. Dezember 2014 um 05:46 Uhr
Von: Chris Adams chris.ad...@qinetic.com.au
An: Gunnar Roth gunnar.r...@gmx.de
Cc: Simon Hausmann simon.hausm...@theqtcompany.com,
development@qt-project.org development@qt-project.org
Betreff: Re: Re: [Development] QML instantiation performance
Hi Gunnar,
Can you get
: Chris Adams chris.ad...@qinetic.com.au
An: Gunnar Roth gunnar.r...@gmx.de
Cc: Simon Hausmann simon.hausm...@theqtcompany.com,
development@qt-project.org development@qt-project.org
Betreff: Re: Re: [Development] QML instantiation performance
Hi Gunnar,
Can you get a backtrace which might
Hi Gunnar,
But what about the QQml_removeValueTypeProvider: was asked to remove
provider 0x2b4c2ac but it was not found
messages? Don't you get that with 5.4? Did you change something in your patch
for 5.4?
I get this on the 5.3.1 (not on 5.1.1) on my linux, right at the end of test:
PASS
);
+}
+
QT_END_NAMESPACE
*Gesendet:* Montag, 01. Dezember 2014 um 05:51 Uhr
*Von:* Chris Adams chris.ad...@qinetic.com.au
*An:* Simon Hausmann simon.hausm...@theqtcompany.com
*Cc:* development@qt-project.org development@qt-project.org
*Betreff:* Re: [Development] QML instantiation performance
See
-project.org
Betreff:Re: [Development] QML instantiation performance
See https://codereview.qt-project.org/#/c/101048/ - huge thanks to Matt Vogt for his help with debugging and fixing the issues.
Juha, does that resolve the issues for you? Gunnar Roth, does that fix the issues on Windows? I havent
Juha, does that resolve the issues for you? Gunnar Roth, does that fix the
Yes it does, tests pass on 5.3.1. Big thanks Chris and Matt!
Going back to the original issue I'd need reference from 5.1.1 where
tests won't run, I get same issue as Gunnar:
file:///data/item.1.qml:3:1: Item is not a
I don't recall, but a quick grep of the git log turned up the following
three changes:
https://codereview.qt-project.org/#/c/60447/
https://codereview.qt-project.org/#/c/60446/
https://codereview.qt-project.org/#/c/60445/
I guess this was the most important bit:
The loop was due to the QtQuick2 plugin's defineModule() code initializing
the value type providers on load, but not deinitializing them on dtor,
resulting in future loads of that plugin adding the same ptr to the linked
list as the head's next ptr (so, cycle).
The attached diff fixes that issue,
See https://codereview.qt-project.org/#/c/101048/ - huge thanks to Matt
Vogt for his help with debugging and fixing the issues.
Juha, does that resolve the issues for you? Gunnar Roth, does that fix the
issues on Windows? I haven't tested there.
www.qinetic.com.au - Qt And QML User Experience
Hi Gunnar,
Note that the results between versions may not be directly comparable.
The benchmark does some very specific things, including destroying the
engine and clearing all type registrations between each run. This is
intended to simulate application startup performance conditions, but is
Hi.
on windows that is even worse as the
qtdeclarative/tests/benchmarks/qml/librarymetrics_performance benchmark is
crashing in many testcases. 5.2.1 runs some more cases than 5.3.2 and 5.4
beta.
the test cases which crash in 5.3.2 and 5.4 beta are:
//QTest::newRow(039) listView - with
Hi,
No worries. Interestingly, when I ran the benchmark against 5.3, I found
that it hangs after performing several tests.
Before each test, the benchmark will clear the QML type registrations to
ensure that the cached compiled type data isn't re-used during the run,
tainting the results. It
Hello,
It appears that the QQuickValueTypeProvider instances are continuously
recreated (spinning at 100% and locking the incubator thread, which the main
thread is waiting on). Lars or Simon, are you able to guess what might
cause this issue? The implementation of qmlClearTypeRegistrations
On Friday 28. November 2014 12.20.29 Chris Adams wrote:
Hi,
No worries. Interestingly, when I ran the benchmark against 5.3, I found
that it hangs after performing several tests.
Before each test, the benchmark will clear the QML type registrations to
ensure that the cached compiled type
Hello,
apologies for cross-posting. I find this such a fundamental issue that I feel
at least having the best possible understanding of it worth it (if any
exists, which is also
a valuable information in itself). I am happy to help if there is
anything I could check.
thanks,
Juha
---
Hello,
Hi Juha,
For some more light on this issue, are you able to run the
qtdeclarative/tests/benchmarks/qml/librarymetrics_performance benchmark on
both 5.1 and 5.3 (you may have to backport to 5.1 as IIRC I made some
changes to the way the components were instantiated for 5.2 before the
benchmark was
Hey Chris,
Thanks heaps. I'll have a look and get back with any findings (may
take a day or two before I get to it).
cheers,
Juha
2014-11-27 3:51 GMT+02:00 Chris Adams chris.ad...@qinetic.com.au:
Hi Juha,
For some more light on this issue, are you able to run the
17 matches
Mail list logo