mara004

I did `gh pr checkout 396` and ran the test suite. It currently shows failures: ``` FAILED tests/test_cli.py::test_images - TypeError: PdfObject.__init__() got an unexpected keyword argument 'textpage' FAILED tests/test_cli.py::test_pageobjects[images] -...

Oh, I think I see the problem. When `PdfObject` doesn't resolve to `PdfTextObj`, the `textpage` argument is not accepted. I hadn't thought about that...

FYI, I filed https://issues.chromium.org/issues/464315000 for this

I see... The [Readme](https://github.com/bwesterb/py-demandimport?tab=readme-ov-file#attribution) says Matt Mackall is the original author.[^1] Has anyone yet attempted to ask them whether they would be willing to relicense this portion of code? [^1]:...

Regardless of viewer issues, I feel like image transparency should generally be handled via `SMask`s instead of encoding to JP2. The JP2 alpha in PDF tends to be a niche...

> > > The practical downside to this is that https://github.com/python-pillow/Pillow/pull/8097 increases the PDF size of this image's issue from 15,721 bytes to 26,604 bytes. > > providing file sizes...

@sandeepb2013 Sorry for the late reply, I missed this thread. > raised concern regarding the same with PDFium community. > [https://github.com/pypdfium2-team/pypdfium2/issues/313] I think you may have mixed up pypdfium2 and...

> @mara004 Any findings regarding this issue? I'm not aware of any activity on this, and searching pdfium's bug tracker for "s390x" did not yield any entries. mgiessing recently created...

There are good news: apparently, @tiran has managed to build pdfium on s390x natively: https://issues.chromium.org/issues/402282789 https://issues.chromium.org/issues/402354437 https://github.com/tiran/libpdfium/

It looks like this is not just hardcoding new platform handlers as in #187, but finally a more portable approach. The shortcomings of pdfium's default non-portable, bloated build setup have...