gecko-dev/intl/hyphenation/nsHyphenator.cpp

123 lines
3.4 KiB
C++

/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#include "nsHyphenator.h"
#include "nsIFile.h"
#include "nsUTF8Utils.h"
#include "nsUnicodeProperties.h"
#include "nsUnicharUtilCIID.h"
#include "nsIURI.h"
#include "hyphen.h"
nsHyphenator::nsHyphenator(nsIURI *aURI)
: mDict(nullptr)
{
nsCString uriSpec;
nsresult rv = aURI->GetSpec(uriSpec);
if (NS_FAILED(rv)) {
return;
}
mDict = hnj_hyphen_load(uriSpec.get());
#ifdef DEBUG
if (mDict) {
printf("loaded hyphenation patterns from %s\n", uriSpec.get());
}
#endif
}
nsHyphenator::~nsHyphenator()
{
if (mDict != nullptr) {
hnj_hyphen_free((HyphenDict*)mDict);
mDict = nullptr;
}
}
bool
nsHyphenator::IsValid()
{
return (mDict != nullptr);
}
nsresult
nsHyphenator::Hyphenate(const nsAString& aString,
FallibleTArray<bool>& aHyphens)
{
if (!aHyphens.SetLength(aString.Length())) {
return NS_ERROR_OUT_OF_MEMORY;
}
memset(aHyphens.Elements(), false, aHyphens.Length() * sizeof(bool));
bool inWord = false;
uint32_t wordStart = 0, wordLimit = 0;
uint32_t chLen;
for (uint32_t i = 0; i < aString.Length(); i += chLen) {
uint32_t ch = aString[i];
chLen = 1;
if (NS_IS_HIGH_SURROGATE(ch)) {
if (i + 1 < aString.Length() && NS_IS_LOW_SURROGATE(aString[i+1])) {
ch = SURROGATE_TO_UCS4(ch, aString[i+1]);
chLen = 2;
} else {
NS_WARNING("unpaired surrogate found during hyphenation");
}
}
nsIUGenCategory::nsUGenCategory cat = mozilla::unicode::GetGenCategory(ch);
if (cat == nsIUGenCategory::kLetter || cat == nsIUGenCategory::kMark) {
if (!inWord) {
inWord = true;
wordStart = i;
}
wordLimit = i + chLen;
if (i + chLen < aString.Length()) {
continue;
}
}
if (inWord) {
const char16_t *begin = aString.BeginReading();
NS_ConvertUTF16toUTF8 utf8(begin + wordStart,
wordLimit - wordStart);
nsAutoTArray<char,200> utf8hyphens;
utf8hyphens.SetLength(utf8.Length() + 5);
char **rep = nullptr;
int *pos = nullptr;
int *cut = nullptr;
int err = hnj_hyphen_hyphenate2((HyphenDict*)mDict,
utf8.BeginReading(), utf8.Length(),
utf8hyphens.Elements(), nullptr,
&rep, &pos, &cut);
if (!err) {
// Surprisingly, hnj_hyphen_hyphenate2 converts the 'hyphens' buffer
// from utf8 code unit indexing (which would match the utf8 input
// string directly) to Unicode character indexing.
// We then need to convert this to utf16 code unit offsets for Gecko.
const char *hyphPtr = utf8hyphens.Elements();
const char16_t *cur = begin + wordStart;
const char16_t *end = begin + wordLimit;
while (cur < end) {
if (*hyphPtr & 0x01) {
aHyphens[cur - begin] = true;
}
cur++;
if (cur < end && NS_IS_LOW_SURROGATE(*cur) &&
NS_IS_HIGH_SURROGATE(*(cur-1)))
{
cur++;
}
hyphPtr++;
}
}
}
inWord = false;
}
return NS_OK;
}