On 02/25/2008 10:45 PM, Joe Orton wrote:
On Mon, Feb 25, 2008 at 09:49:55PM +0100, Ruediger Pluem wrote:
On 02/25/2008 09:09 PM, [EMAIL PROTECTED] wrote:
Author: jorton
Date: Mon Feb 25 12:09:38 2008
New Revision: 630974

-    if (mc->szSessionCacheDataFile == NULL) {
-        ap_log_error(APLOG_MARK, APLOG_ERR, 0, s, "SSLSessionCache required");
-        return APR_EINVAL;
-    }
Why don't we check any if arg == NULL as a replacement for the if statement 
above?

Sorry, I should have mentioned this in the changelog message. In both cases of this, the code will always be passed a non-NULL arg parameter

Hm, but you are supplying sep + 1 which is not NULL but can point to '\0'.
Shouldn't we check for this?

unless a pstrdup fails (which by policy will be ignored anyway). I think these checks may have just been a copy'n'paste legacy from the shmcb code, which has genuine config argument parsing failure cases.

Again, thanks for the detailed review!

You are welcome.

Regards

RĂ¼diger

Reply via email to