Maybe you would want to add to the documentation / change log that DFA
matching may now use more stack than Perl matching. At least to me this
is somewhat surprising because we have so far been used to the exact
opposite. But it is a nice showcase of the good work you have done!
Ralf
On 15.03.2017 18:20, ph10@??? wrote:
>> Nevertheless, this test case from testinput6 still uses lots of
>> stack (and exceeds it on a Windows machine): > Indeed, but that is using pcre2_dfa_match(), which has *not* been
> changed. I don't know how widely used pcre2_dfa_match() is used, but
> I'm sure it is far less than the main pcre2_match() function, which
> is the one that gives Perl-compatibility. It would be possible to
> refactor pcre2_dfa_match() in due course if there is enough demand,
> but it uses function recursion much less than pcre2_match() used to,
> and mainly for handling recursion within the pattern.
