# comments can appear anywhere
# groups are (?:...), captures are ()
# regex for a.b+
START CAPTURE 0
MATCH 1 CHAR A
MATCH 1 WILD
MATCH + CHAR B
END CAPTURE 0
# my dfa definition
s0->s1;char1,char2,char3
# nfa is same, with epsilon
# regex, dots used to simplify parsing
(a.b.c*)|a
implement efficent regex by compiling to my simple regex, then nfa, then dfa, then pruning dfa and building a FSM in C We can handle wild cards with special states in the DFA
feature
- powerset construction
- fix parser
- allow basic concat, union, kleene, and grouping
- add
+,? - add
()as capturing groups - add
(?:), supports non capturing groups - add
[] - add
. - non greedy matching
+?and*? - add
[^] - add
{n} - add
{min,} - add
{,max} - add
{min,max} - add
^and$, supports search vs match - back references, maybe doing a limited max of 9 or soemething
- possessive matching
- look ahead and look behind
- consider predefined character groups
- add support to extract groups
- add a search vs match mode
- jitted regex
software engineering todo
- add my suite of makefile and scripts
- separate out files
- ensure better const correctness
- toC, toNasm, etc using visiotor pattern
- add iterator support for states and transitions