| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |
| Who | What | When |
|---|---|---|
| mchehab | sailus: you sent me a txt->yaml patch against 5.9. Why should it be applied for 5.9, instead of 5.10? | [08:06] |
| ........ (idle for 38mn) | ||
| sailus | mchehab: There's a patch that added the supply names in 5.9.
If it's not merged now, we have a release with upper case supply names, and changing them after that is a no-no. I.e. we have to change it now, not later. There's also a corresponding driver patch in review but that can be merged later as it's just a regular "bug". The driver didn't use regulators earlier so there's nothing that broke here either. | [08:44] |
| .... (idle for 17mn) | ||
| mchehab | sailus: ok. was it reviewed by DT people? (didn't check the PR) | [09:03] |
| sailus | Yes. | [09:06] |
| mchehab | Reviewed-by Rob Herring <robh@kernel.org>
OK I'll merge it via the fixes tree | [09:06] |
| sailus | https://lore.kernel.org/linux-devicetree/20200915200851.GA2422660@bogus/
(I didn't get the Reviewed-by: immediately, our mail server ate it. :-P) I hope it is fixed by now. | [09:07] |
| mchehab | patch applied, thanks. | [09:09] |
| sailus | mchehab: Obrigado! | [09:10] |
| mchehab | anytime!
btw, next time, IMO the better would be to have two separate patches... one with the DT fix, and the other one with the conversion makes clearer | [09:10] |
| sailus | mchehab: Agreed.
I think that was the intention but somehow they got merged, and with tight schedule we just proceeded with it. | [09:11] |
| mchehab | also, the way it is, I'll probably need to write a more detailed description when sending the PR upstream, to explain why it would make sense to merge the conversion so late as a fix | [09:12] |
| sailus | Would the text in my PR be of any use? | [09:12] |
| mchehab | still there is a chance that this might got rejected by chief maintainer | [09:12] |
| sailus | Yes. | [09:12] |
| mchehab | I should be sending the PR later this week, probably on Thursday
feel free to ping me if I forget ;-) | [09:13] |
| sailus | Sure! | [09:15] |
| ..... (idle for 22mn) | ||
| pinchartl | mchehab: hello :-) | [09:37] |
| mchehab | pinchartl: hi | [09:44] |
| pinchartl | how are you doing ? | [09:46] |
| mchehab | fine, and you? | [09:48] |
| pinchartl | I'm doing good, thanks
I was wondering if you had seen my e-mail about the libcamera build failure in jenkins | [09:50] |
| ......................................... (idle for 3h20mn) | ||
| mchehab | no, sorry, I haven't seen yet
let me check changing the build option right now changed to android=auto worked | [13:11] |
| pinchartl | thank you :-) | [13:14] |
| mchehab | please send me later the packages that the builder needs in order to build it in ful
00:00:05.097 Dependency libexif skipped: feature android disabled 00:00:05.097 Dependency libjpeg skipped: feature android disabled 00:00:05.097 Found CMake: NO btw, are you planning to change from mason to cmake? | [13:14] |
| pinchartl | cmake isn't needed
it's used by meson as a backend to locate packages, in addition to pkg-config it's nothing we handle ourselves, it's internal to meson, and it's not required here for the android camera HAL to build, we need libexif and libjpeg we use libjpeg-turbo, in theory it should work with the original libjpeg too, but I don't think we've tested that | [13:15] |
| mchehab | $ sudo apt install -y libjpeg-turbo
Reading package lists... Done Building dependency tree Reading state information... Done E: Unable to locate package libjpeg-turbo | [13:18] |
| .... (idle for 18mn) | ||
| funny that meson requires cmake
libjpeg-dev is already the newest version (1:1.5.2-2). libexif-dev is already the newest version (0.6.21-5.1+deb10u4). pinchartl: no idea why it is disabling android build https://builder.linuxtv.org/job/libcamera/299/console | [13:36] | |
| 00:00:04.536 Run-time dependency gstreamer-video-1.0 found: NO (tried pkgconfig and cmake) 00:00:04.536 Run-time dependency gstreamer-allocators-1.0 found: NO (tried pkgconfig and cmake)
it is also requiring some gst packages libgstreamer1.0-dev is not enough | [13:46] | |
| pinchartl | I'm starting a meeting in 4 minutes, I'll have a look right after
mchehab: you need to add -Dandroid=auto to the meson setup line meson setup -Dandroid=auto build the option is disabled by default, as most users don't need the Android camera HAL and for gstreamer, you also need libgstreamer-plugins-base1.0-dev it's in the README file if that's not enough, please let me know | [13:49] |
| ........ (idle for 38mn) | ||
| mchehab | rm -rf build
meson setup build meson configure -Dandroid=auto -Dv4l2=true build ninja -C build pinchartl: ^ that's the build script it runs | [14:30] |
| ........... (idle for 52mn) | ||
| pinchartl | mchehab: that's strange. with the same version of meson and the libcamera master branch, the first thing that ninja does when running those above commands on my machine is to reconfigure the project, as the configuration options have changed. that doesn't happen on the build server according to the log, I don't know why
as a workaround you can use meson setup -Dandroid=auto -Dv4l2=true build and drop the meson configure line that also simplifies the build script :-) | [15:22] |
| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |