returning pointer into const argument

[System Administrator]

I have a function

int match(struct regexp *regexp, char *string, char **tail)

which matches a string against a regular expression, and if it
succeeds sets *tail to be the unmatched tail of the string.

I'd like the string argument to be const char *, to document the fact
that the match function doesn't modify it, but then I can't assign the
tail without an explicit de-constifying cast.

Should I give up the idea of using const here?

-- Richard

[Jeremy Yallop]

System Administrator wrote:

I have a function

int match(struct regexp *regexp, char *string, char **tail)

which matches a string against a regular expression, and if it
succeeds sets *tail to be the unmatched tail of the string.

I'd like the string argument to be const char *, to document the fact
that the match function doesn't modify it, but then I can't assign the
tail without an explicit de-constifying cast.

Adding the const (and the cast) seems reasonable here. That's how
the strtol() family handle the same issue.

Jeremy.

[Stephen Sprunk]

"System Administrator" <ro**@cogsci.ed.ac.uk> wrote in message
news:c6***********@pc-news.cogsci.ed.ac.uk...

Aside: doing any user-type tasks (e.g. posting to usenet) as root is poor
form -- you should only be root when necessary for a specific task.

I have a function

int match(struct regexp *regexp, char *string, char **tail)

which matches a string against a regular expression, and if it
succeeds sets *tail to be the unmatched tail of the string.

I'd like the string argument to be const char *, to document the fact
that the match function doesn't modify it, but then I can't assign the
tail without an explicit de-constifying cast.

Should I give up the idea of using const here?

Does "char * const * tail" remove the need for a cast?

S

--
Stephen Sprunk "Stupid people surround themselves with smart
CCIE #3723 people. Smart people surround themselves with
K5SSS smart people who disagree with them." --Aaron Sorkin

[Xenos]

"Stephen Sprunk" <st*****@sprunk.org> wrote in message
news:06******************************@news.teranew s.com...

"System Administrator" <ro**@cogsci.ed.ac.uk> wrote in message
news:c6***********@pc-news.cogsci.ed.ac.uk...

Aside: doing any user-type tasks (e.g. posting to usenet) as root is poor
form -- you should only be root when necessary for a specific task.

I have a function

int match(struct regexp *regexp, char *string, char **tail)

which matches a string against a regular expression, and if it
succeeds sets *tail to be the unmatched tail of the string.

I'd like the string argument to be const char *, to document the fact
that the match function doesn't modify it, but then I can't assign the
tail without an explicit de-constifying cast.

Should I give up the idea of using const here?

Does "char * const * tail" remove the need for a cast?

S

--
Stephen Sprunk "Stupid people surround themselves with smart
CCIE #3723 people. Smart people surround themselves with
K5SSS smart people who disagree with them." --Aaron Sorkin

don't you mean: const char** tail (or even: char const ** tail)? Its the
char that is constant, not the pointer.