Here's a git pre-commit hook. In .git/hooks/pre-commit:
====
#!/bin/bash
# Check if any files were touched under test_lists.
diff_cmd="git diff-index --cached HEAD -- webkit/tools/layout_tests/test_lists"
if $diff_cmd | grep -q .; then
echo "Test lists changed; running linter..."
./webkit/tools/layout_tests/run_webkit_tests.sh --lint-test-files
exit $?
fi
exit 0
====
However, the 30s lint time is kind of killer. If we get time one of
us may fix the linter.
On Wed, Dec 10, 2008 at 11:17 AM, Ojan Vafai <[EMAIL PROTECTED]> wrote:
> There's been significant set of changes to the test expectations lists
> recently. There seems to still be some confusion over the state of the
> world. So, hopefully I can clarify a bit.
> Firstly, there is now a lint option that will parse the test list for each
> platform (win/mac/linux) and each build target (debug/release). This way you
> can know that you at least got the syntax right. So, please consider running
> this when you make test list changes. It takes ~30 seconds on my machine and
> saves you from closing the tree. Also, if you feel moved to improve our
> tools a bit, add a pre-commit hook to gcl/git-cl to do this for us
> (please!).
> $ ./run_webkit_tests.sh --lint-test-files
> 081210 11:04:34 run_webkit_tests.py:142 INFO Found: 10244 tests
> 081210 11:04:59 run_webkit_tests.py:614 INFO SUCCESS
>
> As for expecations, the new format is:
> METADATA : path_to_test = EXPECTATIONS
> METADATA and EXPECTATIONS are space-separated lists.
> Tests can have any combination of the following metadata associated
> with them (optional):
> SKIP : We don't want to run the test (perhaps because it hangs).
> DEBUG : Expectations apply only to the Debug build.
> RELEASE : Expectations apply only to the Release build.
> LINUX/WIN/MAC: Expectations apply only to the listed platform(s).
> GOOD EXAMPLES
> // Test fails for all platforms.
> LayoutTests/path/to/test.html = FAIL
>
> // Test fails for linux only.
> LINUX : LayoutTests/path/to/test.html = FAIL
> // Test times out for linux, fails on win-release but crashes on win-debug
> LINUX : LayoutTests/path/to/test.html = TIMEOUT
> WIN RELEASE : LayoutTests/path/to/test.html = FAIL
> WIN DEBUG : LayoutTests/path/to/test.html = CRASH
> BAD EXAMPLES
> // Fails on win-release but crashes on win-debug
> // This ends up defining two sets of expectations for win-debug.
> WIN : LayoutTests/path/to/test.html = FAIL
> WIN DEBUG : LayoutTests/path/to/test.html = CRASH
>
> // Fails on linux and times out on windows.
> // This won't throw an error, but loses information.
> // Now if windows starts failing, or linux starts timing out
> // we won't notice.
> LINUX WIN : LayoutTests/path/to/test.html = FAIL TIMEOUT
> So, as shown above, please use the most specific metadata available to you.
> If you have reason to believe a test fails on all platforms (e.g. it's a new
> test from a webkit merge or it started failing from a change to non-platform
> specific code) feel free to make it apply to all platforms.
> If there are ways to make any of this simpler, feel free to chip in. :)
> Ojan
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Chromium-dev" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/chromium-dev?hl=en
-~----------~----~----~----~------~----~------~--~---
