1269011 - Lots of "SkCairoFTTypeface::onFilterRec unimplemented" spam during pdf.js debug test runs

Status: RESOLVED FIXED

(Core :: Graphics, defect)

Description

[Ryan VanderMeulen [:RyanVM]]

I haven't gone through the logs from other test suites to see how prevalent this is elsewhere, but at least in the pdf.js tests, there's a *LOT* of "SkCairoFTTypeface::onFilterRec unimplemented" spam.

What does it mean and can we somehow shut it up?

http://archive.mozilla.org/pub/firefox/try-builds/[e-mail]-bfecb53c1626b14dab001068ddfe3fd686854802/try-linux64-debug/try_ubuntu64_vm-debug_test-mochitest-browser-chrome-4-bm131-tests1-linux64-build198.txt.gz

Comment 1

[Ryan VanderMeulen [:RyanVM]]

15,111 instances in that log, FWIW. Maybe Eric could speak to whether it affects other tests as well.

Comment 2

[Eric Rahm [:erahm]]

(In reply to Ryan VanderMeulen [:RyanVM] from comment #1)
> 15,111 instances in that log, FWIW. Maybe Eric could speak to whether it
> affects other tests as well.

Looks like just pdf.js and devtools, full details below:

> 35672 SkCairoFTTypeface::onFilterRec unimplemented

This warning shows up in the following test suites:

>  16168 - [TC] Linux64 mochitest-browser-chrome-4 bc4
>  15336 - [TC] Linux64 mochitest-browser-chrome-e10s-2 bc2
>   2689 - [TC] Linux64 mochitest-devtools-chrome-2 dt2
>    812 - [TC] Linux64 mochitest-devtools-chrome-6 dt6
>    165 - [TC] Linux64 reftest-e10s-2 R-e10s2
>    165 - [TC] Linux64 reftest-3 R3
>     78 - [TC] Linux64 mochitest-devtools-chrome-3 dt3
>     50 - [TC] Linux64 web-platform-tests-e10s-1 1
>     50 - [TC] Linux64 web-platform-tests-1 1
>     48 - [TC] Linux64 mochitest-devtools-chrome-7 dt7
>     24 - [TC] Linux64 mochitest-gl gl
>     18 - [TC] Linux64 reftest-e10s-5 R-e10s5
>     16 - [TC] Linux64 mochitest-chrome-1 c1
>     14 - [TC] Linux64 reftest-8 R8
>      9 - [TC] Linux64 mochitest-2 2
>      9 - [TC] Linux64 mochitest-plain-e10s-2 2
>      8 - [TC] Linux64 reftest-7 R7
>      5 - [TC] Linux64 crashtest C
>      5 - [TC] Linux64 crashtest e10s C
>      1 - [TC] Linux64 reftest-5 R5
>      1 - [TC] Linux64 mochitest-a11y a11y
>      1 - [TC] Linux64 reftest-e10s-3 R-e10s3

It shows up in 253 tests. A few of the most prevalent:

>   9945 -        browser/extensions/pdfjs/test/browser_pdfjs_zoom.js
>   9945 - [e10s] browser/extensions/pdfjs/test/browser_pdfjs_zoom.js
>   6223 -        browser/extensions/pdfjs/test/browser_pdfjs_navigation.js
>   5391 - [e10s] browser/extensions/pdfjs/test/browser_pdfjs_navigation.js
>    812 -        devtools/client/shared/test/browser_flame-graph-05.js
>    162 -        devtools/client/performance/test/browser_perf-console-record-08.js
>    158 -        devtools/client/performance/test/browser_perf-recording-selected-04.js
>    100 -        devtools/client/performance/test/browser_perf-console-record-06.js
>     67 -        devtools/client/performance/test/browser_perf-recording-selected-03.js
>     63 -        devtools/client/performance/test/browser_perf-loading-02.js

Comment 3

[Lee Salzman [:lsalzman]]

Attachment: silence SkCairoFTTypeface::onFilterRec warnings

The onFilterRec method is only used to override settings in the scaler context (such as hinting and formats). But since this is our own custom font host and we do most of the hinting and other setting tweaking to the SkPaint in DrawTargetSkia, it should be safe for now that this is empty.

Hence, let us just remove this warning about it and make the world a better place for developers.

Comment 4

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/3166fa3db323

Comment 5

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/3166fa3db323