William Langford
e84d17196c
Read additional bytes from the file to complete the UTF-8 sequence so the bytes in it don't get converted to U+FFFD replacement characters.
63 lines
1.7 KiB
C
63 lines
1.7 KiB
C
|
|
#include <errno.h>
|
|
#include <stdio.h>
|
|
#include <stdlib.h>
|
|
#include <string.h>
|
|
#include "jv.h"
|
|
#include "jv_unicode.h"
|
|
|
|
jv jv_load_file(const char* filename, int raw) {
|
|
FILE* file = fopen(filename, "r");
|
|
struct jv_parser* parser;
|
|
jv data;
|
|
if (!file) {
|
|
return jv_invalid_with_msg(jv_string_fmt("Could not open %s: %s",
|
|
filename,
|
|
strerror(errno)));
|
|
}
|
|
if (raw) {
|
|
data = jv_string("");
|
|
} else {
|
|
data = jv_array();
|
|
parser = jv_parser_new(0);
|
|
}
|
|
|
|
// To avoid mangling UTF-8 multi-byte sequences that cross the end of our read
|
|
// buffer, we need to be able to read the remainder of a sequence and add that
|
|
// before appending.
|
|
const int max_utf8_len = 4;
|
|
char buf[4096+max_utf8_len];
|
|
while (!feof(file) && !ferror(file)) {
|
|
size_t n = fread(buf, 1, sizeof(buf)-max_utf8_len, file);
|
|
int len = 0;
|
|
if (jvp_utf8_backtrack(buf+(n-1), buf, &len) && len > 0) {
|
|
if (!feof(file) && !ferror(file)) {
|
|
n += fread(buf+n, 1, len, file);
|
|
}
|
|
}
|
|
|
|
if (raw) {
|
|
data = jv_string_append_buf(data, buf, n);
|
|
} else {
|
|
jv_parser_set_buf(parser, buf, n, !feof(file));
|
|
jv value;
|
|
while (jv_is_valid((value = jv_parser_next(parser))))
|
|
data = jv_array_append(data, value);
|
|
if (jv_invalid_has_msg(jv_copy(value))) {
|
|
jv_free(data);
|
|
data = value;
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
if (!raw)
|
|
jv_parser_free(parser);
|
|
int badread = ferror(file);
|
|
if (fclose(file) != 0 || badread) {
|
|
jv_free(data);
|
|
return jv_invalid_with_msg(jv_string_fmt("Error reading from %s",
|
|
filename));
|
|
}
|
|
return data;
|
|
}
|