I created a doc <https://docs.google.com/document/d/1IwpQ_Od3EiWmzgEXPzPNf_V9ZM9Yq-Y9sy3anfAxnMI/edit?usp=sharing&resourcekey=0-2B2NrXBGxpcRKJ_X9-kjtg> with information about diagnosing the cause of flakiness as threaded compositing, and debugging tests that are failing or becoming flaky due to threaded compositing. This is a living document and we will add more information about debugging these kind of tests as we learn more. I hope you find it useful. (Link to the doc - https://docs.google.com/document/d/1IwpQ_Od3EiWmzgEXPzPNf_V9ZM9Yq-Y9sy3anfAxnMI/edit?usp=sharing&resourcekey=0-2B2NrXBGxpcRKJ_X9-kjtg)
Yotam On Monday, June 12, 2023 at 1:26:08 PM UTC-7 Morten Stenshorne wrote: > Sure - bug filed: > https://bugs.chromium.org/p/chromium/issues/detail?id=1454242 > > On Mon, 12 Jun 2023 at 16:35, Stefan Zager <sza...@chromium.org> wrote: > >> On Mon, Jun 12, 2023 at 2:59 AM Morten Stenshorne <msten...@chromium.org> >> wrote: >> >>> Yotam Hacohen <yo...@chromium.org> writes: >>> >>> > tl;dr - web_tests will switch to running with threaded compositing by >>> default. >>> > [...] >>> > This change shouldn't increase failures, flakiness, or test runtime, >>> > and should not increase the gardener's burden. >>> >>> This does affect test runtime. It seems especially bad when combining >>> WPT tests and non-WPT tests - about twice as slow as before. >>> >>> Without crrev.com/c/4599846 it takes 33 seconds. With it, 59 seconds. >>> >>> The command I used: >>> tools/run_web_tests.py -t Release -f external/wpt/css/css-multicol >>> fast/multicol >>> >> >> Morten, can you file a bug about this? >> >> We did analyze try bot runtimes before launching, and found little to no >> effect. >> We haven't yet had a chance to analyze runtimes across the CI fleet after >> launching, but we intend to. >> >> The test runner spins up a new content_shell for each unique permutation >> of command-line >> flags it encounters. It's likely that the test runner is creating more >> content_shells with this feature >> enabled, and I wonder if it's leading to thrashing on your system. >> >> Thanks, >> >> Stefan >> >> >>> -- >>> Morten Stenshorne, Software developer, >>> Blink/Layout, Google, Oslo, Norway >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "blink-dev" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to blink-dev+unsubscr...@chromium.org. >>> To view this discussion on the web visit >>> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/ozzi4jndc86x.fsf%40aeneas.osl.corp.google.com >>> . >>> >> -- You received this message because you are subscribed to the Google Groups "blink-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscr...@chromium.org. To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/00e455e4-2b96-4296-a9d2-bcd0b838ab74n%40chromium.org.
