开发者

Does Python use NFAs for regular expression evaluation in the re module?

Does anybody know if Python (any version) used NFAs (Non-Deterministic Finite Automata) to evaluate regular expressions or does it use some other mechanism? Please provide links/refe开发者_如何学Pythonrence if available.


This should take less than a ms on a DFA:

$ time python3 -c 'import re; re.match("a?"*25+"a"*25, "a"*25)'
real    0m7.273s

Change 25 with 100 and it won't terminate for a lifetime.

Here is how it looks on a DFA (grep):

$ time echo "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa" |grep "a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?a\?aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa"
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
real    0m0.063s

There is a great discussion of the topic at http://swtch.com/~rsc/regexp/regexp1.html


NFA.

See Friedl's Mastering Regular Expressions, 3rd edition, chapter 4 - table 4-1, page 145.

Google books has a preview to it.

0

上一篇:

下一篇:

精彩评论

暂无评论...
验证码 换一张
取 消

最新问答

问答排行榜