RGlob.h

Glob syntax implementation over POSIX ERE regex.

License: GPL-3.0-or-later.

Preprocessor

This glob is implemented as a preprocessor and a series of wrappers around POSIX regex API. This preprocessor translate glob patter to valid and equal regex pattern. This includes both escaping regex’s special characters and translating glob’s special characters to equal sequences of regexes.

Pattern format

In general any valid glob patter can be used, it support this “operators”:

? - match any single character. (Translates to .).
* - match any sequence of characters. (Translates to .*).
[...] - matches any character from character class. (Translates to itself literally).
! as first element of character class to invert character class. (Translates to ‘^’).

It also supports some syntax extensions:

Often you want to match some specific character multiple characters. In regex you can just write [...]*. This glob engine support same syntax. If ]* sequence is found it will not be translated in any way.
Capture groups are exposed in an API but have no special pattern syntax.
POSIX ERE support \ as escape character. Glob also support is so no preprocessing is needed here. But this preprocessor supports its own escape character too - #. Next character after it will be literally copied into output regex pattern.

These characters are escaped with \ by default in generated regex:

(.
).
{.
}.
^.
$.
..
|.
+.

By default pattern is wrapped in ^...$ to match full string.

cbuild_glob_t

typedef struct cbuild_glob_t {
struct __cbuild_glob_res_t {
const char* res;
size_t res_idx;
regmatch_t captures[CBUILD_GLOB_CAPTURE_COUNT];
} *data;
size_t size;
size_t capacity;
regex_t regex;
regmatch_t captures[CBUILD_GLOB_CAPTURE_COUNT];
} cbuild_glob_t;

Glob context.

__cbuild_glob_res_t* data Data for dynamic array of matches.
- const char* data.res Match text. Points to original data.
- size_t data.res_idx Match index.
- regmatch_t[CBUILD_GLOB_CAPTURE_COUNT] data.captures Copy of capture groups for that match.
size_t capacity Capacity for dynamic array of matches.
size_t size Count of elements for dynamic array of matches.
regex_t regex Cache for compiled regex.
regmatch_t[CBUILD_GLOB_CAPTURE_COUNT] captures Scratchpad captures for regex engine.

cbuild_glob_opts_t

struct cbuild_glob_opts_t {
bool partial_match;
};

Options for pattern compilation

bool partial_match Do not wrap pattern in ^...$.

cbuild_glob_compile_opt

CBUILDDEF bool cbuild_glob_compile_opt(cbuild_glob_t* glob, const char* pattern,
struct cbuild_glob_opts_t opts);

Compile new glob patter.

This is a two-step process, first it will transform glob expression into a regx one and then it will call regcomp. POSIX ERE will be used.

cbuild_glob_compile

#define cbuild_glob_compile(glob, pattern, ...)

Compile new glob patter.

This is a two-step process, first it will transform glob expression into a regx one and then it will call regcomp. POSIX ERE will be used.

cbuild_glob_t* glob Glob object.
const char* pattern Pattern to compile.
...cbulid_glob_opts_t … Fields of configuration structure in initializer-list form.

cbuild_glob_match

CBUILDDEF bool cbuild_glob_match(cbuild_glob_t* glob, const char** list, size_t size);

Find matches in a list.

bool true if at least one match was found. false if no matches were found.

cbuild_glob_match_single

CBUILDDEF bool cbuild_glob_match_single(cbuild_glob_t* glob, const char* elem);

Match single element with a glob.

This function resets glob matches but does not set them. You should check return value to know if glob matched. To get capture groups values you can use outer captures array.

bool true if at least one match was found. false if no matches were found.

cbuild_glob_free

CBUILDDEF void cbuild_glob_free(cbuild_glob_t* glob);

Clean underlying regex and free matches array.