FreeTDS Development Group

["Lowden, James K" <LowdenJK AT bernstein.com>]

> From: Sam Denton [mailto:denton AT wantec.com]
> Sent: May 23, 2002 1:59 PM

>> Is it okay to re-use CS_COMMAND structures:  i.e.
>> allocate them once, and then re-use them after
>> ct_results returns something other than CS_SUCCEED?

> Patient:  "Doctor, it hurts when I do *this*."
> Doctor:  "Well, then, don't do that any more."

Musta been a C doctor.  C++ doctors recommend *this, and tell you that
eventually you'll like it.  

> In ct_cmd_alloc:
>         *cmd = (CS_COMMAND *) malloc(sizeof(CS_COMMAND));
>         memset(*cmd,'\0',sizeof(CS_COMMAND));

OK.  I think both lines could be replaced by:

          *cmd = (CS_COMMAND *) calloc(1, sizeof(CS_COMMAND));

but maybe that wouldn't be any more efficient.  

> In ct_command:
>         if (cmd->query) free(cmd->query);
>         cmd->query = (char *) malloc(query_len + 1);

Assumes cmd->query will be null if it had been freed earlier.  

if (cmd->query) free... is an unnecessary test (in theory), because
according to free(3), "If ptr is NULL, no action occurs."  (What does yours
say?)

> In ct_cmd_drop:
>         if (cmd) {
>                 if (cmd->query) free(cmd->query);
>                 free(cmd);
>         }

Doesn't set cmd = NULL; may be freed again later, with disappointing
results.  

> In ct_dynamic:
>         if (cmd->query) free(cmd->query);
>         cmd->query = (char *) malloc(query_len + 1);

Same caveats as in ct_command.

NetBSD uses /etc/malloc.conf to help debug malloc issues.  Maybe your system
has something similar.  Might turn something up.  

Regards, 
--jkl