Bug 290230
| Summary: | REGRESSION(291808@main): [macOS Debug]: ASSERTION FAILED: !index in CSSFontSelector::fallbackFontAt on html5lib/generated/run-entities01-data.html (290012) | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Vitor Roriz <vitor.roriz> |
| Component: | Text | Assignee: | Nobody <webkit-unassigned> |
| Status: | RESOLVED DUPLICATE | ||
| Severity: | Normal | CC: | fantasai.bugs, vitor.roriz |
| Priority: | P2 | Keywords: | InRadar |
| Version: | Other | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
Vitor Roriz
Settings control how we resolve generic font-family names.
This bug happens because we were setting setFontFallbackPrefersPictographs(true)
from another test, on the same process, before running run-entities01-data.html.
This made the cached FontCascadeFont object to have its m_lastRealizedFallbackIndex
expanded beyound effectiveFamilyCount(), making fontSelectorFallbackIndex turning 1.
When running run-entities01-data.html, m_lastRealizedFallbackIndex is still
retaining that value. However, fontSelector->fallbackFontCount() that comes from Settings
is now 0 and we won't pass the check `fontSelectorFallbackIndex == fontSelector->fallbackFontCount()`.
This lead us to a inconsistent state in which we call fallbackFontAt for resolving the pictographFontFamily,
although we are stating by the index != 0 that we already did.
We can fix this issue by considering the generic font families configuration from settings for building
FontCascadeCache key. That way, we won't have false positive hits when settings differ.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Alexey Proskuryakov
*** This bug has been marked as a duplicate of bug 290231 ***