[Webkit-unassigned] [Bug 95346] Get rid of device/host clock synchronization.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=95346


Peter Beverloo <peter at chromium.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |peter at chromium.org




--- Comment #5 from Peter Beverloo <peter at chromium.org>  2012-08-30 05:56:27 PST ---
(In reply to comment #4)
> (In reply to comment #3)
> > Note that you can use self.path_from_chromium_base() in the python code instead of relying on CHROME_SRC. That will work reliably in both a webkit-in-chromium or chromium-in-webkit checkout.
> 
> Nice. Going back to your comment Peter, can I use what Dirk suggested to include the android_commands module and use AndroidCommands.PushIfNeeded() or do we still really want to avoid this dependency (which would make this patch much larger)? My understanding is that if we have a way to fetch chromium's path, it might not be totally alien to depend on some code coming from it.

There are a number of similar dependencies WebKit has on Chromium already. The implication of starting to rely on test-runner scripts is that refactoring / "API" changes would need the new method to be introduced on the Chromium side, then usage of the old versions removed within WebKit, before the old code can be removed. A cross-project three-way patch is significantly more painful than a single patch.

Given the amount of random build breakages due to Chromium infrastructure changes recently, we should try to minimize the number of dependencies until the code stabilizes. While android_commands.py probably will be more stable, our branch still has a 165 line diff and it's not unthinkable more refactoring will happen too.

As discussed on chat, adding a dependency to md5sum in DumpRenderTree's gyp configuration is perfectly fine. It'll introduce some additional code for now, but we'll be able to clean that up later again.

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