Re: Error handling for ShmemInitStruct and ShmemInitHash

Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> none of the other half are printing messages that are more useful
> than "out of shared memory" (which isn't even necessarily
> correct).

I think the messages in the locking area are a bit more useful than
"out of shared memory", but it would be trivial to build the
equivalent message in the ShmemInitHash function, based on the first
parameter.

LockMethodProcLockHash = ShmemInitHash("PROCLOCK hash",
init_table_size,
max_table_size,
&info,
hash_flags);
if (!LockMethodProcLockHash)
elog(FATAL, "could not initialize proclock hash table");

Presumably the ShmemInitHash function could add other information
which would make the message *more* useful. (Perhaps other
parameter information or maybe even the actual *cause* of the
failure.)

> I suggest making these functions throw their own errors rather
> than returning NULL on failure, and removing the redundant error
> reports from the callers that have 'em.

+1 It would be low priority if the return value was currently being
consistently checked for NULL; but since that's not the case we have
to do something, and what you suggest sounds best, long term.

-Kevin