https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #18 from Krystian Nowak ---
Thanks for https://svn.apache.org/viewvc?view=revision=1902732 for
those pathological patterns, Yann!
When +/- will you be planning to port it to 2.4 branch?
--
You are receiving this mail because:
You
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #17 from Krystian Nowak ---
(In reply to Yann Ylavic from comment #8)
> Created attachment 38318 [details]
> Honor nmatch for the vector passed to pcre_exec()
Yann, will your patch be planned to get to 2.4 branch maybe?
--
You
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #16 from Krystian Nowak ---
Additionally, Alpine's package (by default) still uses PCRE1:
https://git.alpinelinux.org/aports/tree/main/apache2/APKBUILD#n15
--
You are receiving this mail because:
You are the assignee for the bug.
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #15 from Yann Ylavic ---
> Does it crash when 2.4.54 is compiled against PCRE2?
No it does not, PCRE2 uses its own (re)allocation logic in the passed in
pcre2_match_data context, and stack usage is better constrained.
It's still
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #14 from Ruediger Pluem ---
Does it crash when 2.4.54 is compiled against PCRE2?
--
You are receiving this mail because:
You are the assignee for the bug.
-
To
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #13 from Krystian Nowak ---
And you are right Yann, in 2.4.51 it segfaults with:
AH00052: child pid 44 exit signal Segmentation fault (11)
when having configured:
RedirectMatch "^((?!/errors/).)*$" "http://www.example.com$1
and
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #12 from Krystian Nowak ---
Got it, so at least this patch fixes the NULL vector case (as in this specific
issue) - so then I can't wait being it applied and merged to 2.4 branch for
next release - thanks for finding the fix for
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #11 from Yann Ylavic ---
As I said:
> but there is nothing we can do at httpd level to prevent some (badly
> written) regexes from exhausting the stack in all cases, notably when
> captures are to be used like in the RedirectMatch
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #10 from Krystian Nowak ---
Yann, thanks for the patch!
It seems it got fixed for the case described in this issue (after the patch it
stopped segfaulting), but not in the case in issue 66021 (that one still fails
with segfault).
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
Yann Ylavic changed:
What|Removed |Added
CC||szymek@gmail.com
--- Comment #9
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #8 from Yann Ylavic ---
Created attachment 38318
--> https://bz.apache.org/bugzilla/attachment.cgi?id=38318=edit
Honor nmatch for the vector passed to pcre_exec()
Since 2.4.52 (r1898467) we always provide a cached (per-thread)
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #7 from Krystian Nowak ---
(In reply to Ruediger Pluem from comment #6)
> > > Can you please modify the regex from
> > > ^((?!/errors/).)*$
> > >
> > > to
> > >
> > > ^(?!/errors/)(.*)$
> > >
> I understand, but for further
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #6 from Ruediger Pluem ---
(In reply to Krystian Nowak from comment #5)
> (In reply to Ruediger Pluem from comment #4)
> > Can you please modify the regex from
> > ^((?!/errors/).)*$
> >
> > to
> >
> > ^(?!/errors/)(.*)$
> >
>
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #5 from Krystian Nowak ---
(In reply to Ruediger Pluem from comment #4)
> Can you please modify the regex from
> ^((?!/errors/).)*$
>
> to
>
> ^(?!/errors/)(.*)$
>
> This is not about discussing which regular expression could
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #4 from Ruediger Pluem ---
Can you please modify the regex from
^((?!/errors/).)*$
to
^(?!/errors/)(.*)$
This is not about discussing which regular expression could be better or more
"correct". It should check if just the
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
--- Comment #3 from Krystian Nowak ---
(In reply to Eric Covener from comment #1)
> >
>
> Do you create 1 capture per character on purpose, or did intend the * to
> come after the wildcard?
It is a similar case (regarding capture groups) as
https://bz.apache.org/bugzilla/show_bug.cgi?id=66119
Krystian Nowak changed:
What|Removed |Added
Summary|Segmentation fault in |Segmentation fault in
17 matches
Mail list logo
