[webkit-reviews] review requested: [Bug 47220] new-run-webkit-tests - enable cygwin support for chromium win : [Attachment 70702] Patch

[bugzilla-daemon at webkit.org]

Dirk Pranke <dpranke at chromium.org> has asked  for review:
Bug 47220: new-run-webkit-tests - enable cygwin support for chromium win
https://bugs.webkit.org/show_bug.cgi?id=47220

Attachment 70702: Patch
https://bugs.webkit.org/attachment.cgi?id=70702&action=review

------- Additional Comments from Dirk Pranke <dpranke at chromium.org>
This is a revised version of the patch that implements a slow, conservative
version, a fast version, and a version with a long-running cygpath process
open.

It's a bit unclear which version is the best. The fast version is, well, fast,
but may not work correctly in some cygwin configurations. The slow version is
slow (test-webkitpy is almost 2x slower). The long-running cygpath version
seems to work fine but python coughs up an error at the end of running
test-webkitpy that I can't track down a reason for.

The error is something like:

Exception AttributeError: "'NoneType' object has no attribute 'error'" in
<bound
 method Popen.__del__ of <subprocess.Popen object at 0x7e6e966c>> ignored

which makes me think something odd is happening during cleanup at process exit.
This doesn't seem to happen except when running test-webkitpy, which changes
the logging system in such a way as to make me think that this message is
normally suppressed.

At any rate, It would be great if I can get some feedback on the different
approaches. Adam, in particular it would be great if you can confirm that this
patch works on your system and the benchmarking results are similar to mine.