[Webkit-unassigned] [Bug 48236] Add support for loading each test several times in chromium Python test runner

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Tue Oct 26 00:54:16 PDT 2010


https://bugs.webkit.org/show_bug.cgi?id=48236





--- Comment #6 from Søren Gjesse <sgjesse at chromium.org>  2010-10-26 00:54:15 PST ---
(In reply to comment #5)
> What are these flags used for (or what are they going to be used for)? AFAIK, old-run-webkit-tests has a --repeat-each flag, but that's implemented by looping in the script, not in DRT, and doesn't have a --js-flags script.
> 

The main purpose is the enhance testing as we have seen bugs related to V8 which could only be revealed by loading the tests several times. While testing this new option the bug http://code.google.com/p/chromium/issues/detail?id=60135 was discovered.

> Does upstream DRT support these flags, or just Chromium DRT?
> 

No this is a chromium DRT only option for now.

> How does --multiple-loads actually interact with the framework? Do we only dump output for the first load, or the last load, or something?
> 

The --multiple-loads and --js-files are passed unmodified to DRT. Currently DRT will just dump the output for the last load. See https://bugs.webkit.org/show_bug.cgi?id=48233.

> As far as the actual coding of the patch goes ...
> 
> 1) Is it actually possible to pass an option in that preserves the quotation marks, or will you need to actually enclose the option in quotes every time (i.e., can both branches of that test execute)?
> 

Well, only if you really want to e.g. --js-flags="\"--xxx --yyy\"", so I will remove that check.

> 2) I'm not wild about using --multiple-loads=0 to indicate "use the default value". It seems like --multiple-loads=0 should turn off the feature, and I'm not sure what the savings would be over just typing --multiple-loads=5 ?

Good point. I was thinking of having different defaults for release and debug (e.g. 5 and 2), but I could not get optparse of type "int" to handle just --multiple-loads (without equals some number). I have discarded the idea of a default for now.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.


More information about the webkit-unassigned mailing list