Affected versions

• 92.0a1 (2021-07-27)
• 91.0b7
• 90.0.2

Affected platforms

• macOS 11.5
• macOS 10.15

Steps to reproduce

1. Open https://bug1260585.bmoattachments.org/attachment.cgi?id=8736063
2. Hit CTRL + P
3. Save to Pdf or print the page to paper
4. Observe the output

Expected result

• The output is rendered correctly

Actual result

• The output is not rendered correctly

Regression range

• Last good revision: d9e30da70def144c22f40e7debc5b190a31eb1f1
• First bad revision: c00df77af50e0703d37188be7c57032713e6efec
• Pushlog:
  https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=d9e30da70def144c22f40e7debc5b190a31eb1f1&tochange=c00df77af50e0703d37188be7c57032713e6efec

Other notes

• Reproducible with the old modal too
• Windows and Ubuntu are not affected

There are a couple of printing changes in there, but I'm guessing this is a regression from bug 739096.

This looks like a pretty extreme torture-test PDF (note how slowly that graph renders to the screen...). I guess the cairo-quartz backend must be failing in some way. I'll see if I can narrow down what's happening at all.

I think we can safely assume this is related to the cairo update.

Jonathan, should that be moved to core:graphics?

I suppose so, though AFAIK we're only keeping cairo there at all for the sake of printing output. Still, I guess the actual issue here is (probably) within the graphics library rather than in the gecko printing code that makes use of it.