BUG #18797: pointer te->defn not checked for NULL
The following bug has been logged on the website:
Bug reference: 18797
Logged by: Daniel Elishakov
Email address: dan-eli@mail.ru
PostgreSQL version: 16.6
Operating system: ubuntu 20.04
Description:
src/interfaces/ecpg/pgtypeslib/dt_common.c:2662
src/interfaces/ecpg/pgtypeslib/dt_common.c:2787
src/interfaces/ecpg/pgtypeslib/dt_common.c:2795
src/interfaces/ecpg/pgtypeslib/dt_common.c:2840
in function PGTYPEStimestamp_defmt_scan there is a usage of pgtypes_alloc().
In other functions the result of pgtypes_alloc is being checked for NULL.
For example, in /src/interfaces/ecpg/pgtypeslib/numeric.c. I think in
PGTYPEStimestamp_defmt_scan it is also worth checking for NULL before
dereferencing in strcpy, because PGTYPEStimestamp_defmt_scan is being called
recursively and if there won't be any available memory to allocate, it can
be a real problem in terms of data consistency.
PG Bug reporting form <noreply@postgresql.org> writes:
in function PGTYPEStimestamp_defmt_scan there is a usage of pgtypes_alloc().
In other functions the result of pgtypes_alloc is being checked for NULL.
Thanks for the report! It looks like this was already dealt with
during the v17 cycle:
Author: Michael Paquier <michael@paquier.xyz>
Branch: master Release: REL_17_BR [0a9118ccc] 2024-02-19 09:05:51 +0900
ecpg: Fix error handling on OOMs when parsing timestamps
pgtypes_alloc() can return NULL when failing an allocation, which is
something that PGTYPEStimestamp_defmt_asc() has forgotten about when
translating a timestamp for 'D', 'r', 'R' and 'T' as these require a
temporary allocation.
This is unlikely going to be a problem in practice, so no backpatch is
done.
Author: Oleg Tselebrovskiy
Discussion: /messages/by-id/bf47888585149f83b276861a1662f7e4@postgrespro.ru
I tend to agree with Michael's conclusion that an error here is
improbable enough to not justify a back-patch.
regards, tom lane