small regex engine
Regex patterns can be compiled to byte arrays at compile time, which can be used at runtime to avoid re-compiling the regex and reducing code size by a lot, since all the compilation logic is not needed.
inaccurate benchmarks of mathing stirng green car pattern \s*?(red|green|blue)?\s*?(car|train)\s*?:
| engine | time |
|---|---|
| tpre | 0.___163 ms |
| pcre2 | 0.___785 ms |
| pcre2 (jit) | 0.____80 ms |
| GNU C++ regex | 0.___805 ms |
match any char in the range
example: 0-9
| pattern | description |
|---|---|
. |
any char, except for line breaks |
\s |
any whitespace / line break |
\S |
not a \s |
\d |
digit from 0 to 9 |
\D |
not a \d |
\w |
letter, digit, or underscore |
\W |
not a \w |
example: \n to match a line break
example: \. to match a literal .
match the previous pattern repeated, until the next pattern matches.
example: h*?i to match for example hhhhhhhi
match the previous pattern repeated (at least one time), until the next pattern matches.
example: h+?i to match for example hhhhhhhi
try to match the previous pattern.
example: h? matches both h and a.
used to group multiple patterns together
example: (?:hi)? matches both hi, and an empty string.
example: (?# this is ignored)
matches any of the inner patterns.
example: [12(?:hi)3] matches either 1, 2, hi, or 3
matches either the left pattern or the right pattern
example: hi|bye matches either hi or bye
example: ([a-zA-Z]+) matches Alex, and stores Alex in the next capture group ID, beginning with 1.
example: (?'name'[a-zA-Z]+) matches Alex, and stores Alex in the capture group name.
the limit on name length is 20 chars.
example: ^ matches beginning of string.
example: $ matches end of string.