OS/2, eCS & ArcaOS - Technical > Programming

Qt6 Application Testing

<< < (173/177) > >>

Paul Smedley:

--- Quote from: David McKenna on May 08, 2025, 12:58:40 pm --- Yes, this one seems to work as well as the previous working version - with the same limitations (mostly occasional random system hangs). Any thought on trying to build the 6.2.12 version of webengine?

--- End quote ---

I'm currently stuck on 6.2.8 - which is the same webengine that came as part of Qt 6.4.x. The code compiles, but pretty much any site with javascript crashes. Any newer than 6.2.8 and we hit the webengine that came with Qt 6.5 - and we encounter https://github.com/bitwiseworks/node-os2/issues/1 where Dmik and I both got stuck at basically the same crash.

So in short, No.

Paul Smedley:

--- Quote from: Dave Yeo on May 08, 2025, 06:01:48 pm ---What I'm wondering is whether there is an equivalent Qt5 Webengine to 6.2.6 and how to merge it.
Trying to clone Qt5 from the official sources, init repository fails,

--- End quote ---

In short....
Qt 6.2.6 - https://github.com/psmedley/qt6-webengine-os2/blob/main/CHROMIUM_VERSION

--- Code: ---Based on Chromium version:                                      94.0.4606.126
Patched with security patches up to Chromium version:           104.0.5112.81
--- End code ---

Qt 5.15.19 (latest tag at https://github.com/qt/qtwebengine/blob/5.15.19/CHROMIUM_VERSION

--- Code: ---Based on Chromium version:                                      87.0.4280.144
Patched with security patches up to Chromium version:           135.0.7049.95
--- End code ---

So the latest 5.15 version is older than what we have. We could however, try and cherry-pick the security fixes.

Dave Yeo:
There's something seriously wrong with code pages and perhaps icu. Using dooble-qt6, simply opening it, displaying the release notes and closing it results in this,

--- Code: ---[107:19:0511/115608.818000:ERROR:text_encoding_registry.cc(139)] alias KS_C_5601-1987 maps to EUC-KR already, but someone is trying to make it map to cp1363
[107:19:0511/115608.818000:ERROR:text_encoding_registry.cc(139)] alias KSC_5601 maps to EUC-KR already, but someone is trying to make it map to cp1363
[107:19:0511/115608.818000:ERROR:text_encoding_registry.cc(139)] alias 5601 maps to EUC-KR already, but someone is trying to make it map to cp1363
[107:19:0511/115608.818000:ERROR:text_encoding_registry.cc(139)] alias windows-949 maps to cp1363 already, but someone is trying to make it map to EUC-KR
[107:19:0511/115608.818000:ERROR:text_encoding_registry.cc(139)] alias KS_C_5601-1989 maps to cp1363 already, but someone is trying to make it map to EUC-KR
[107:19:0511/115608.818000:ERROR:text_encoding_registry.cc(139)] alias csKSC56011987 maps to cp1363 already, but someone is trying to make it map to EUC-KR
[107:19:0511/115608.818000:ERROR:text_encoding_registry.cc(139)] alias korean maps to cp1363 already, but someone is trying to make it map to EUC-KR
[107:19:0511/115608.818000:ERROR:text_encoding_registry.cc(139)] alias iso-ir-149 maps to cp1363 already, but someone is trying to make it map to EUC-KR

--- End code ---

As far as I can see, the release notes are straight ASCII.
I've also seen errors related to ICU.
For Mozilla, had to re-add cp850 at one point when they cleaned up the codepage section.
Doing the same with dooble-qt5 results in no messages to stdout/stderr
Another observation is here, going to github often sees dooble-qt6 close pretty quick. Redirecting stdout/stderr (dooble 2>&1 | tee dooble.log) sees dooble display the pages. Timing issue I assume.

David McKenna:
Dave,

  I see quick hangs or closes in Dooble QT6 with github too. Just thought it was because of the age of our WebEngine, but maybe not?

Regards,

Dave Yeo:
Try running "dooble 2>&1 | tee dooble.log"
Unluckily github has changed its code again and even dooble-qt6 is choking on various stuff. Filed an issue yesterday, https://github.com/textbrowser/dooble/issues/261 and had to use my phone to do the initial post. Luckily can continue the conversation with email.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version