VV8 LUCI CQ[test] Remove --testing-d8-test-runner
| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
[test] Infer status-file variables from build flags This auto-generates all build-time flags for status-file configuration from the flags specified in V8's BUILD.gn file. This makes makes adding such a flag a one-line change in the future, instead of >10 lines in >10 files as in https://crrev.com/c/4521710. This also raises an improved exception when a build flag is missing either in production or in tests, showing a hint with the location where to add it. This also removes redundant test data and instead bootstraps V8's build configuration for testing from a single blueprint. Cq-Include-Trybots: luci.v8.try:v8_numfuzz_dbg Cq-Include-Trybots: luci.v8.try:v8_numfuzz_rel Cq-Include-Trybots: luci.v8.try:v8_numfuzz_tsan_rel Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel Cq-Include-Trybots: luci.v8.try:v8_android_arm64_n5x_rel Cq-Include-Trybots: luci.v8.try:v8_linux64_msan_rel Cq-Include-Trybots: luci.v8.try:v8_linux64_ubsan_rel Bug: chromium:1132088 Change-Id: I53c002857fcacf7368c23c249209170e7af6291b Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/4528239 Reviewed-by: Liviu Rau <liviurau@google.com> Commit-Queue: Michael Achenbach <machenbach@chromium.org> Cr-Commit-Position: refs/heads/main@{#87674} | 3 年前 | |
[test] Infer status-file variables from build flags This auto-generates all build-time flags for status-file configuration from the flags specified in V8's BUILD.gn file. This makes makes adding such a flag a one-line change in the future, instead of >10 lines in >10 files as in https://crrev.com/c/4521710. This also raises an improved exception when a build flag is missing either in production or in tests, showing a hint with the location where to add it. This also removes redundant test data and instead bootstraps V8's build configuration for testing from a single blueprint. Cq-Include-Trybots: luci.v8.try:v8_numfuzz_dbg Cq-Include-Trybots: luci.v8.try:v8_numfuzz_rel Cq-Include-Trybots: luci.v8.try:v8_numfuzz_tsan_rel Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel Cq-Include-Trybots: luci.v8.try:v8_android_arm64_n5x_rel Cq-Include-Trybots: luci.v8.try:v8_linux64_msan_rel Cq-Include-Trybots: luci.v8.try:v8_linux64_ubsan_rel Bug: chromium:1132088 Change-Id: I53c002857fcacf7368c23c249209170e7af6291b Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/4528239 Reviewed-by: Liviu Rau <liviurau@google.com> Commit-Queue: Michael Achenbach <machenbach@chromium.org> Cr-Commit-Position: refs/heads/main@{#87674} | 3 年前 | |
[test] Infer status-file variables from build flags This auto-generates all build-time flags for status-file configuration from the flags specified in V8's BUILD.gn file. This makes makes adding such a flag a one-line change in the future, instead of >10 lines in >10 files as in https://crrev.com/c/4521710. This also raises an improved exception when a build flag is missing either in production or in tests, showing a hint with the location where to add it. This also removes redundant test data and instead bootstraps V8's build configuration for testing from a single blueprint. Cq-Include-Trybots: luci.v8.try:v8_numfuzz_dbg Cq-Include-Trybots: luci.v8.try:v8_numfuzz_rel Cq-Include-Trybots: luci.v8.try:v8_numfuzz_tsan_rel Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel Cq-Include-Trybots: luci.v8.try:v8_android_arm64_n5x_rel Cq-Include-Trybots: luci.v8.try:v8_linux64_msan_rel Cq-Include-Trybots: luci.v8.try:v8_linux64_ubsan_rel Bug: chromium:1132088 Change-Id: I53c002857fcacf7368c23c249209170e7af6291b Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/4528239 Reviewed-by: Liviu Rau <liviurau@google.com> Commit-Queue: Michael Achenbach <machenbach@chromium.org> Cr-Commit-Position: refs/heads/main@{#87674} | 3 年前 | |
[test] Refactor testrunner (3) - removed some 'feature envy' instances: - created a AugmentedOptions class to take care of non trivial option look-ups - moved some calculation closer the the actual data - promoted parameter that was passed around a lot (options) to filed in the offending class - similar object creation looks a bit more similar - CommandContext provides a wrapper that ensures resource de-allocation - separate tests from standard and num_fuzz runners - added a couple of more tests to improve coverage This is still in flux. But further changes risk creating a disconnect between the original implementation and further refactorings. Bug: v8:12785 Change-Id: I0ec2e9c6a81604a5cd1d4a80982dd3329c1b48db Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3686411 Commit-Queue: Liviu Rau <liviurau@chromium.org> Reviewed-by: Michael Achenbach <machenbach@chromium.org> Cr-Commit-Position: refs/heads/main@{#81148} | 4 年前 | |
[test] Infer status-file variables from build flags This auto-generates all build-time flags for status-file configuration from the flags specified in V8's BUILD.gn file. This makes makes adding such a flag a one-line change in the future, instead of >10 lines in >10 files as in https://crrev.com/c/4521710. This also raises an improved exception when a build flag is missing either in production or in tests, showing a hint with the location where to add it. This also removes redundant test data and instead bootstraps V8's build configuration for testing from a single blueprint. Cq-Include-Trybots: luci.v8.try:v8_numfuzz_dbg Cq-Include-Trybots: luci.v8.try:v8_numfuzz_rel Cq-Include-Trybots: luci.v8.try:v8_numfuzz_tsan_rel Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel Cq-Include-Trybots: luci.v8.try:v8_android_arm64_n5x_rel Cq-Include-Trybots: luci.v8.try:v8_linux64_msan_rel Cq-Include-Trybots: luci.v8.try:v8_linux64_ubsan_rel Bug: chromium:1132088 Change-Id: I53c002857fcacf7368c23c249209170e7af6291b Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/4528239 Reviewed-by: Liviu Rau <liviurau@google.com> Commit-Queue: Michael Achenbach <machenbach@chromium.org> Cr-Commit-Position: refs/heads/main@{#87674} | 3 年前 | |
[test] Infer status-file variables from build flags This auto-generates all build-time flags for status-file configuration from the flags specified in V8's BUILD.gn file. This makes makes adding such a flag a one-line change in the future, instead of >10 lines in >10 files as in https://crrev.com/c/4521710. This also raises an improved exception when a build flag is missing either in production or in tests, showing a hint with the location where to add it. This also removes redundant test data and instead bootstraps V8's build configuration for testing from a single blueprint. Cq-Include-Trybots: luci.v8.try:v8_numfuzz_dbg Cq-Include-Trybots: luci.v8.try:v8_numfuzz_rel Cq-Include-Trybots: luci.v8.try:v8_numfuzz_tsan_rel Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel Cq-Include-Trybots: luci.v8.try:v8_android_arm64_n5x_rel Cq-Include-Trybots: luci.v8.try:v8_linux64_msan_rel Cq-Include-Trybots: luci.v8.try:v8_linux64_ubsan_rel Bug: chromium:1132088 Change-Id: I53c002857fcacf7368c23c249209170e7af6291b Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/4528239 Reviewed-by: Liviu Rau <liviurau@google.com> Commit-Queue: Michael Achenbach <machenbach@chromium.org> Cr-Commit-Position: refs/heads/main@{#87674} | 3 年前 | |
[test] Enable running the standard runner with numfuzz behavior The two frameworks num_fuzzer and standard_runner tweak testing behavior slightly, which can cause problems when flake bisect runs for cases detected with num_fuzzer. Flake bisect is only implemented for the standard runner. The differences are: - Numfuzz uses a few more extra default flags. - Numfuzz uses an additional test processor that ignores timeouts and expected failures. - Numfuzz passes an extra file in the Mjsunit test suite. This change enables running the standard runner with this specific numfuzz behavior, which makes flake bisect the same way as the original run. The --framework cmd-line flag needs to be passed. On the infra side, the framework name is already available for failures. In a follow up, we'll pass it as a cmd-line in fitting revisions during flake bisect. Bug: 42202137 Change-Id: I77703ba2c7f2ee345f2af2925370da14f3149ac6 Cq-Include-Trybots: luci.v8.try:v8_numfuzz_dbg,v8_numfuzz_tsan_rel Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/6063807 Reviewed-by: Alexander Schulze <alexschulze@chromium.org> Commit-Queue: Michael Achenbach <machenbach@chromium.org> Cr-Commit-Position: refs/heads/main@{#97616} | 1 年前 | |
[numfuzz] Silence presubmit test that shows a spurious failure ResultDB was interpreting the fake test errors from the test-framework test as real failures, which showed up on all CLs. For the purpose of the given test, it wasn't needed to actually run failing tests, so this change just makes them passing. Bug: 417413668 Change-Id: I2ed51e8b9be7d08f325d0ce7f31d820cd7eb1a38 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/6781050 Reviewed-by: Clemens Backes <clemensb@chromium.org> Commit-Queue: Michael Achenbach <machenbach@chromium.org> Cr-Commit-Position: refs/heads/main@{#101610} | 11 个月前 | |
[test] Remove --testing-d8-test-runner This flag is contained to V8 and its test runner. Since all usages were removed already, it's safe to remove it. Some tests passed --no-testing-d8-test-runner. Those tests were essentially dead code as the flag and the negation just created an expected trivial flag contradiction. This avoids passing --fuzzing and instead repairs existing test cases that lack a %Prepare call. This also flushed out a couple of tests with OptimizeOsr() in dead code. This repairs most of these with best effort. Tests for clean-up were identified with the following cmd line: git grep OptimizeOsr -- test/mjsunit | \ cut -d':' -f1 | uniq | sort | \ xargs grep -L PrepareFunctionForOptimization The test osr-elements-kind, which was dead code before, fails now in some variants, now tracked as https://crbug.com/453967876. Bug: 450577983, 453967876 Change-Id: I452548918a8926d0ee9fbadae58470bdc6129e52 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/7032166 Commit-Queue: Michael Achenbach <machenbach@chromium.org> Reviewed-by: Marja Hölttä <marja@chromium.org> Cr-Commit-Position: refs/heads/main@{#103252} | 8 个月前 | |
[test] Remove --testing-d8-test-runner This flag is contained to V8 and its test runner. Since all usages were removed already, it's safe to remove it. Some tests passed --no-testing-d8-test-runner. Those tests were essentially dead code as the flag and the negation just created an expected trivial flag contradiction. This avoids passing --fuzzing and instead repairs existing test cases that lack a %Prepare call. This also flushed out a couple of tests with OptimizeOsr() in dead code. This repairs most of these with best effort. Tests for clean-up were identified with the following cmd line: git grep OptimizeOsr -- test/mjsunit | \ cut -d':' -f1 | uniq | sort | \ xargs grep -L PrepareFunctionForOptimization The test osr-elements-kind, which was dead code before, fails now in some variants, now tracked as https://crbug.com/453967876. Bug: 450577983, 453967876 Change-Id: I452548918a8926d0ee9fbadae58470bdc6129e52 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/7032166 Commit-Queue: Michael Achenbach <machenbach@chromium.org> Reviewed-by: Marja Hölttä <marja@chromium.org> Cr-Commit-Position: refs/heads/main@{#103252} | 8 个月前 | |
[test] Set environment variables for sandbox hardware support Bug: 428152522, 428179540 Change-Id: If1e4de62401e4bb0d4fa73f67fdfaa71589ff9fd Cq-Include-Trybots: luci.v8.try:v8_linux64_pku_dbg Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/6680632 Reviewed-by: Samuel Groß <saelo@chromium.org> Commit-Queue: Michael Achenbach <machenbach@chromium.org> Cr-Commit-Position: refs/heads/main@{#101096} | 1 年前 |