[Webkit-unassigned] [Bug 41842] Add feature detection support for NRWT

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=41842


Eric Seidel <eric at webkit.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #64576|review+                     |review-
               Flag|                            |




--- Comment #20 from Eric Seidel <eric at webkit.org>  2010-08-17 19:36:00 PST ---
(From update of attachment 64576)
WebKitTools/Scripts/webkitpy/layout_tests/port/base.py:662
 +      def _path_to_library(self):
This should really be _path_to_webcore, since these symbols are searched for in WebCore, not WebKit on ports which don't build all-as one.

WebKitTools/Scripts/webkitpy/layout_tests/port/webkit.py:228
 +              symbol_list = os.popen(driver_path + " --print-supported-features 2>&1").readlines()
We actually want to move to this model for all ports.

WebKitTools/Scripts/webkitpy/layout_tests/port/webkit.py:221
 +      def _get_supported_features_symbols(self):
This name doesnt' make a lot of sense since it's not symbols on win32.  Seems like we should have a separate path for ports which use symbols vs. supported features, no?

WebKitTools/Scripts/webkitpy/layout_tests/port/webkit.py:239
 +          if sys.platform in ('cygwin', 'win32'):
Yeah, these are tottally separate code paths.  Lets not conflate them.

One unit test is not sufficient for this complexity of code.

This looks fine in general.  But please separate out the runtime list lookup vs. the symbol lookup. Basically we should try running "print-supported-features" and if that erros out, fall back to nm on all platforms (would be my suggestion).

Thank you so much for looking at this!

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