Magic-Mirror/gecko-dev
1
0
Fork 0
mirror of https://github.com/mozilla/gecko-dev.git synced 2025-02-03 20:49:27 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

bug 1048899 - don't end ::first-letter in the middle of a ligature for Indic and SEAsian scripts. r=smontagu

Browse Source
This commit is contained in:
Jonathan Kew 2014-09-01 21:08:52 +01:00
parent e79dc99076
commit 4572da3eb7
1 changed files with 69 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

73
layout/generic/nsTextFrame.cpp
View File

@ -2970,17 +2970,22 @@ static void FindClusterStart(gfxTextRun* aTextRun, int32_t aOriginalStart,
}
/**
* Finds the offset of the last character of the cluster containing aPos
* Finds the offset of the last character of the cluster containing aPos.
* If aAllowSplitLigature is false, we also check for a ligature-group
* start.
*/
static void FindClusterEnd(gfxTextRun* aTextRun, int32_t aOriginalEnd,
gfxSkipCharsIterator* aPos)
gfxSkipCharsIterator* aPos,
bool aAllowSplitLigature = true)
{
NS_PRECONDITION(aPos->GetOriginalOffset() < aOriginalEnd,
"character outside string");
aPos->AdvanceOriginal(1);
while (aPos->GetOriginalOffset() < aOriginalEnd) {
if (aPos->IsOriginalCharSkipped() ||
aTextRun->IsClusterStart(aPos->GetSkippedOffset())) {
(aTextRun->IsClusterStart(aPos->GetSkippedOffset()) &&
(aAllowSplitLigature ||
aTextRun->IsLigatureGroupStart(aPos->GetSkippedOffset())))) {
break;
}
aPos->AdvanceOriginal(1);
@ -6965,8 +6970,68 @@ FindFirstLetterRange(const nsTextFragment* aFrag,
}
// consume another cluster (the actual first letter)
// For complex scripts such as Indic and SEAsian, where first-letter
// should extend to entire orthographic "syllable" clusters, we don't
// want to allow this to split a ligature.
bool allowSplitLigature;
switch (unicode::GetScriptCode(aFrag->CharAt(aOffset + i))) {
default:
allowSplitLigature = true;
break;
// For now, lacking any definitive specification of when to apply this
// behavior, we'll base the decision on the HarfBuzz shaping engine
// used for each script: those that are handled by the Indic, Tibetan,
// Myanmar and SEAsian shapers will apply the "don't split ligatures"
// rule.
// Indic
case MOZ_SCRIPT_BENGALI:
case MOZ_SCRIPT_DEVANAGARI:
case MOZ_SCRIPT_GUJARATI:
case MOZ_SCRIPT_GURMUKHI:
case MOZ_SCRIPT_KANNADA:
case MOZ_SCRIPT_MALAYALAM:
case MOZ_SCRIPT_ORIYA:
case MOZ_SCRIPT_TAMIL:
case MOZ_SCRIPT_TELUGU:
case MOZ_SCRIPT_SINHALA:
case MOZ_SCRIPT_BALINESE:
case MOZ_SCRIPT_LEPCHA:
case MOZ_SCRIPT_REJANG:
case MOZ_SCRIPT_SUNDANESE:
case MOZ_SCRIPT_JAVANESE:
case MOZ_SCRIPT_KAITHI:
case MOZ_SCRIPT_MEETEI_MAYEK:
case MOZ_SCRIPT_CHAKMA:
case MOZ_SCRIPT_SHARADA:
case MOZ_SCRIPT_TAKRI:
case MOZ_SCRIPT_KHMER:
// Tibetan
case MOZ_SCRIPT_TIBETAN:
// Myanmar
case MOZ_SCRIPT_MYANMAR:
// Other SEAsian
case MOZ_SCRIPT_BUGINESE:
case MOZ_SCRIPT_NEW_TAI_LUE:
case MOZ_SCRIPT_CHAM:
case MOZ_SCRIPT_TAI_THAM:
// What about Thai/Lao - any special handling needed?
// Should we special-case Arabic lam-alef?
allowSplitLigature = false;
break;
}
iter.SetOriginalOffset(aOffset + i);
FindClusterEnd(aTextRun, endOffset, &iter);
FindClusterEnd(aTextRun, endOffset, &iter, allowSplitLigature);
i = iter.GetOriginalOffset() - aOffset;
if (i + 1 == length)
return true;