Re: pset_quoted_string is broken
| От | David Rowley |
|---|---|
| Тема | Re: pset_quoted_string is broken |
| Дата | |
| Msg-id | CAApHDvrjCULr1NH75BK6qso70K8vQeooo0r3RTv7eqSD2iz_pA@mail.gmail.com обсуждение исходный текст |
| Ответ на | Re: pset_quoted_string is broken (Tom Lane <tgl@sss.pgh.pa.us>) |
| Список | pgsql-hackers |
On Mon, Oct 27, 2014 at 12:20 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
David Rowley <dgrowleyml@gmail.com> writes:
> It seems the buffer created in pset_quoted_string is just 1 char too small.
Yeah, that's a bug. Fix pushed, thanks!
Thanks for committing.
> This breaks psql's \pset for me, though I've no idea why the buildfarm is
> not complaining a bit more.
I think in most cases, maxalign padding of the malloc allocation would
result in there being room for another byte without trashing anything
important. You must be using a libc that notices and complains about
even 1-byte buffer overruns.
I'm using MSVC.
After a bit of reading it seems like when compiled in debug mode that malloc() uses something called _malloc_dbg() which allocates a bit more memory to allow for more strict checking of buffer overruns.
I guess all the MSVC buildfarm members must be compiled in release mode then? I wonder if it would be worth changing one to build with debug as it seem like none of the buildfarm animals picked this up despite there being a regression test to ensure \pset works.
Regards
David Rowley
В списке pgsql-hackers по дате отправления: