RESOLVED FIXED 119266
Add back c++11 features removed by buildfixes after all ports did the switch
https://bugs.webkit.org/show_bug.cgi?id=119266
Summary Add back c++11 features removed by buildfixes after all ports did the switch
Csaba Osztrogonác
Reported 2013-07-30 09:21:48 PDT
Patch is coming soon.
Attachments
Patch (12.04 KB, patch)
2013-07-30 10:35 PDT, Csaba Osztrogonác
no flags
Patch (12.71 KB, patch)
2013-07-30 10:42 PDT, Csaba Osztrogonác
no flags
Patch (12.71 KB, patch)
2013-07-30 11:12 PDT, Csaba Osztrogonác
no flags
Patch (14.53 KB, patch)
2013-08-01 10:31 PDT, Csaba Osztrogonác
no flags
Patch (6.41 KB, patch)
2013-09-12 04:58 PDT, Csaba Osztrogonác
no flags
Patch (5.36 KB, patch)
2013-09-12 05:11 PDT, Csaba Osztrogonác
no flags
Csaba Osztrogonác
Comment 1 2013-07-30 10:35:06 PDT
Created attachment 207746 [details] Patch WIP patch for EWS bots to check if all ports are ready for c++11 in JSC
Early Warning System Bot
Comment 2 2013-07-30 10:40:05 PDT
Early Warning System Bot
Comment 3 2013-07-30 10:40:42 PDT
EFL EWS Bot
Comment 4 2013-07-30 10:42:08 PDT
Csaba Osztrogonác
Comment 5 2013-07-30 10:42:42 PDT
Created attachment 207747 [details] Patch WIP patch for EWS bots, let's try to enable c++11 for whole QtWebKit
EFL EWS Bot
Comment 6 2013-07-30 10:55:41 PDT
Csaba Osztrogonác
Comment 7 2013-07-30 11:12:50 PDT
Created attachment 207750 [details] Patch one more round after killing the Qt EWS bots :) (clean build worked for me locally)
EFL EWS Bot
Comment 8 2013-07-30 11:20:44 PDT
Allan Sandfeld Jensen
Comment 9 2013-07-30 16:06:06 PDT
(In reply to comment #7) > Created an attachment (id=207750) [details] > Patch > > one more round after killing the Qt EWS bots :) (clean build worked for me locally) Did anyone actually enable C++11 on Qt?
Csaba Osztrogonác
Comment 10 2013-07-30 23:56:31 PDT
(In reply to comment #9) > (In reply to comment #7) > > Created an attachment (id=207750) [details] [details] > > Patch > > > > one more round after killing the Qt EWS bots :) (clean build worked for me locally) > > Did anyone actually enable C++11 on Qt? No, I tried to enable it by the uploaded patch. But unfortunately the EWS bot couldn't handle this situation and needed a clean build.
Csaba Osztrogonác
Comment 11 2013-08-01 10:31:26 PDT
Created attachment 207935 [details] Patch updated patch for EWS bots, Qt fix removed (Qt switched to c++11), cmake fixes added to make EFL bots happy
Build Bot
Comment 12 2013-08-11 03:30:08 PDT
Allan Sandfeld Jensen
Comment 13 2013-08-13 01:47:53 PDT
(In reply to comment #11) > Created an attachment (id=207935) [details] > Patch > > updated patch for EWS bots, Qt fix removed (Qt switched to c++11), cmake fixes added to make EFL bots happy Actually I would prefer if we didn't add these back. We might have enabled C++11, but using nullptr limits WebKit from GCC 4.3+ to GCC 4.6+, and considering how little using nullptr brings, I would prefer we avoid them.
Allan Sandfeld Jensen
Comment 14 2013-08-13 02:29:08 PDT
(In reply to comment #13) > (In reply to comment #11) > > Created an attachment (id=207935) [details] [details] > > Patch > > > > updated patch for EWS bots, Qt fix removed (Qt switched to c++11), cmake fixes added to make EFL bots happy > > Actually I would prefer if we didn't add these back. > > We might have enabled C++11, but using nullptr limits WebKit from GCC 4.3+ to GCC 4.6+, and considering how little using nullptr brings, I would prefer we avoid them. Btw, the reason support older GCC still makes sense is mostly Red Hat. Since Debian stable updated earlier this year, it is now RHEL6 that restraining gcc on linux. The latest Red Hat Enterprise Linux ships with GCC 4.4.
Csaba Osztrogonác
Comment 15 2013-08-20 03:33:56 PDT
(In reply to comment #13) > (In reply to comment #11) > > Created an attachment (id=207935) [details] [details] > > Patch > > > > updated patch for EWS bots, Qt fix removed (Qt switched to c++11), cmake fixes added to make EFL bots happy > > Actually I would prefer if we didn't add these back. > > We might have enabled C++11, but using nullptr limits WebKit from GCC 4.3+ to GCC 4.6+, and considering how little using nullptr brings, I would prefer we avoid them. It's all the same for me, but it would be great to reply to the original webkit-dev message about it. Otherwise I don't think if there is any buildbot or EWS with GCC 4.4 to ensure nobody adds nullptr again and again.
Csaba Osztrogonác
Comment 16 2013-09-12 04:58:48 PDT
Created attachment 211419 [details] Patch updated patch to ToT, Qt and EFL buildsystem hacks aren't needed, both of them builds with c++11
Build Bot
Comment 17 2013-09-12 05:02:11 PDT
EFL EWS Bot
Comment 18 2013-09-12 05:02:13 PDT
Early Warning System Bot
Comment 19 2013-09-12 05:02:52 PDT
EFL EWS Bot
Comment 20 2013-09-12 05:04:05 PDT
Csaba Osztrogonác
Comment 21 2013-09-12 05:04:11 PDT
I grepped to nullptr and it seems there are zillion everywhere: - 8 nullptr in JavaScriptCore - 71 nullptr in WTF - 383 nullptr in WebCore - 99 nullptr in WebKit2 So I don't see why shouldn't we add them back.
Early Warning System Bot
Comment 22 2013-09-12 05:04:21 PDT
Csaba Osztrogonác
Comment 23 2013-09-12 05:11:04 PDT
Zan Dobersek
Comment 24 2013-09-12 08:09:04 PDT
If you're taking requests here, feel free to also include the std::nullptr reintroduction in ChildProcessProxy in bug #109162.
Csaba Osztrogonác
Comment 25 2013-09-12 08:27:17 PDT
Comment on attachment 211421 [details] Patch Clearing flags on attachment: 211421 Committed r155613: <http://trac.webkit.org/changeset/155613>
Csaba Osztrogonác
Comment 26 2013-09-12 08:27:25 PDT
All reviewed patches have been landed. Closing bug.
Note You need to log in before you can comment on or make changes to this bug.