[Releasing] Meeting minutes: Qt release team meeting 12.01.2016
Lars.Knoll at theqtcompany.com
Thu Jan 14 19:24:46 CET 2016
On 14/01/16 18:57, "Thiago Macieira" <thiago.macieira at intel.com> wrote:
>On Thursday 14 January 2016 09:50:51 Knoll Lars wrote:
>> Hi Thiago,
>> Can you explain in more detail what’s blocking 5.7 on QNX? As far as I
>> understand, it’s mainly that std::atomic is not properly implemented in
>> their libstdc++, right?
>Right. QNX folks gave us a patch, but it hasn't been applied to the toolchain
>in the CI yet.
>> Do we have any other options apart from patching the QNX SDK, as I don’t
>> believe that to be a very good option?
>Not easily. I'd have to hunt down the uses of QAtomicPointer that fail to
>compile and replace with QAtomicInteger<quintptr>, then do manual conversions.
>I know only of one in qlogging.cpp, the one that first causes a compilation
>error. There may be more, we just don't know.
>I'd rather apply the fix instead.
Well, this works in 5.6, I guess because we're not using std::atomic there. Why aren't we continuing to use that code path until QNX has released a fixed tool chain?
>Note that they're also working on an updated version of their SDK, which
>should happen soon but not in time for our branching. So this problem will go
>away in time and users won't have to patch for long, if at all (the updated
>SDK could happen before the 5.7 final).
>[The same thing happened back in Qt 4 days for QAtomicPointer<const Type>,
>which caused compilation errors in many platforms. Instead of working around
>in the QtDBus code that used that construct, I insisted on having the atomic
>code (all ours, in src/corelib/arch) be fixed]
>Thiago Macieira - thiago.macieira (AT) intel.com
> Software Architect - Intel Open Source Technology Center
More information about the Releasing