Possible NULL pointer deferenced (src/interfaces/libpq/fe-exec.c (line 563)
Hi Mark,
please, can you take a look?
This possible bug was appeared before, see at:
1. https://bugzilla.redhat.com/show_bug.cgi?id=879803
The trap still persist, in HEAD see:
src/interfaces/libpq/fe-exec.c (line 563)
/* If there's enough space in the current block, no problem. */
if (nBytes <= (size_t) res->spaceLeft)
{
space = res->curBlock->space + res->curOffset;
res->curOffset += nBytes;
res->spaceLeft -= nBytes;
return space;
}
The res->curBlock pointer possibly, can be NULL here (line 563).
See at:
src/interfaces/libpq/fe-exec.c (line 585)
if (res->curBlock)
The path is res->curBlock be NULL and res->spaceLeft > nBytes.
If res->curBlock it not can be NULL, inside pqResultAlloc function, why is
verified against NULL at line 585?
regards,
Ranier Vilela
Ranier Vilela <ranier.vf@gmail.com> writes:
The res->curBlock pointer possibly, can be NULL here (line 563).
No, it can't.
To get to that line, nBytes has to be > 0, which means res->spaceLeft
has to be > 0, which cannot happen while res->curBlock is NULL.
regards, tom lane
Em dom., 21 de jun. de 2020 às 02:16, Tom Lane <tgl@sss.pgh.pa.us> escreveu:
Ranier Vilela <ranier.vf@gmail.com> writes:
The res->curBlock pointer possibly, can be NULL here (line 563).
No, it can't.
To get to that line, nBytes has to be > 0, which means res->spaceLeft
has to be > 0, which cannot happen while res->curBlock is NULL.
Hi Tom, thanks for answer.
regards,
Ranier Vilela