A tiny, sensible subset of regex.
- Zero dependencies (unless you count the C standard library)
- Completely cross-platform
- Tiny implementation
- No allocations or global state
- Concise, safe API
picomatch supports a wide subset of regex, including..
- Anchors
^ $ \b \B - Grouping and capturing
() (?:) - Sets, including ranges
[...] [^...] [a-zA-z] - Character classes
\s \S \w \W \d \D - Escaped characters
\n \r \t \0 \\ - Quantifiers
+ +? * *? {1} {1,} {1,5} ?
Unlike most other small regex libraries, picomatch's quantifiers support limited backtracking as well
Here's a basic example using picomatch to identify email addresses
// capture both the name and the provider
const char* regex = "^([\\w.\\-]{0,25})@(yahoo|hotmail|gmail)\\.com$";
// get the required allocation size for the compiled regex
// you can skip this step if you provide a sensible allocation,
// and picomatch will gracefully error if it's not enough
const char* err = NULL;
size_t compiled_size = pm_expsize(regex, &err);
if (compiled_size == 0) {
printf("Failed to size: %s\n", err);
return 1;
}
// Compile the regex!
pm_Regex* expr = malloc(compiled_size);
if (!pm_compile(expr, compiled_size, regex)) {
printf("Failed to compile: %s\n", pm_geterror(expr));
}
// Get the number of capture groups required;
int num_groups = pm_getgroups(expr);
pm_Group* capture_groups = malloc(sizeof(pm_Group) * num_groups);
memset(capture_groups, 0, sizeof(pm_Group) * num_groups);
const char* source = "example-mail@yahoo.com";
printf("\nMatches: %d\n", pm_match(expr, source, 0, capture_groups, num_groups, 0));
for (int i = 0; i < num_groups; i++) {
printf("\t%d - '%.*s'\n", i, capture_groups[i].length, source + capture_groups[i].start);
}
// matches: 1
// 0 - 'example-mail@yahoo.com'
// 1 - 'example-mail'
// 2 - 'yahoo'For more complex usage, refer to the documentation in picomatch.h
picomatch is licensed under MIT, see the notice in the source files or LICENSE for details.