From: =?iso-8859-1?Q?Ga=EBl?= PORTAY Date: Fri, 2 Mar 2018 06:49:01 -0500 Subject: [Buildroot] [PATCH 1/1] qt5webkit: fix build issue with 32-bits armv8-a In-Reply-To: <20180302112652.GB8805@momiji> References: <20180302105833.6828-1-gael.portay@savoirfairelinux.com> <20180302112652.GB8805@momiji> Message-ID: <20180302114901.s7crdeo6tinrn6um@archlinux> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Hi Adrian, On Fri, Mar 02, 2018 at 11:26:52AM +0000, Adrian Perez de Castro wrote: > > The patch is an adaptation of an upstream fix in version 5.212 of > > qtwebkit[1]. > > I fixed this upstream a while ago: > > https://bugs.webkit.org/show_bug.cgi?id=174425 > > Note that you might need also the following patch for bmalloc: > > https://bugs.webkit.org/show_bug.cgi?id=174424 > > For WebKitGTK+ those patches were merged upstream since version 2.16.6; so > I expect that newer versions of qt5webkit based on some upstream WebKit > release from around July 2017 should be including the two patches linked > above. > It seems your modifications has not yet landed into qt5webkit (even in 5.212). > That being said, I understand that your patch here targets versions of > qt5webkit based on *older* WebKit revisions, right? That's fine of course > and I think it would be desirable to check both for __ARM_ARCH_8__ and > __ARM_ARCH_8A__ like upstream WebKit does: > > #if defined(__ARM_ARCH_8__) || defined(__ARM_ARCH_8A__) > // ... > I will add this double check so the fix is closer that the upstream version of WebKit. > Of course if qt5webkit does NOT use bmalloc, then you do not need the > corresponding fix for it ? sorry, I don't know what's the qt5webkit situation > in this regard, I have not been following it much since the Qt port was > removed from the upstream WebKit source tree. > It looks like qt5webkit does not use bmalloc. > > Reviewed-by: Adrian Perez de Castro > > Patch looks good to me, but please double-check the comments above, in > particular the situation with bmalloc which might need a similar fix. > Thanks for your review.