[Development] Regression in QtQml

Robin Burchell robin+qt at viroteck.net
Mon Jul 16 08:09:23 CEST 2012


Ah, ok! I didn't even think to check that. Perhaps the message should
make a mention of qmldir files, or something?

Thanks, and sorry to waste your time.

On Mon, Jul 16, 2012 at 12:53 AM,  <matthew.vogt at nokia.com> wrote:
> Hi Robin.
>
> The problem appears to be that the file 'imports/com/Nokia/extras/qmldir' contains numerous entries which are repetitions of earlier entries.
> As of the change you identified, this condition is now reported as an error - the old behavior was to import the first entry encountered, and to silently ignore the remainder.
>
> Thanks,
> Matt
>
>> -----Original Message-----
>> From: robin at viroteck.net [mailto:robin at viroteck.net] On Behalf Of ext
>> Robin Burchell
>> Sent: Monday, July 16, 2012 5:46 AM
>> To: Vogt Matthew (Nokia-MP/Brisbane)
>> Cc: development at qt-project.org
>> Subject: Regression in QtQml
>>
>> Hi,
>>
>> I was porting the QtQuick2 branch of MeeGo components to track recent Qt
>> 5 changes (removal of QT += declarative, old plugin loaders, etc).
>> I got it building, but I was running into errors running
>> examples/meego/QmlComponentGallery about things having "already been
>> defined".
>>
>> Upon further inspection, the error appears to have been caused by
>> 57e3325affbe8bbb8edcc2c6d072db28aee46438.
>>
>> I don't know precisely why this is, although this loop looks a bit strange to
>> me:
>>
>>                 ConstIterator cend = qmldircomponents.constEnd();
>>                 for (ConstIterator cit = qmldircomponents.constBegin(); cit != cend;
>> ++cit) {
>>                     for (ConstIterator cit2 = qmldircomponents.constBegin(); cit2 != cit;
>> ++cit2) {
>>                         if ((cit2->typeName == cit->typeName) &&
>>                             (cit2->majorVersion == cit->majorVersion) &&
>>                             (cit2->minorVersion == cit->minorVersion)) {
>>                             // This is entry clashes with a predecessor
>>
>> That is, two loops wandering over the same container's start, is a little
>> strange, but I don't know whether this is really wrong.
>>
>> Can anyone (Matthew?) take a look into this and tell me what the
>> appropriate fix is? For now, I've reverted this commit locally.
>>
>> My components tree is at
>> https://github.com/nemomobile/qt-components/tree/qtquick2 - and can be
>> configured with (for example) ./configure -no-xdamage -meego which
>> should work on OS X.
>>
>> BR,
>>
>> Robin



More information about the Development mailing list