[webkit-dev] More thoughts on cleaning up the root directory
levin at chromium.org
Mon Dec 27 19:35:38 PST 2010
On Mon, Dec 27, 2010 at 7:17 PM, Maciej Stachowiak <mjs at apple.com> wrote:
> On Dec 27, 2010, at 1:11 PM, Sam Weinig wrote:
> On Mon, Dec 27, 2010 at 11:14 AM, Adam Barth <abarth at webkit.org> wrote:
>> > Is it really a good idea to move platform out of WebCore? Lots of stuff
>> > there seems quite WebCore related.
>> There seem to be a couple people who aren't sold on moving platform
>> out of WebCore. It sounds like we should hold off on doing that and
>> discuss it separately down the road.
>> On Mon, Dec 27, 2010 at 2:47 AM, Hajime Morita <morrita at google.com>
>> >> Platform/ (was WebCore/platform)
>> > I'd like to keep platform directory under WebCore if there is no strong
>> Ok. I think different people have slightly different ideas about what
>> should go into this folder. That sounds like a complex topic that we
>> might need to discuss more later.
> I think moving Platform out from WebCore is great long term goal, but right
> now, there is simply too many layering violations for it to be feasible. For
> those curious, the intent is for nothing in Platform to be dependent on
> anything else in WebCore (eg. dom, html, rendering, loader), so something
> like platform/qt/RenderThemeQt.cpp would be considered a layering violation.
> There are bugs filed on many of these violations, but the work has not be
> Indeed, that's the reason I suggested to Adam that we should move platform
> out of WebCore eventually. It would make the layering intent more clear and
> would let us enforce the layering by making it a compile-time error to
> depend on other parts of WebCore inside the platform directory.
fwiw, this change
finished would help a lot with enforcing it. (The enforcement in there is
likely too aggressive as previously discussed).
> However, I don't think we should make this change part of the initial
> reorg. It's something we could do down the line once we have had time to fix
> up more of the layering violations. Note: we could also whitelist specific
> files with known layering violations if we want to make this change before
> we eliminate all layering violations.
> webkit-dev mailing list
> webkit-dev at lists.webkit.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the webkit-dev