00:17 <alyssa> tomeu: We should look into running dEQP with --deqp-surface-type=fbo

00:17 <alyssa> It's slower (expect CI to take like 1.5x time) for unknown reasons; that needs to be sorted unrelated to CI since FBOs being slow is a bad thing.

00:18 <alyssa> That said it's what the tests actually want.... results will be more reliable. A lot of tests are skipped with =window. Other tests pass "automatically" without testing the actual thing.

00:18 <alyssa> Once that's changed, expect regressions. But at this point tests that work on one but not the other were never really working at all.

00:19 <alyssa> (It turns out stenciling is quite broken, for instance.)

06:06 <tomeu> alyssa: sounds good

06:06 <tomeu> btw, why would it be slower?

06:09 <tomeu> alos: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/master/client/site_tests/graphics_dEQP/graphics_dEQP.py#77

06:15 <tomeu> so ChromeOS seems to only run with fbo or pbuffer, but android seems to always run with window: https://android.googlesource.com/platform/external/deqp/+/master/android/cts/AndroidTest.xml

06:15 <tomeu> anyway, will look at that on monday

12:40 <hanetzer> alyssa: good on ya, just read the gnome/panfrost collabra post :D

13:01 <HdkR> https://www.cnx-software.com/2019/07/06/beelink-gt1-mini-2-amlogic-s905x3-tv-box/ Another cute Bifrost device

15:38 <hanetzer> bifrost is one of the panfrost capable devices, yes?

15:40 <hanetzer> wifi 5, what is that? 5ghz wifi?

15:41 <bnieuwenhuizen> apparently wifi 5 is an alternate name of 802.11ac?

15:41 <hanetzer> news to me

15:41 <bnieuwenhuizen> well, it is new: https://www.accessagility.com/blog/wifi-6-wifi-5-wifi-4-new-generational-wifi-names-from-the-wifi-alliance

15:41 <bnieuwenhuizen> (well, okay, not *that* new)

17:17 <hanetzer> ah, interesting.