/*
* Read the next line from a tsearch data file (expected to be in UTF-8), and
* convert it to database encoding if needed. The returned string is palloc'd.
* NULL return means EOF.
*/
char *
tsearch_readline(tsearch_readline_state *stp)
{
char *result;
stp->lineno++;
stp->curline = NULL;
result = t_readline(stp->fp);
stp->curline = result;
return result;
}
/*
* Reads a stop-word file. Each word is run through 'wordop'
* function, if given. wordop may either modify the input in-place,
* or palloc a new version.
*/
void
readstoplist(const char *fname, StopList *s, char *(*wordop) (const char *))
{
char **stop = NULL;
s->len = 0;
if (fname && *fname)
{
char *filename = get_tsearch_config_filename(fname, "stop");
FILE *hin;
char *line;
int reallen = 0;
if ((hin = AllocateFile(filename, "r")) == NULL)
ereport(ERROR,
(errcode(ERRCODE_CONFIG_FILE_ERROR),
errmsg("could not open stop-word file \"%s\": %m",
filename)));
while ((line = t_readline(hin)) != NULL)
{
char *pbuf = line;
/* Trim trailing space */
while (*pbuf && !t_isspace(pbuf))
pbuf++;
*pbuf = '\0';
/* Skip empty lines */
if (*line == '\0')
{
pfree(line);
continue;
}
if (s->len >= reallen)
{
if (reallen == 0)
{
reallen = 64;
stop = (char **) palloc(sizeof(char *) * reallen);
}
else
{
reallen *= 2;
stop = (char **) repalloc((void *) stop,
sizeof(char *) * reallen);
}
}
if (wordop)
{
stop[s->len] = wordop(line);
if (stop[s->len] != line)
pfree(line);
}
else
stop[s->len] = line;
(s->len)++;
}
FreeFile(hin);
pfree(filename);
}
s->stop = stop;
/* Sort to allow binary searching */
if (s->stop && s->len > 0)
qsort(s->stop, s->len, sizeof(char *), comparestr);
}