Guillaume Ayoub

It’s probably a bug in Cairo: https://gitlab.freedesktop.org/cairo/cairo/-/issues/530

> 2 years later, still looking forward to seeing this fix merged! You can read [this blog article](https://www.courtbouillon.org/blog/00023-the-future-of-cairosvg) to understand why this pull request is not merged yet.

> Hi, I had the same issue since exported image.png likely jpeg compressed . Hi! What you see is not caused by JPEG compression, it caused by [anti-aliasing](https://en.wikipedia.org/wiki/Spatial_anti-aliasing). If you...

> I think the above should close this issue. If someone wants to open a PR with code that’s equivalent to [what’s done in WeasyPrint](https://github.com/Kozea/WeasyPrint/blob/1c1365639abfd09f4ef0f5148aeb8b0ea51578fe/weasyprint/text/ffi.py#L402), we will probably merge it.

The documentation will be rewritten for 2.0.

Hi! Thanks for the report. Two (!) pull requests have been opened, but I don’t know if we’ll manage to get an easy and clean fix.

We already tried to fix this problem in https://github.com/Kozea/cairocffi/issues/85, but it looks like it’s back. We sometimes have the problem on GitHub Actions, but only on Windows. If it fails...

> @MeggyCal is my colleague at SUSE, so obviously this has nothing to do with Windows. Yes, I’ve seen the tracebacks, and that’s really sad because I didn’t see this...

Hi! Thanks for the report. Instead of trying to find wrong data lengths and stride values, we should rely on Cairo. Removing the condition is not enough, as we want...

Hi! Thanks for opening this issue. Could you please open a pull request to fix this link?