Re: [Patch, libstdc++/64441] Fix out of range match_results submatch access

[Jakub Jelinek <jakub@redhat.com>]

On Tue, Mar 10, 2015 at 02:13:49PM +0000, Jonathan Wakely wrote:
> On 9 March 2015 at 10:56, Tim Shen wrote:
> > I guess this patch doesn't break abi compatibility, so if everything
> > is Ok, I'm gonna patch it to 4.9 too.
> 
> It doesn't change the ABI directly, but it does change the layout of
> the match_results' data on the heap. It mean that instantiations of
> e.g. __regex_algo_impl compiled with GCC 4.9.2 will still call
> __res.resize(n+2) when poopulating a match_results whereas code
> compiled with 4.9.3 would expect to be able to find n+3 sub_match
> objects in the match_results object.
> 
> I think it's safer to not change the branch.
> 
> > I'm not sure if this is a "regression fix" though; if it's
> > inappropriate for trunk, then I can simply wait.
> 
> For the reasons given above, I don't think we want to change this
> between 5.1 and 5.2, but also don't want to wait for 6.0 to fix these
> bugs, so I've CC'd the release managers for their OK.
> 
> RMs: this fixes two bugs in std::regex, which was new in 4.9.0 and is
> still stabilising, as shown by this fix. For the reasons given above I
> think we should fix it on trunk now. The patch has my approval, OK to
> commit?

Ok for trunk.

	Jakub