corysanin/tubestation
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Bug 1719554 - Unify most of nsUnicodeProperties.h; r=platform-i18n-reviewers,jfkthame,gregtatum,necko-reviewers,valentin

Browse Source 
This unifies most of the calls in nsUnicodeProperties.h. CharType and Script
will be handled in subsequent patches on this bug.

Differential Revision: https://phabricator.services.mozilla.com/D132273
This commit is contained in:
Dan Minor
2021-12-06 18:15:49 +00:00
parent 57984a4b4d
commit 1244a99641
13 changed files with 286 additions and 114 deletions
Download Patch File Download Diff File Expand all files Collapse all files

103
intl/unicharutil/util/nsUnicodeProperties.h
View File

@@ -7,14 +7,13 @@
#ifndef NS_UNICODEPROPERTIES_H
#define NS_UNICODEPROPERTIES_H
#include "mozilla/intl/UnicodeProperties.h"
#include "nsBidiUtils.h"
#include "nsUGenCategory.h"
#include "nsUnicodeScriptCodes.h"
#include "harfbuzz/hb.h"
#include "unicode/uchar.h"
#include "unicode/uscript.h"
const nsCharProps2& GetCharProps2(uint32_t aCh);
namespace mozilla {
@@ -57,32 +56,17 @@ const uint32_t kEmojiSkinToneLast = 0x1f3ff;
extern const hb_unicode_general_category_t sICUtoHBcategory[];
inline uint32_t GetMirroredChar(uint32_t aCh) { return u_charMirror(aCh); }
inline bool HasMirroredChar(uint32_t aCh) { return u_isMirrored(aCh); }
inline uint8_t GetCombiningClass(uint32_t aCh) {
return u_getCombiningClass(aCh);
}
inline uint8_t GetGeneralCategory(uint32_t aCh) {
return sICUtoHBcategory[u_charType(aCh)];
return sICUtoHBcategory[intl::UnicodeProperties::CharType(aCh)];
}
inline nsCharType GetBidiCat(uint32_t aCh) {
return nsCharType(u_charDirection(aCh));
}
inline int8_t GetNumericValue(uint32_t aCh) {
UNumericType type =
UNumericType(u_getIntPropertyValue(aCh, UCHAR_NUMERIC_TYPE));
return type == U_NT_DECIMAL || type == U_NT_DIGIT
? int8_t(u_getNumericValue(aCh))
: -1;
}
inline uint8_t GetLineBreakClass(uint32_t aCh) {
return u_getIntPropertyValue(aCh, UCHAR_LINE_BREAK);
return intl::UnicodeProperties::GetIntPropertyValue(
aCh, intl::UnicodeProperties::IntProperty::LineBreak);
}
inline Script GetScriptCode(uint32_t aCh) {
@@ -104,28 +88,22 @@ inline uint32_t GetScriptTagForCode(Script aScriptCode) {
}
inline PairedBracketType GetPairedBracketType(uint32_t aCh) {
return PairedBracketType(
u_getIntPropertyValue(aCh, UCHAR_BIDI_PAIRED_BRACKET_TYPE));
return PairedBracketType(intl::UnicodeProperties::GetIntPropertyValue(
aCh, intl::UnicodeProperties::IntProperty::BidiPairedBracketType));
}
inline uint32_t GetPairedBracket(uint32_t aCh) {
return u_getBidiPairedBracket(aCh);
}
inline uint32_t GetUppercase(uint32_t aCh) { return u_toupper(aCh); }
inline uint32_t GetLowercase(uint32_t aCh) { return u_tolower(aCh); }
inline uint32_t GetTitlecaseForLower(
uint32_t aCh) // maps LC to titlecase, UC unchanged
{
return u_isULowercase(aCh) ? u_totitle(aCh) : aCh;
return intl::UnicodeProperties::IsLowercase(aCh)
? intl::UnicodeProperties::ToTitle(aCh)
: aCh;
}
inline uint32_t GetTitlecaseForAll(
uint32_t aCh) // maps both UC and LC to titlecase
{
return u_totitle(aCh);
return intl::UnicodeProperties::ToTitle(aCh);
}
inline uint32_t GetFoldedcase(uint32_t aCh) {
@@ -135,62 +113,22 @@ inline uint32_t GetFoldedcase(uint32_t aCh) {
if (aCh == 0x0130 || aCh == 0x0131) {
return 'i';
}
return u_foldCase(aCh, U_FOLD_CASE_DEFAULT);
}
inline bool IsEastAsianWidthFHWexcludingEmoji(uint32_t aCh) {
switch (u_getIntPropertyValue(aCh, UCHAR_EAST_ASIAN_WIDTH)) {
case U_EA_FULLWIDTH:
case U_EA_HALFWIDTH:
return true;
case U_EA_WIDE:
return u_hasBinaryProperty(aCh, UCHAR_EMOJI) ? false : true;
case U_EA_AMBIGUOUS:
case U_EA_NARROW:
case U_EA_NEUTRAL:
return false;
}
return false;
}
inline bool IsEastAsianWidthAFW(uint32_t aCh) {
switch (u_getIntPropertyValue(aCh, UCHAR_EAST_ASIAN_WIDTH)) {
case U_EA_AMBIGUOUS:
case U_EA_FULLWIDTH:
case U_EA_WIDE:
return true;
case U_EA_HALFWIDTH:
case U_EA_NARROW:
case U_EA_NEUTRAL:
return false;
}
return false;
}
inline bool IsEastAsianWidthFW(uint32_t aCh) {
switch (u_getIntPropertyValue(aCh, UCHAR_EAST_ASIAN_WIDTH)) {
case U_EA_FULLWIDTH:
case U_EA_WIDE:
return true;
case U_EA_AMBIGUOUS:
case U_EA_HALFWIDTH:
case U_EA_NARROW:
case U_EA_NEUTRAL:
return false;
}
return false;
return intl::UnicodeProperties::FoldCase(aCh);
}
inline bool IsDefaultIgnorable(uint32_t aCh) {
return u_hasBinaryProperty(aCh, UCHAR_DEFAULT_IGNORABLE_CODE_POINT);
return intl::UnicodeProperties::HasBinaryProperty(
aCh, intl::UnicodeProperties::BinaryProperty::DefaultIgnorableCodePoint);
}
inline EmojiPresentation GetEmojiPresentation(uint32_t aCh) {
if (!u_hasBinaryProperty(aCh, UCHAR_EMOJI)) {
if (!intl::UnicodeProperties::HasBinaryProperty(
aCh, intl::UnicodeProperties::BinaryProperty::Emoji)) {
return TextOnly;
}
if (u_hasBinaryProperty(aCh, UCHAR_EMOJI_PRESENTATION)) {
if (intl::UnicodeProperties::HasBinaryProperty(
aCh, intl::UnicodeProperties::BinaryProperty::EmojiPresentation)) {
return EmojiDefault;
}
return TextDefault;
@@ -271,11 +209,6 @@ uint32_t CountGraphemeClusters(const char16_t* aText, uint32_t aLength);
// to the values we care about at runtime.
bool IsCombiningDiacritic(uint32_t aCh);
// Keep this function in sync with is_math_symbol in base_chars.py.
inline bool IsMathOrMusicSymbol(uint32_t aCh) {
return u_charType(aCh) == U_MATH_SYMBOL || u_charType(aCh) == U_OTHER_SYMBOL;
}
// Remove diacritics from a character
uint32_t GetNaked(uint32_t aCh);