/ | 09:42 |
situ | Stskeeps: surfaceflinger uses virtual displays and we don't. qpa needs to be fixed to use it. | 09:53 |
Stskeeps | ok | 09:53 |
Stskeeps | well, at least sailfishos looks beautiful on this device.. | 10:33 |
vgrade | :) | 10:35 |
* vgrade hopes he can get some time to port to his one+ | 10:36 | |
morphis | Stskeeps: right, there should be no need to modify the HAL bits if we're using them correctly the same way as Android does | 10:54 |
*** tbr has joined #sailfishos-porters | 11:02 | |
*** swex has joined #sailfishos-porters | 12:23 | |
*** tbr has quit IRC | 12:29 | |
*** Sage_ has joined #sailfishos-porters | 12:30 | |
*** shallow has quit IRC | 12:30 | |
*** Sage___ has quit IRC | 12:30 | |
*** hexo has joined #sailfishos-porters | 12:31 | |
*** hexo is now known as shallow | 12:32 | |
*** tbr has joined #sailfishos-porters | 12:59 | |
*** sledges has quit IRC | 13:04 | |
*** sledges has joined #sailfishos-porters | 13:05 | |
*** vgrade has quit IRC | 13:29 | |
sledges | thanks morphis; any objections anyone? https://github.com/mer-hybris/qt5-qpa-hwcomposer-plugin/pull/21#issuecomment-48167979 | 13:50 |
morphis | sledges, thp: what is with merging https://github.com/mer-hybris/qt5-qpa-hwcomposer-plugin/pull/20 and https://github.com/mer-hybris/qt5-qpa-hwcomposer-plugin/pull/19 too? | 14:08 |
situ | sledges: I belive this will cause problems on Nexus 5. | 14:13 |
situ | I will test and confirm. | 14:15 |
sledges | situ: thanks | 14:15 |
sledges | morphis: #20 merged, #19 needs testing | 14:19 |
sledges | situ: don't forget to revert num_displays | 14:22 |
situ | sledges: I didn't get it. | 14:22 |
sledges | https://github.com/siteshwar/android_hardware_qcom_display/commit/95f7ed38297e505f59d45c820d109bb01797ba4e | 14:23 |
situ | ok | 14:23 |
sledges | it should not be needed with 13 fix | 14:23 |
morphis | sledges: have pulled 19 into our branch, works fine so far | 14:27 |
sledges | thx | 14:28 |
situ | build fails | 14:29 |
situ | [ 67s] hwcomposer_backend_v11.cpp:42:37: fatal error: android/android-version.h: No such file or directory | 14:29 |
situ | This commit breaks compilation https://github.com/mer-hybris/qt5-qpa-hwcomposer-plugin/commit/3f201846fd3123d0aad43c76496d8bc960f7cda0 | 14:31 |
sledges | android-version is provided by dhd iirc | 14:35 |
sledges | can't understand how can not including #include "hwcomposer_backend_v0.h" for v13 would make difference? | 14:39 |
situ | sledges: oops sorry, it's this pull request https://github.com/mer-hybris/qt5-qpa-hwcomposer-plugin/pull/21/files | 14:45 |
situ | +#include <android/android-version.h> | 14:45 |
situ | lbt's commit had also broke compilation, but that was for differnt reasons. | 14:45 |
*** vgrade has joined #sailfishos-porters | 14:48 | |
situ | Looks like it was fixed by https://github.com/mer-hybris/qt5-qpa-hwcomposer-plugin/commit/c9d6a4a70634dafab25dfe87802c8dce7240b454 | 14:48 |
situ | sledges: droid-hal-hammerhead-devel package provides '/usr/lib/droid-devel/droid-headers/android-version.h'. | 15:21 |
* sledges nods | 15:25 | |
situ | morphis: How did you build this package ? | 15:28 |
morphis | situ: which one? | 15:30 |
situ | morphis: qpa | 15:30 |
morphis | situ: /home/morphis/work/wop/webos-ports/tmp-eglibc/work/tenderloin-webos-linux-gnueabi/linux-hp-tenderloin/3.0.101-20140707-61-r0/ :) | 15:31 |
morphis | ah damn it | 15:32 |
morphis | https://github.com/webOS-ports/meta-webos-ports/blob/dora/recipes-qt/qt5/qt5-qpa-hwcomposer-plugin_git.bb is the one | 15:32 |
morphis | not using Mer :) | 15:32 |
situ | sledges: Look at this line https://github.com/mer-hybris/qt5-qpa-hwcomposer-plugin/blob/master/hwcomposer/hwcomposer_backend_v11.cpp#L83 | 15:34 |
situ | sledges: and look at how surfaceflinger does this https://github.com/CyanogenMod/android_frameworks_native/blob/cm-11.0/services/surfaceflinger/DisplayHardware/HWComposer.cpp#L122 | 15:35 |
situ | It only initializes 1 display. | 15:35 |
situ | it doesn't matter what numDisplays is. That's what qpa plugin should do too. | 15:36 |
situ | Even if we manage to compile morphis's changes, qpa plugin will still be broken on Nexus 5. | 15:37 |
sledges | right | 15:38 |
situ | morphis: Is there any specific reason you want to use virtual displays ? | 15:40 |
morphis | you see the hwcHasApiVersion call before? | 15:40 |
morphis | it verifies you have atlest 1.0 | 15:40 |
morphis | before using numDisplays | 15:41 |
morphis | you only had one display in times before 1.0 | 15:41 |
situ | Ok. | 15:47 |
morphis | situ: I would assume that the nexus 5 should just work with that change | 16:01 |
situ | morphis: ok, I will try to get your changes compiled. | 16:02 |
*** vgrade has quit IRC | 17:18 | |
situ | sledges: I am trying to build from this repository https://github.com/siteshwar/qt5-qpa-hwcomposer-plugin/commits/hwcomposer-1.3-support | 19:48 |
situ | and stuck at this error : | 19:48 |
situ | [ 66s] /home/abuild/rpmbuild/BUILD/qt5-qpa-hwcomposer-plugin-5.1.0.1+hwcomposer.1.3.support.47.11.ga378b61/hwcomposer/hwcomposer_backend.cpp:119: undefined reference to `HwComposerBackend_v0::HwComposerBackend_v0(hw_module_t*, hw_device_t*)' | 19:48 |
situ | Fixed through https://github.com/siteshwar/qt5-qpa-hwcomposer-plugin/commit/935778d0f8949b374041eef52060be204f66c3d3 | 19:54 |
situ | morphis: I managed to compile your changes, but I start seeing tearing after 30 seconds or so. I am using this repository https://github.com/siteshwar/qt5-qpa-hwcomposer-plugin/commits/hwcomposer-1.3-support | 20:04 |
vgrade_ | situ: same tearing as before | 20:06 |
situ | Right | 20:06 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!