ref: 81c9ede53f6f9ebd67e7134e664c98e08b9fcca6
dir: /sys/man/2/regexp/
.TH REGEXP 2
.SH NAME
regcomp, regcomplit, regcompnl, regexec, regsub, rregexec, rregsub, regerror \- regular expression
.SH SYNOPSIS
.B #include <u.h>
.br
.B #include <libc.h>
.br
.B #include <regexp.h>
.PP
.ta \w'\fLRegprog 'u
.B
Reprog *regcomp(char *exp)
.PP
.B
Reprog *regcomplit(char *exp)
.PP
.B
Reprog *regcompnl(char *exp)
.PP
.nf
.B
int regexec(Reprog *prog, char *string, Resub *match, int msize)
.PP
.nf
.B
void regsub(char *source, char *dest, int dlen, Resub *match, int msize)
.PP
.nf
.B
int rregexec(Reprog *prog, Rune *string, Resub *match, int msize)
.PP
.nf
.B
void rregsub(Rune *source, Rune *dest, int dlen, Resub *match, int msize)
.PP
.B
void regerror(char *msg)
.SH DESCRIPTION
.I Regcomp
compiles a
regular expression and returns
a pointer to the generated description.
The space is allocated by
.IR malloc (2)
and may be released by
.IR free .
Regular expressions are exactly as in
.IR regexp (6).
.PP
.I Regcomplit
is like
.I regcomp
except that all characters are treated literally.
.I Regcompnl
is like
.I regcomp
except that the
.B .
metacharacter matches all characters, including newlines.
.PP
.I Regexec
matches a null-terminated
.I string
against the compiled regular expression in
.IR prog .
If it matches,
.I regexec
returns
.B 1
and fills in the array
.I match
with character pointers to the substrings of
.I string
that correspond to the
parenthesized subexpressions of
.IR exp :
.BI match[ i ].sp
points to the beginning and
.BI match[ i ].ep
points just beyond
the end of the
.IR i th
substring.
(Subexpression
.I i
begins at the
.IR i th
left parenthesis, counting from 1.)
Pointers in
.B match[0]
pick out the substring that corresponds to
the whole regular expression.
Unused elements of
.I match
are filled with zeros.
Matches involving
.LR * ,
.LR + ,
and
.L ?
are extended as far as possible.
The number of array elements in
.I match
is given by
.IR msize .
The structure of elements of
.I match
is:
.IP
.EX
typedef struct {
union {
char *sp;
Rune *rsp;
};
union {
char *ep;
Rune *rep;
};
} Resub;
.EE
.LP
If
.B match[0].sp
is nonzero on entry,
.I regexec
starts matching at that point within
.IR string .
If
.B match[0].ep
is nonzero on entry,
the last character matched is the one
preceding that point.
.PP
.I Regsub
places in
.I dest
a substitution instance of
.I source
in the context of the last
.I regexec
performed using
.IR match .
Each instance of
.BI \e n\f1,
where
.I n
is a digit, is replaced by the
string delimited by
.BI match[ n ].sp
and
.BI match[ n ].ep\f1.
Each instance of
.L &
is replaced by the string delimited by
.B match[0].sp
and
.BR match[0].ep .
The substitution will always be null terminated and
trimmed to fit into dlen bytes.
.PP
.IR Regerror ,
called whenever an error is detected in
.IR regcomp ,
writes the string
.I msg
on the standard error file and exits.
.I Regerror
can be replaced to perform
special error processing.
If the user supplied
.I regerror
returns rather than exits,
.I regcomp
will return 0.
.PP
.I Rregexec
and
.I rregsub
are variants of
.I regexec
and
.I regsub
that use strings of
.B Runes
instead of strings of
.BR chars .
With these routines, the
.I rsp
and
.I rep
fields of the
.I match
array elements should be used.
.SH SOURCE
.B /sys/src/libregexp
.SH "SEE ALSO"
.IR grep (1)
.SH DIAGNOSTICS
.I Regcomp
returns
.B 0
for an illegal expression
or other failure.
.I Regexec
returns 0
if
.I string
is not matched.
.SH BUGS
There is no way to specify or match a NUL character; NULs terminate patterns and strings.
The size of a character class and the number of sub-expression matches are hard-coded
limits. The library uses the worst-case space estimate for allocating VM runtime threads.
.SH HISTORY
.IR Regexp (2)
first appeared in Plan 9 from Bell Labs. This implementation was written from
scratch for 9front (May, 2016).