Matching of regexps with ^ could be much faster

[Eric Seidel (no email)]

Matching of regexps with ^ could be much faster

try /^bar/ on a really long string and watch Safari go off the deep end...  It should immediately be able to detect that ^bar doesn't match.

[Eric Seidel (no email)]

Created attachment 20049 [details]
Make matching of regexps using ^ much faster

 JavaScriptCore/pcre/pcre_compile.cpp |   20 ++++++++++++--------
 JavaScriptCore/pcre/pcre_exec.cpp    |   32 ++++++++++++++++++++++++--------
 JavaScriptCore/pcre/pcre_internal.h  |    2 ++
 3 files changed, 38 insertions(+), 16 deletions(-)

[Eric Seidel (no email)]

*** Bug 18087 has been marked as a duplicate of this bug. ***

[Eric Seidel (no email)]

Created attachment 20051 [details]
Test case which shows bug (warning, will hang Safari!)

[Eric Seidel (no email)]

I'm not sure where to put the test case.  I guess I could put it in LayoutTests/fast/js/
(and make it "pass" if it completes w/o timing out).

[Eric Seidel (no email)]

Created attachment 20054 [details]
Make matching of regexps using ^ much faster

 JavaScriptCore/pcre/pcre_compile.cpp               |   20 +++++++-----
 JavaScriptCore/pcre/pcre_exec.cpp                  |   32 +++++++++++++++-----
 JavaScriptCore/pcre/pcre_internal.h                |    2 +
 LayoutTests/ChangeLog                              |   10 ++++++
 .../fast/js/ascii-regexp-subject-expected.txt      |    5 +++
 LayoutTests/fast/js/ascii-regexp-subject.html      |   13 ++++++++
 .../fast/js/resources/ascii-regexp-subject.js      |   18 +++++++++++
 7 files changed, 84 insertions(+), 16 deletions(-)

[Darin Adler]

Comment on attachment 20054 [details]
Make matching of regexps using ^ much faster

r=me

[Darin Adler]

Assigning to Eric, assuming that he will check it in.

[Eric Seidel (no email)]

r31454