Bug 146194 - ASSERTION FAILED: typesettingFeatures & (Kerning | Ligatures) in WebCore::applyFontTransforms
Summary: ASSERTION FAILED: typesettingFeatures & (Kerning | Ligatures) in WebCore::app...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: WebCore Misc. (show other bugs)
Version: 528+ (Nightly build)
Hardware: PC Linux
: P2 Normal
Assignee: Myles C. Maxfield
URL:
Keywords:
: 150638 (view as bug list)
Depends on:
Blocks:
 
Reported: 2015-06-21 08:45 PDT by Michael Catanzaro
Modified: 2015-10-29 02:09 PDT (History)
5 users (show)

See Also:

Attachments
backtrace (264.94 KB, text/plain)
2015-06-21 08:45 PDT, Michael Catanzaro 		no flags Details
Patch (7.50 KB, patch)
2015-09-09 13:49 PDT, Myles C. Maxfield 		no flags Details | Formatted Diff | Diff
Patch (9.50 KB, patch)
2015-09-09 14:40 PDT, Myles C. Maxfield 		dino: review+
Details | Formatted Diff | Diff
Show Obsolete (1) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Michael Catanzaro 2015-06-21 08:45:26 PDT 
Created attachment 255326 [details]
backtrace

ASSERTION FAILED: typesettingFeatures & (Kerning | Ligatures)

In WebCore::applyFontTransforms, at ../../Source/WebCore/platform/graphics/WidthIterator.cpp:102, loading www.vox.com on a debug build of master, GTK port. Seems to happen 100%. Backtrace attached.
Comment 1 Carlos Alberto Lopez Perez 2015-09-04 10:29:25 PDT 
This also happens on https://www.scaleway.com/pricing/ (Uploaded copy of the site, just in case it changes to http://people.igalia.com/clopez/wkbug/146194/ )
Comment 2 Myles C. Maxfield 2015-09-09 13:23:27 PDT 
Unfortunately http://people.igalia.com/clopez/wkbug/146194/ doesn't cause an assert in a debug build on Yosemite as of r189543 :(
Comment 3 Myles C. Maxfield 2015-09-09 13:25:08 PDT 
I'm also not getting an assert on http://www.vox.com/ :(
Comment 4 Myles C. Maxfield 2015-09-09 13:32:55 PDT 
From inspection, it must be that the following is true

glyphBuffer && glyphBuffer->size() == lastGlyphCount + 1 && isSoftBankEmoji(previousCharacter)

but there are no m_typesettingFeatures.
Comment 5 Myles C. Maxfield 2015-09-09 13:33:42 PDT 
This is a case where the assert is wrong.
Comment 6 Myles C. Maxfield 2015-09-09 13:49:34 PDT 
Created attachment 260871 [details]
Patch
Comment 7 Myles C. Maxfield 2015-09-09 14:05:32 PDT 
Comment on attachment 260871 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=260871&action=review

> Source/WebCore/ChangeLog:10
> +        No new tests because there is no behavior change.

This isn't quite true. I should still create a test.
Comment 8 Myles C. Maxfield 2015-09-09 14:40:44 PDT 
Created attachment 260874 [details]
Patch
Comment 9 Dean Jackson 2015-09-09 14:46:55 PDT 
Comment on attachment 260874 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=260874&action=review

> Source/WebCore/platform/graphics/WidthIterator.cpp:105
> +    if (glyphBuffer && glyphBuffer->size() == lastGlyphCount + 1 && isSoftBankEmoji(previousCharacter))

Could you put () around lastGlyphCount + 1?
Comment 10 Myles C. Maxfield 2015-09-09 15:42:59 PDT 
Committed r189557: <http://trac.webkit.org/changeset/189557>
Comment 11 Carlos Garcia Campos 2015-10-29 02:09:16 PDT 
*** Bug 150638 has been marked as a duplicate of this bug. ***