translate: share vstr_add_chr's implementation of utf8 encoding

this saves ~44 bytes on trinket_m0
This commit is contained in:
Jeff Epler 2023-11-23 12:04:10 -06:00
parent 3c8f9f4dab
commit a38cd711ce
No known key found for this signature in database
GPG Key ID: D5BF15AB975AB4DE
1 changed files with 17 additions and 28 deletions

View File

@ -57,37 +57,23 @@ STATIC void get_word(int n, const mchar_t **pos, const mchar_t **end) {
*end = *pos + len;
}
STATIC int put_utf8(char *buf, int u) {
STATIC void put_utf8(vstr_t *vstr, int u) {
if (u >= translation_offstart) {
u += translation_offset;
}
if (u <= 0x7f) {
*buf = u;
return 1;
} else if (word_start <= u && u <= word_end) {
if (word_start <= u && u <= word_end) {
uint n = (u - word_start);
const mchar_t *pos, *end;
get_word(n, &pos, &end);
int ret = 0;
// note that at present, entries in the words table are
// guaranteed not to represent words themselves, so this adds
// at most 1 level of recursive call
for (; pos < end; pos++) {
int len = put_utf8(buf, *pos);
buf += len;
ret += len;
put_utf8(vstr, *pos);
}
return ret;
} else if (u <= 0x07ff) {
*buf++ = 0b11000000 | (u >> 6);
*buf = 0b10000000 | (u & 0b00111111);
return 2;
} else { // u <= 0xffff
*buf++ = 0b11100000 | (u >> 12);
*buf++ = 0b10000000 | ((u >> 6) & 0b00111111);
*buf = 0b10000000 | (u & 0b00111111);
return 3;
return;
}
vstr_add_char(vstr, u);
}
uint16_t decompress_length(mp_rom_error_text_t compressed) {
@ -123,15 +109,15 @@ static int get_nbits(bitstream_state_t *st, int n) {
return r;
}
char *decompress(mp_rom_error_text_t compressed, char *decompressed) {
static void decompress_vstr(mp_rom_error_text_t compressed, vstr_t *decompressed) {
bitstream_state_t b = {
.ptr = &(compressed->data) + (compress_max_length_bits >> 3),
.bit = 1 << (7 - ((compress_max_length_bits) & 0x7)),
};
uint16_t length = decompress_length(compressed);
size_t alloc = decompressed->alloc - 1;
// Stop one early because the last byte is always NULL.
for (uint16_t i = 0; i < length - 1;) {
for (; decompressed->len < alloc;) {
uint32_t bits = 0;
uint8_t bit_length = 0;
uint32_t max_code = lengths[0];
@ -148,16 +134,19 @@ char *decompress(mp_rom_error_text_t compressed, char *decompressed) {
int v = values[searched_length + bits - max_code];
if (v == 1) {
qstr q = get_nbits(&b, translation_qstr_bits) + 1; // honestly no idea why "+1"...
for (const char *qc = qstr_str(q); *qc;) {
decompressed[i++] = *qc++;
}
vstr_add_str(decompressed, qstr_str(q));
} else {
i += put_utf8(decompressed + i, v);
put_utf8(decompressed, v);
}
}
}
decompressed[length - 1] = '\0';
return decompressed;
char *decompress(mp_rom_error_text_t compressed, char *decompressed) {
vstr_t vstr;
vstr_init_fixed_buf(&vstr, decompress_length(compressed), decompressed);
decompress_vstr(compressed, &vstr);
return vstr_null_terminated_str(&vstr);
}
#if CIRCUITPY_TRANSLATE_OBJECT == 1