bool handle_line(const char *data, size_t datalen, void *pw)
{
line_ctx *ctx = (line_ctx *) pw;
if (data[0] == '#') {
if (ctx->inexp) {
/* This marks end of testcase, so run it */
if (ctx->buf[ctx->bufused - 1] == '\n')
ctx->bufused -= 1;
if (ctx->exp[ctx->expused - 1] == '\n')
ctx->expused -= 1;
run_test(ctx);
ctx->buf[0] = '\0';
ctx->exp[0] = '\0';
ctx->bufused = 0;
ctx->expused = 0;
ctx->exp_ret = PARSERUTILS_OK;
}
if (strncasecmp(data+1, "data", 4) == 0) {
parserutils_charset_codec_optparams params;
const char *ptr = data + 6;
ctx->indata = true;
ctx->inexp = false;
if (strncasecmp(ptr, "decode", 6) == 0)
ctx->dir = DECODE;
else if (strncasecmp(ptr, "encode", 6) == 0)
ctx->dir = ENCODE;
else
ctx->dir = BOTH;
ptr += 7;
if (strncasecmp(ptr, "LOOSE", 5) == 0) {
params.error_mode.mode =
PARSERUTILS_CHARSET_CODEC_ERROR_LOOSE;
ptr += 6;
} else if (strncasecmp(ptr, "STRICT", 6) == 0) {
params.error_mode.mode =
PARSERUTILS_CHARSET_CODEC_ERROR_STRICT;
ptr += 7;
} else {
params.error_mode.mode =
PARSERUTILS_CHARSET_CODEC_ERROR_TRANSLIT;
ptr += 9;
}
assert(parserutils_charset_codec_setopt(ctx->codec,
PARSERUTILS_CHARSET_CODEC_ERROR_MODE,
(parserutils_charset_codec_optparams *) ¶ms)
== PARSERUTILS_OK);
} else if (strncasecmp(data+1, "expected", 8) == 0) {
ctx->indata = false;
ctx->inexp = true;
ctx->exp_ret = parserutils_error_from_string(data + 10,
datalen - 10 - 1 /* \n */);
} else if (strncasecmp(data+1, "reset", 5) == 0) {
ctx->indata = false;
ctx->inexp = false;
parserutils_charset_codec_reset(ctx->codec);
} else if (strncasecmp(data+1, "enc", 3) == 0) {
const char *enc = data + 5;
const char *end;
char *enc_name;
for (end = enc; !isspace(*end); end++)
;
enc_name = malloc(end - enc + 1);
memcpy(enc_name, enc, end - enc);
enc_name[end - enc] = 0;
assert(parserutils_charset_codec_create(enc_name,
myrealloc, NULL, &ctx->codec) ==
PARSERUTILS_OK);
ctx->hadenc = true;
free(enc_name);
}
} else {
if (ctx->indata) {
memcpy(ctx->buf + ctx->bufused, data, datalen);
ctx->bufused += datalen;
}
if (ctx->inexp) {
memcpy(ctx->exp + ctx->expused, data, datalen);
ctx->expused += datalen;
}
}
return true;
}
bool handle_line(const char *data, size_t datalen, void *pw)
{
line_ctx *ctx = (line_ctx *) pw;
if (data[0] == '#') {
if (ctx->inexp) {
/* This marks end of testcase, so run it */
if (ctx->buf[ctx->bufused - 1] == '\n')
ctx->bufused -= 1;
if (ctx->exp[ctx->expused - 1] == '\n')
ctx->expused -= 1;
run_test(ctx);
ctx->buf[0] = '\0';
ctx->exp[0] = '\0';
ctx->bufused = 0;
ctx->expused = 0;
ctx->exp_ret = PARSERUTILS_OK;
}
if (strncasecmp(data+1, "data", 4) == 0) {
parserutils_charset_codec_optparams params;
const char *ptr = data + 6;
ctx->indata = true;
ctx->inexp = false;
if (strncasecmp(ptr, "decode", 6) == 0)
ctx->dir = DECODE;
else if (strncasecmp(ptr, "encode", 6) == 0)
ctx->dir = ENCODE;
else
ctx->dir = BOTH;
ptr += 7;
if (strncasecmp(ptr, "LOOSE", 5) == 0) {
params.error_mode.mode =
PARSERUTILS_CHARSET_CODEC_ERROR_LOOSE;
ptr += 6;
} else if (strncasecmp(ptr, "STRICT", 6) == 0) {
params.error_mode.mode =
PARSERUTILS_CHARSET_CODEC_ERROR_STRICT;
ptr += 7;
} else {
params.error_mode.mode =
PARSERUTILS_CHARSET_CODEC_ERROR_TRANSLIT;
ptr += 9;
}
assert(parserutils_charset_codec_setopt(ctx->codec,
PARSERUTILS_CHARSET_CODEC_ERROR_MODE,
(parserutils_charset_codec_optparams *) ¶ms)
== PARSERUTILS_OK);
} else if (strncasecmp(data+1, "expected", 8) == 0) {
ctx->indata = false;
ctx->inexp = true;
ctx->exp_ret = parserutils_error_from_string(data + 10,
datalen - 10 - 1 /* \n */);
} else if (strncasecmp(data+1, "reset", 5) == 0) {
ctx->indata = false;
ctx->inexp = false;
parserutils_charset_codec_reset(ctx->codec);
}
} else {
if (ctx->indata) {
/* Process "&#xNNNN" as 16-bit code units. */
while (datalen) {
uint16_t nCodePoint;
if (data[0] == '\n') {
ctx->buf[ctx->bufused++] = *data++;
--datalen;
continue;
}
assert(datalen >= sizeof ("&#xNNNN")-1 \
&& data[0] == '&' && data[1] == '#' \
&& data[2] == 'x' && isxdigit(data[3]) \
&& isxdigit(data[4]) && isxdigit(data[5]) \
&& isxdigit(data[6]));
/* UTF-16 code is always host endian (different
than UCS-32 !). */
nCodePoint = (hex2digit(data[3]) << 12) |
(hex2digit(data[4]) << 8) |
(hex2digit(data[5]) << 4) |
hex2digit(data[6]);
*((uint16_t *) (void *) (ctx->buf + ctx->bufused)) =
nCodePoint;
ctx->bufused += 2;
data += sizeof ("&#xNNNN")-1;
datalen -= sizeof ("&#xNNNN")-1;
}
}
if (ctx->inexp) {
/* Process "&#xXXXXYYYY as 32-bit code units. */
while (datalen) {
uint32_t nCodePoint;
if (data[0] == '\n') {
ctx->exp[ctx->expused++] = *data++;
--datalen;
continue;
}
assert(datalen >= sizeof ("&#xXXXXYYYY")-1 \
&& data[0] == '&' && data[1] == '#' \
&& data[2] == 'x' && isxdigit(data[3]) \
&& isxdigit(data[4]) && isxdigit(data[5]) \
&& isxdigit(data[6]) && isxdigit(data[7]) \
&& isxdigit(data[8]) && isxdigit(data[9]) \
&& isxdigit(data[10]));
/* UCS-4 code is always big endian, so convert
host endian to big endian. */
nCodePoint =
htonl((hex2digit(data[3]) << 28)
| (hex2digit(data[4]) << 24)
| (hex2digit(data[5]) << 20)
| (hex2digit(data[6]) << 16)
| (hex2digit(data[7]) << 12)
| (hex2digit(data[8]) << 8)
| (hex2digit(data[9]) << 4)
| hex2digit(data[10]));
*((uint32_t *) (void *) (ctx->exp + ctx->expused)) =
nCodePoint;
ctx->expused += 4;
data += sizeof ("&#xXXXXYYYY")-1;
datalen -= sizeof ("&#xXXXXYYYY")-1;
}
}
}
return true;
}
