Re: Memory bug in dsnowball_lexize

Mark Dilger <hornschnorter(at)gmail(dot)com> writes:
> On Thu, May 23, 2019 at 8:46 AM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> Mark Dilger <hornschnorter(at)gmail(dot)com> writes:
>>> In src/backend/snowball/dict_snowball.c, 'dsnowball_lexize'
>>> calls 'SN_set_current' and ignores the return value, thereby
>>> failing to notice the error, if any.

>> Hm. This seems like possibly a bug, in that even if we cover the
>> malloc issue, there's no API guarantee that OOM is the only possible
>> reason for reporting failure.

> Ok, that sounds fair. Since the memory is being palloc'd, I suppose
> it would be safe to just ereport when the return value is -1?

Yeah ... I'd just make it an elog really, since whatever it is
would presumably not be a user-facing error.

>> Fair complaint --- do you want to propose some new wording that
>> references what header.h does?

> Perhaps something along these lines?

Seems reasonable, please include in patch covering the other thing.

regards, tom lane