Nico Weber
Nico Weber
Red CI is unrelated (`Fail: /Users/runner/work/1/s/Tests/LibWeb/Text/input/clipboard.html`).
Lagom_Linux_NoFuzz redness is failing for a different reason that's also unrelated to this PR.
Here's a before/after for the second commit. Before: After:
(Why does this add a new DateTime struct instead of using Core::DateTime? Because Core::DateTime does a conversion to a timestamp, which I'm explicitly trying not to do. The printing code...
Red CI is unrelated.
Before: After: Before: After:
Reduced repro: [umlaut.pdf](https://github.com/SerenityOS/serenity/files/14373857/umlaut.pdf)
I went through a few PDFs to look at CCITTs, and they decode without a diag but look wrong (all 1s) for at least these: Build/lagom/bin/pdf --render out.png /Users/thakis/Downloads/0000/0000031.pdf Build/lagom/bin/pdf...
Here are a few that do work and look much better than before: Build/lagom/bin/pdf --render out.png /Users/thakis/Downloads/0000/0000769.pdf --page 6 Build/lagom/bin/pdf --render out.png /Users/thakis/Downloads/0000/0000769.pdf --page 7 Build/lagom/bin/pdf --render out.png /Users/thakis/Downloads/0000/0000711.pdf --page...
…looks like the ones that don't decode right are exactly those with `row == 0` (which I think means "keep decoding rows while you have data"). So maybe we can...