Qt6 Development

[Paul Smedley]

Thanks, Paul... https://html5test.com works now - Nice detective work! Oddly, the QT6 version of Dooble gives a slightly lower score than the QT5 one...

It was comparatively easy once I knew which function was crashing!

I have Dooble installed on linux which is also using Qt 6.2.4 - and am getting 472 points of 555 - on both OS/2 and linux.... What does dooble score on Qt5?

FWIW
Dooble on Qt 6.2.4 under OS/2 - http://html5te.st/affa3a605cef5c6d
Dooble on Qt 6.2.4 under linux - http://html5te.st/d58d0b605d64a0ad

[David McKenna]

Paul,

  I'm proud to say that here Dooble on OS/2 using QT5 gets a score of 522! Dooble using QT6 gets 519... those Linux guys need to step up their game :-)

Regards,

[Paul Smedley]

Hey Dave - are you able to click the 'Save' button and post the URL for your Qt5 result? I'm guessing the difference is due to some codecs that a) I don't have installed on my linux partition, and b) may not yet be working quite right in out Qt6 port.

Edit: actually if you can save both results, and I'll work out how to get my Dooble/2 to your level

[Eugene Tucker]

Dave not fair they don't have Paul Smedly developing for them.

[Paul Smedley]

Dave not fair they don't have Paul Smedly developing for them.

Whilst I appreciate the compliment, in reality, I'm developing very little, I'm just hacking around to get things working. If it wasn't for the efforts of LOTS of people working on Qt and Chromium for other platforms, we'd have nothing.

[Dave Yeo]

I get 519 when using https, and 472 when using http at html5test.com with the security improving from 21/32 to 29/32. I guess the problem with Dooble is it is happy with a HTTP address instead of aggressively trying for https

[Paul Smedley]

I didn't even notice the URL - I'm so used to chrome trying https first..

Dooble under linux Qt 6.2.4 - https - https://html5te.st/ab2700605eccae48 - 519 / 555
Dooble under OS/2 Qt 6.2.4 - https://html5te.st/584092605e962288 - 519 / 555

Dooble under OS/2 Qt 5.15.x - https://html5te.st/903da8605ebf9a6d - 522/555 - the +3 is from 'Web Applications' which is interesting

[TeLLie]

Hi all,
What i try i dont get any output out off the html5test site""

Even with Qt5 dooble it dont give any output

[David McKenna]

Hi Tellie,

  Does it just show a blank page? Are you sure Javascript is enabled? The site needs Javascript to run, but at least here, if not enabled I get a message saying to enable it. Might want to try a new profile too.

Regards,

[TeLLie]

Hi David

I see the this...
And then Notting happens
Also tried a new profile, also JavaScript is enabled.

[David McKenna]

 OK - I see that too briefly, then the score comes up. Are you using any kind of blocker or HOSTS file that might interfere, or proxy? If so, maybe disable temporarily to test...

Regards,

[roberto]

I see the this...

Hi Tellie
You are run http and is https
Saludos

[Paul Smedley]

I see the this...

Hi Tellie
You are run http and is https
Saludos

Both should work - but https will give a better score...

[Paul Smedley]

Started work to enable assembly for openh264 - got the following - not the error I was expecting

Code: [Select]

FAILED: obj/third_party/openh264/openh264_processing_asm/downsample_bilinear.obj
C:/USR/BIN/python2.7.exe ../../../../../src/3rdparty/chromium/build/gn_run_binary.py nasm.exe -DPREFIX -fobj -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/api/svc/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/common/arm/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/common/inc/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/common/src/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/common/x86/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/interface/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/adaptivequantization/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/backgrounddetection/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/common/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/complexityanalysis/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/denoise/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/downsample/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/imagerotate/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/scenechangedetection/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/scrolldetection/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/vaacalc/ -I../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/common/x86/ -I./ -I../../../../../src/3rdparty/chromium/ -Igen/ -DX86_32 -DX86_32_PICASM -MD obj/third_party/openh264/openh264_processing_asm/downsample_bilinear.obj.d -o obj/third_party/openh264/openh264_processing_asm/downsample_bilinear.obj ../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/x86/downsample_bilinear.asm
../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/x86/downsample_bilinear.asm:754: error: short jump is out of range
../../../../../src/3rdparty/chromium/third_party/openh264/src/codec/processing/src/x86/downsample_bilinear.asm:754: warning: byte data exceeds bounds [-w+number-overflow]
nasm.exe failed with exit code 1
ninja: build stopped: subcommand failed.

Note: I haven't even tried modifying "asm_inc.asm" yet to add support for the obj object format - I thought I'd just patch openh264\BUILD.gn to add is_os2 to the conditions to enable assembly and see what happens. Will try spend some time on this later this week...

[Dave Yeo]

Short jump is a 16 bit thing, staying in the same segment. Need USE32