// Return a STRING containing debug information on the unichar, including
// the id_to_unichar, its hex unicodes and the properties.
STRING UNICHARSET::debug_str(UNICHAR_ID id) const {
if (id == INVALID_UNICHAR_ID) return STRING(id_to_unichar(id));
const CHAR_FRAGMENT *fragment = this->get_fragment(id);
if (fragment) {
return fragment->to_string();
}
const char* str = id_to_unichar(id);
STRING result = debug_utf8_str(str);
// Append a for lower alpha, A for upper alpha, and x if alpha but neither.
if (get_isalpha(id)) {
if (get_islower(id))
result += "a";
else if (get_isupper(id))
result += "A";
else
result += "x";
}
// Append 0 if a digit.
if (get_isdigit(id)) {
result += "0";
}
// Append p is a punctuation symbol.
if (get_ispunctuation(id)) {
result += "p";
}
return result;
}
// Sets all the properties for this unicharset given a src unicharset with
// everything set. The unicharsets don't have to be the same, and graphemes
// are correctly accounted for.
void UNICHARSET::PartialSetPropertiesFromOther(int start_index,
const UNICHARSET& src) {
for (int ch = start_index; ch < size_used; ++ch) {
const char* utf8 = id_to_unichar(ch);
UNICHAR_PROPERTIES properties;
if (src.GetStrProperties(utf8, &properties)) {
// Setup the script_id, other_case, and mirror properly.
const char* script = src.get_script_from_script_id(properties.script_id);
properties.script_id = add_script(script);
const char* other_case = src.id_to_unichar(properties.other_case);
if (contains_unichar(other_case)) {
properties.other_case = unichar_to_id(other_case);
} else {
properties.other_case = ch;
}
const char* mirror_str = src.id_to_unichar(properties.mirror);
if (contains_unichar(mirror_str)) {
properties.mirror = unichar_to_id(mirror_str);
} else {
properties.mirror = ch;
}
unichars[ch].properties.CopyFrom(properties);
set_normed_ids(ch);
} else {
tprintf("Failed to get properties for index %d = %s\n", ch, utf8);
}
}
}
// Expands the tops and bottoms and widths for this unicharset given a
// src unicharset with ranges in it. The unicharsets don't have to be the
// same, and graphemes are correctly accounted for.
void UNICHARSET::ExpandRangesFromOther(const UNICHARSET& src) {
for (int ch = 0; ch < size_used; ++ch) {
const char* utf8 = id_to_unichar(ch);
UNICHAR_PROPERTIES properties;
if (src.GetStrProperties(utf8, &properties)) {
// Expand just the ranges from properties.
unichars[ch].properties.ExpandRangesFrom(properties);
}
}
}
// Sets the normed_ids vector from the normed string. normed_ids is not
// stored in the file, and needs to be set when the UNICHARSET is loaded.
void UNICHARSET::set_normed_ids(UNICHAR_ID unichar_id) {
unichars[unichar_id].properties.normed_ids.truncate(0);
if (unichar_id == UNICHAR_SPACE && id_to_unichar(unichar_id)[0] == ' ') {
unichars[unichar_id].properties.normed_ids.push_back(UNICHAR_SPACE);
} else if (!encode_string(unichars[unichar_id].properties.normed.string(),
true, &unichars[unichar_id].properties.normed_ids,
NULL, NULL)) {
unichars[unichar_id].properties.normed_ids.truncate(0);
unichars[unichar_id].properties.normed_ids.push_back(unichar_id);
}
}
const char* const UNICHARSET::id_to_unichar_ext(UNICHAR_ID id) const {
if (id == INVALID_UNICHAR_ID) {
return INVALID_UNICHAR;
}
ASSERT_HOST(id < this->size());
// Resolve from the kCustomLigatures table if this is a private encoding.
if (get_isprivate(id)) {
const char* ch = id_to_unichar(id);
for (int i = 0; kCustomLigatures[i][0] != NULL; ++i) {
if (!strcmp(ch, kCustomLigatures[i][1])) {
return kCustomLigatures[i][0];
}
}
}
// Otherwise return the stored representation.
return unichars[id].representation;
}
// Returns whether the unichar id represents a unicode value in the private use
// area. We use this range only internally to represent uncommon ligatures
// (eg. 'ct') that do not have regular unicode values.
bool UNICHARSET::get_isprivate(UNICHAR_ID unichar_id) const {
UNICHAR uc(id_to_unichar(unichar_id), -1);
int uni = uc.first_uni();
return (uni >= 0xE000 && uni <= 0xF8FF);
}
