1311612 - Assert in gfxFontGroup։։GetDefaultFont on x86

Status: VERIFIED WORKSFORME

(Core :: Graphics: Text, defect, P3)

Description

[SkyLined]

Attachment: repro.svg

The following SVG file will cause an assertion failure on x86 Windows. The assertion does not trigger on x64 Windows.

<svg xmlns="http://www.w3.org/2000/svg" font-size-adjust="1" font-size="1%" style="font-family:crash;box-shadow:2ex 2ex;transition:4s"/>

I don't know exactly why all the font attributes and styles are needed, but apparently they create some kind of edge case that's not handled.

The assert triggered should be here in the source:
https://dxr.mozilla.org/mozilla-central/source/gfx/thebes/gfxTextRun.cpp#1847

Comment 1

[Sotaro Ikeda [:sotaro]]

SkyLined, can you post the actual assert log?

Comment 2

[Sotaro Ikeda [:sotaro]]

:jfkthame, can you comment to the bug?

Comment 3

[SkyLined]

Sotaro: sure, just let me know where I can grab that and I will add it.

Comment 4

[Jonathan Kew [:jfkthame]]

I haven't been able to reproducet this locally (I don't currently have a 32-bit Windows system on hand) to investigate further.

I guess it may be related to the extremely small font size causing something to fail? It might also depend whether we're using the GDI or DirectWrite backends.

Could you please attach the about:support info for the system where it fails? Thanks.

Comment 5

[SkyLined]

Attachment: about support.txt

As requested, the about:support info for the VM on which it crashes.

Comment 6

[Jonathan Kew [:jfkthame]]

OK, this is a VM where graphics acceleration is blocked, and so we end up using the GDI font backend. If you go into about:config and set gfx.font_rendering.directwrite.force-enabled to true, and then restart the browser, does the issue still occur or does that avoid it?

Comment 7

[SkyLined]

Changing that setting does not affect the outcome.

Comment 8

[Rares Doghi, Desktop QA]

I tried to reproduce this issue on a 32bit Windows 10 machine with Intel HD Graphics but it doesnt occur in our latest Release version Fx 85.0 I will close this issue as Works For Me for now but if the issue reoccurs on the reporter side please reopen it.

Comment 9

[Rares Doghi, Desktop QA]

@Skylined does this issue still occur on your end ? and if so are there any steps that might help us reproduce this issue on our side ?

Comment 10

[SkyLined]

No, it no longer crashes on my machine.