[Bug regex/20095] parse_dup_op duplicates the tree exponentially when using repeated +

[Bug regex/20095] parse_dup_op duplicates the tree exponentially when using repeated +

[dpmendenhall at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20095

David Mendenhall <dpmendenhall at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dpmendenhall at gmail dot com

--- Comment #2 from David Mendenhall <dpmendenhall at gmail dot com> ---
*** Bug 25814 has been marked as a duplicate of this bug. ***

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug regex/20095] parse_dup_op duplicates the tree exponentially when using repeated +

[jwakely.gcc at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20095

Jonathan Wakely <jwakely.gcc at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jwakely.gcc at gmail dot com

--- Comment #3 from Jonathan Wakely <jwakely.gcc at gmail dot com> ---
This behaviour can rapidly exhaust memory (Bug 25814, Bug 28864, Bug 20095
, Bug 29642), which seems unhelpful when ".++" is not even a valid regex. POSIX
says it's undefined:
https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap09.html#tag_09_04_06

Why doesn't regcomp just fail to compile it with REG_BADRPT?

Similarly for ".**" etc.

GNU grep seems to have tests for these that expect BADRPT:
https://git.savannah.gnu.org/cgit/grep.git/tree/tests/tests?h=v3.11#n234

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug regex/20095] parse_dup_op duplicates the tree exponentially when using repeated +

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=20095

Adhemerval Zanella <adhemerval.zanella at linaro dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |adhemerval.zanella at linaro dot o
                   |                            |rg

--- Comment #4 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
I am not sure who exactly GNU grep handles this since it also uses gnulib regex
code. Is this code really being tested by GNU grep? I deleted the fil, make
check, and it seems not to affect the test's outcome.

I also tried to sync with gnulib master, but it also does not help.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug regex/20095] parse_dup_op duplicates the tree exponentially when using repeated +

[jwakely.gcc at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=20095

--- Comment #5 from Jonathan Wakely <jwakely.gcc at gmail dot com> ---
(In reply to Adhemerval Zanella from comment #4)
> Is this code really being tested by GNU grep?

No idea - I grepped the grep code for cases of '++' and found those tests, I
don't know if they're actually run or not.

FWIW, Solaris 2.11 and AIX 7.3 both have the same behaviour for "a++"

jwakely@gcc-solaris11:~$ /usr/xpg4/bin/grep -E  a++ <<< a
a

$ /usr/bin/grep -E  a++ <<< a
a


So maybe POSIX says it's undefined to allow for this traditional/common
behaviour.

Glibc's support for it seems poor though, given the memory exhaustion problems.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug regex/20095] parse_dup_op duplicates the tree exponentially when using repeated +

[sh200105 at mail dot ru]

https://sourceware.org/bugzilla/show_bug.cgi?id=20095

Alexander Kernozhitsky <sh200105 at mail dot ru> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sh200105 at mail dot ru

-- 
You are receiving this mail because:
You are on the CC list for the bug.