On Thu, 24 Sep 2026 17:55:10 GMT, Phil Race <[email protected]> wrote:
>> PolyVertTest verifies the drawing of some de-generate polylines and polygons. >> It has been problem-listed on all platforms for some time. >> The test isn't particularly stable - problems with scaled UIs and rounded >> window edges, and how it waits for the surface to be rendered. >> I've updated it to fix all these things and save an image of any failed >> cases. >> Also the test didn't specify any pipelines, so would run with the default. >> On windows, where we use D3D if we can, but often fall back to software, >> this can be misleading >> as the test passes on software (since the same code is used for the >> reference image) but fails with D3D. >> >> So we still need to problem-list the "default" case on windows, in case it >> is D3D. >> I also found that the default case fails with the metal pipeline on macOS. >> Also although OpenGL passes on Windows & macOS it fails on Linux .. >> >> Cases that do pass are >> - xrender=true (default) and xrender=false (x11) on Linux >> - opengl=true on windows & macOS >> - d3d=false on windows >> >> Although each platform has some failure mode, I don't think it can be called >> a "generic" failure. >> i.e I don't think it likely that the reason for an OpenGL failure on Linux >> is the same as D3D on Windows. >> So I've made sub-test cases, and will use separate bugs for each of the 3 >> failure cases. >> >> --------- >> - [x] I confirm that I make this contribution in accordance with the >> [OpenJDK Interim AI Policy](https://openjdk.org/legal/ai). > > Phil Race has updated the pull request incrementally with one additional > commit since the last revision: > > 6986565 Updated. ------------- PR Review: https://git.openjdk.org/jdk/pull/32992#pullrequestreview-5308101822
