HTTP/1.1 100 Continue

[ra...@bellglobal.com]

I think my previous mail on this was lost due to / filling up on taz earlier.

I was wondering if someone could explain this 100 Continue header that is
forced out from the should_client_block() function.  When I do a POST to
a page with IE4 I get these showing up at the top of the document.  I may be
calling should_client_block() incorrectly in my module somehow.

I am using the following bit of code to read post data:

    if (should_client_block(GLOBAL(php3_rqst))) {
        void (*handler) (int);
        int dbsize, len_read, dbpos = 0;
 
        hard_timeout("copy script args", GLOBAL(php3_rqst));
        handler = signal(SIGPIPE, SIG_IGN);     /* Ignore sigpipes for now */
        while ((len_read = get_client_block(GLOBAL(php3_rqst), argsbuffer,
HUGE_STRING_LEN)) > 0) {
            if ((dbpos + len_read) > length)
                dbsize = length - dbpos;
            else
                dbsize = len_read;
            reset_timeout(GLOBAL(php3_rqst)); /* Make sure we don't timeout */
            memcpy(buf + dbpos, argsbuffer, dbsize);
            dbpos += dbsize;
        }
        signal(SIGPIPE, handler);   /* restore normal sigpipe handling */
        kill_timeout(GLOBAL(php3_rqst));    /* stop timeout timer */
    }

Right when should_client_block() is called the HTTP/1.1 100 Continue header is
generated and flushed to the browser.  It looks like IE4 doesn't know what to
do with this header, or it may be because it is being sent in the wrong
sequence.

-Rasmus