The qtwebengine-5.15.13_p20240322 ebuild is missing the ppc64 patches present in the qtwebengine-5.15.12_p20240122 ebuild, and can therefore not be built on ppc64. With the removal of the qtwebengine-5.15.12_p20240122 ebuild, the package is now left in a [??] status on ppc64 where it can not be rebuilt or upgraded. However, this is simple enough to fix: The ppc64 code (and related patches) in the old ebuild can be used as they are, and still work. I have tested this (on big endian, so also including my own patches from https://github.com/zeldin/chromium_be/tree/master/qtwebengine-5.15.11_p20231102) with Calibre. Reproducible: Always Steps to Reproduce: 1.Try to rebuild or update dev-qt/qtwebengine on ppc64 2. 3. Actual Results: The package can not be rebuild because qtwebengine-5.15.12_p20240122.ebuild was removed, and it can not be updated because qtwebengine-5.15.13_p20240322.ebuild is missing the ppc64 fixes. Expected Results: The package can be updated to version 5.15.13_p20240322
Created attachment 892498 [details] A fixed ebuild for qtwebengine-5.15.13_p20240322 with the ppc64 fixed put back
It was dropped because of bug 924936 where nobody was maintaining them. If you're willing to step up, we can reconsider...
> bug 924936 Ah. I have a fix for that in my personal patchset. I could step up if you want; I'd supply patches for big endian as well in that case.
Created attachment 892499 [details, diff] Fix for bug 924936
A new datapoint: dev-qt/qtwebengine-5.15.16_p20241115 builds and runs perfectly on ppc64 (big endian) with _exactly_ the same patchset that I have been using since qtwebengine-5.15.12_p20240122. So practically zero maintenance burden. The only thing I had to do (in addition to carrying _all_ the ppc64 patches instead of just the big endian ones in /etc/portage/patches) was to put the "# we need to generate ppc64 stuff because upstream does not ship it yet" block back in qtwebengine-5.15.16_p20241115.ebuild.
