| T42 | <smatkovi> does anyone know if clang-tools-extra is needed to build gecko-dev? | 07:18 |
|---|---|---|
| T42 | <smatkovi> or asyou whish xulrunner | 07:18 |
| *** amccarthy is now known as Guest6308 | 07:41 | |
| *** amccarthy_ is now known as amccarthy | 07:41 | |
| T42 | <elros34> usually all dependencies should be defined in spec file. You only need to apply ~100 patches (prep step) and build | 07:55 |
| T42 | <smatkovi> Isn't this automatically done when i run mb2 -t <target> -s rpm/clang.spec build? | 09:04 |
| T42 | <smatkovi> But then maybe not, because running this with xulrunner.spec doesn't create the entries in rpm-shared.env | 09:05 |
| T42 | <smatkovi> So instead of build i need to execute prep step? | 09:05 |
| T42 | <smatkovi> Before build. Or prepare as it is called now apparently | 09:06 |
| T42 | <smatkovi> i found these in usage: build-init Initialize build directory; build-requires Update build time dependencies; prepare Prepare sources using the recipe from RPM SPEC file; apply Apply patches listed in the RPM SPEC file | 09:14 |
| T42 | <smatkovi> are they to be applied in this order? | 09:15 |
| T42 | <smatkovi> as i understood i can just run build -p and then also prepare is execetuted | 09:21 |
| CarbonatedCaffeine | Hello, i'm back!! | 09:26 |
| T42 | <smatkovi> wb | 09:29 |
| CarbonatedCaffeine | I took a good break from porting, got too frustrating | 09:30 |
| T42 | <smatkovi> yes it seems like a huge task | 10:44 |
| T42 | <smatkovi> what did you port? | 10:44 |
| CarbontatedCaffeine | I was porting the Google Pixel XL (original) | 11:11 |
| T42 | <Mister_Magister> lol u (re @SailfishFreenodeIRCBridgeBot: <CarbonatedCaffeine>...) | 12:00 |
| CarbontatedCaffeine | Me | 12:25 |
| T42 | <b100dian> mal: that slowness in browser video playback, followed by failure -> graphicBuffer is behaving differently I think for ccodec2 reasons https://github.com/sailfishos/droidmedia/pull/111 This only fixes gecko-camera playback, gst-droid still TBD | 22:43 |
| mal | @b100dian ok, need to test that | 22:48 |
| mal | hopefully that doesn't cause any issues on other devices | 22:49 |
| mal | seems safe but needs verification | 22:49 |
| T42 | <b100dian> yeas, no rush. Am trying to debug gst-droid next, maybe there are some other corner cases. It may be that a new callback needs to be introduced for clients, right now it is only "not leaking native graphic buffers" but a releaseBuffer (not plural) may be needed. To be determined. | 22:53 |
| mal | @b100dian would checking getStrongCount make sense? | 22:55 |
| mal | before doing that decStrong | 22:55 |
| mal | not sure which android versions support getStrongCount, needs checking | 22:55 |
| T42 | <b100dian> I'll check this suggestion. Ideally, no time-of-check/time-of-use race should be introduced, that's why I'm saying that the client should be aware. | 22:57 |
| T42 | <b100dian> rn the client thinks it has a ton of buffers, but most of them were released back to mediacodec, they're just small memory | 22:58 |
| T42 | <b100dian> there's no slot in clients, and that would be complex to introduce | 22:58 |
| mal | just happened to check decStrong implementation in RefBase.cpp and it logs some warning if that is called too many times | 22:59 |
| mal | but that was some random version of RefBase | 23:00 |
| mal | https://android.googlesource.com/platform/system/core/+/master/libutils/RefBase.cpp#488 | 23:01 |
| T42 | <b100dian> I should look if that message appears. Also, if the repo would allow draft merge requests, I would have done a draft until I figure out gst-droid. | 23:06 |
| T42 | <b100dian> ah, it has, converted to draft | 23:08 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!