This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [Patch] Change default executor to DFS in regex


On 10/19/2013 05:53 AM, Tim Shen wrote:
On Fri, Oct 18, 2013 at 9:13 PM, Tim Shen <> wrote:
As the comment in this patch said, DFS approach is faster in simple
regex, but exponentially slower in complex(many quantifier) cases.
Actually I suggest to use DFS where number of quantifiers < 2, to make
this 'optimization' more conservative. Now `split regex`, say "\s+"
can be optimized.
The idea of course makes sense to me, but I think it needs a few small tweaks. First, I don't like much the name _S_anti_auto, sounds like something which goes against auto (?). What about _S_alternate?

The comment:

The _S_auto policy now is as coded,

should probably be:

The _S_auto policy is the following:

In the text, remember to always put a space before open parentheses, it adds clarity.

About the < 2, in general hardcoding a parameter value in the code isn't a nice idea. Why don't we take it out to a macro, say _GLIBCXX_REGEX_NFA_QUANTIFIERS_LIMIT? In stl_deque.h we have something similar and in the present case it would be even safe from the ABI point of view, if I'm not mistaken.

Finally, you mentioned the DFA optimization: I would like to see some details about it. Would it be a big win in all cases? Before we freeze the next ABI we should keep open all the possibilities...


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]