regexec calls with REG_STARTEND find ghost matches for \>

[Benno Schulenberg <bensberg at justemail dot net>]

When calling regexec with the REG_STARTEND flag and providing
an end-of-range value in .rm_eo that points to somewhere in
the middle of a word, regexec will nevertheless find a match
for \> at that offset.  The corresponding case for \<, with a
start-of-range value in .rm_so that points to the middle of a
word, will /not/ find a match for \< there.  The latter is
what I expected, the former was a surprise.

See attached two files for a demonstration of the above two
cases.

I know that REG_STARTEND is undocmented, but why is that?
Because it is not working right?  Or because you don't want
people to use it?  But what is the point of having a function
when it can't be used?

(Please CC; not subscribed.)

Benno

-- 
http://www.fastmail.com - Does exactly what it says on the tin

#include <regex.h>
#include <stdio.h>

int search_tail_until(int limit, char *string)
{
	regex_t rgx;
	regmatch_t match[1];

	regcomp(&rgx, "\\>", 0);

	match[0].rm_so = 0;
	match[0].rm_eo = limit;

	if (regexec(&rgx, string, 1, match, REG_STARTEND) == 0)
		return match[0].rm_so;
	else
		return -1;
}

int main()
{
	int index;
	char *line = "  word.  ";

	index = search_tail_until(8, line);
	if (index >= 0)
		printf("Found tail at %i: '%c'\n", index, line[index]);

	index = search_tail_until(4, line);
	if (index >= 0)
		printf("Found tail at %i: '%c'\n", index, line[index]);
}

#include <regex.h>
#include <stdio.h>
#include <string.h>

int search_head_from(int start, char *string)
{
	regex_t rgx;
	regmatch_t match[1];

	regcomp(&rgx, "\\<", 0);

	match[0].rm_so = start;
	match[0].rm_eo = strlen(string);

	if (regexec(&rgx, string, 1, match, REG_STARTEND) == 0)
		return match[0].rm_so;
	else
		return -1;
}

int main()
{
	int index;
	char *line = "  word.  ";

	index = search_head_from(0, line);
	if (index >= 0)
		printf("Found head at %i: '%c'\n", index, line[index]);

	index = search_head_from(4, line);
	if (index >= 0)
		printf("Found head at %i: '%c'\n", index, line[index]);
}