100789 2012-10-30 14:40:00 -0700 glsl-function-atan.html WebGL conformance test fails after https://bugs.webkit.org/show_bug.cgi?id=99154 2012-10-31 22:56:49 -0700 1 1 1 Unclassified WebKit WebGL 528+ (Nightly build) Unspecified Unspecified RESOLVED FIXED P2 Normal --- 1 roger_fong webkit-unassigned barraclough dino fpizlo hausmann roger_fong yuqiang.xian oldest_to_newest 754670 0 roger_fong 2012-10-30 14:40:00 -0700 Hello, I'm not savvy enough with Javascript, Compilers or MacroAssemblers to know what's causing this but a WebGL conformance test started failing after this patch (https://bugs.webkit.org/show_bug.cgi?id=99154 / r131858) landed. The test is here https://www.khronos.org/registry/webgl/conformance-suites/1.0.1/conformance/glsl/functions/glsl-function-atan.html. In particular the part of the test that is failing is this. tests: [ { source: ["$(output) = vec4(", " $(func)($(input).x * 8.0 - 4.0) / k2PI + 0.5,", " 0.5,", " 0,", " 1);"].join("\n"), generator: function(x, y, z, w) { return [ atan(x * 8.0 - 4.0) / k2PI + 0.5, 0.5, 0, 1 ]; }, }, The "generator" is the reference for the test and it seems to be having trouble with the second (green) element of the array, which is set to .5. I noticed however, that if I set it to anything besides .5 (like .50001) the test passes. Could someone take a look at it? Thanks 754839 1 yuqiang.xian 2012-10-30 17:51:13 -0700 I'm looking. Thanks. 754916 2 171571 yuqiang.xian 2012-10-30 20:41:00 -0700 Created attachment 171571 patch Sorry for introducing the bug. 755013 3 yuqiang.xian 2012-10-31 00:00:38 -0700 Fix landed: https://trac.webkit.org/changeset/132991 755903 4 roger_fong 2012-10-31 22:56:49 -0700 Many thanks! 171571 2012-10-30 20:41:00 -0700 2012-10-30 23:46:04 -0700 patch 100789.0.patch text/plain 1487 yuqiang.xian SW5kZXg6IFNvdXJjZS9KYXZhU2NyaXB0Q29yZS9DaGFuZ2VMb2cKPT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gU291 cmNlL0phdmFTY3JpcHRDb3JlL0NoYW5nZUxvZwkocmV2aXNpb24gMTMyOTc5KQorKysgU291cmNl L0phdmFTY3JpcHRDb3JlL0NoYW5nZUxvZwkod29ya2luZyBjb3B5KQpAQCAtMSwzICsxLDE1IEBA CisyMDEyLTEwLTMwICBZdXFpYW5nIFhpYW4gIDx5dXFpYW5nLnhpYW5AaW50ZWwuY29tPgorCisg ICAgICAgIGdsc2wtZnVuY3Rpb24tYXRhbi5odG1sIFdlYkdMIGNvbmZvcm1hbmNlIHRlc3QgZmFp bHMgYWZ0ZXIgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTk5MTU0Cisg ICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD0xMDA3ODkKKwor ICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBXZSBhY2NpZGVu dGx5IG1pc3NlZCBhIGJpdHdpc2UgZG91YmxlIHRvIGludDY0IGNvbnZlcnNpb24uCisKKyAgICAg ICAgKiBkZmcvREZHU3BlY3VsYXRpdmVKSVQuaDoKKyAgICAgICAgKEpTQzo6REZHOjpTcGVjdWxh dGl2ZUpJVDo6c2lsZW50RmlsbCk6CisKIDIwMTItMTAtMzAgIEpvc2VwaCBQZWNvcmFybyAgPHBl Y29yYXJvQGFwcGxlLmNvbT4KIAogICAgICAgICBbTWFjXSBTeW5jIHVwIEZlYXR1cmVEZWZpbmUg Q29uZmlndXJhdGlvbiBGaWxlcwpJbmRleDogU291cmNlL0phdmFTY3JpcHRDb3JlL2RmZy9ERkdT cGVjdWxhdGl2ZUpJVC5oCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFNvdXJjZS9KYXZhU2NyaXB0Q29yZS9kZmcv REZHU3BlY3VsYXRpdmVKSVQuaAkocmV2aXNpb24gMTMyOTc5KQorKysgU291cmNlL0phdmFTY3Jp cHRDb3JlL2RmZy9ERkdTcGVjdWxhdGl2ZUpJVC5oCSh3b3JraW5nIGNvcHkpCkBAIC01NDMsNyAr NTQzLDcgQEAgcHVibGljOgogICAgICAgICAgICAgbV9qaXQubW92ZSh2YWx1ZU9mSlNDb25zdGFu dEFzSW1tNjQocGxhbi5ub2RlSW5kZXgoKSksIHBsYW4uZ3ByKCkpOwogICAgICAgICAgICAgYnJl YWs7CiAgICAgICAgIGNhc2UgU2V0RG91YmxlQ29uc3RhbnQ6Ci0gICAgICAgICAgICBtX2ppdC5t b3ZlKEltbTY0KHZhbHVlT2ZOdW1iZXJDb25zdGFudChwbGFuLm5vZGVJbmRleCgpKSksIGNhblRy YW1wbGUpOworICAgICAgICAgICAgbV9qaXQubW92ZShJbW02NChyZWludGVycHJldERvdWJsZVRv SW50NjQodmFsdWVPZk51bWJlckNvbnN0YW50KHBsYW4ubm9kZUluZGV4KCkpKSksIGNhblRyYW1w bGUpOwogICAgICAgICAgICAgbV9qaXQubW92ZTY0VG9Eb3VibGUoY2FuVHJhbXBsZSwgcGxhbi5m cHIoKSk7CiAgICAgICAgICAgICBicmVhazsKICAgICAgICAgY2FzZSBMb2FkMzJQYXlsb2FkQm94 SW50Ogo=