Re: fastcgi and ssi

James M. Golovich (james@mail.netlite.com)
Wed, 9 Oct 1996 12:33:14 -0400 (EDT)

Date: Wed, 9 Oct 1996 12:33:14 -0400 (EDT)
From: "James M. Golovich" <james@mail.netlite.com>
To: Mark Brown <mbrown@openmarket.com>
Subject: Re: fastcgi and ssi
In-Reply-To: <199610091336.JAA02346@breckenridge.openmarket.com>
Message-Id: <Pine.LNX.3.91.961009123053.25910A-100000@piglet.pooh-corner.com>

I guess I didn't make my question clear.  Or actually I don't think I 
even really asked a question.  But The BeginHeader function is what is 
printing the headers, and I was wondering if it would break anything if I 
hacked it so these headers werent being printed..



James

# from mod_fastcgi.c

void BeginHeader(FastCgiInfo *reqPtr, char *msg)
{
    time_t now;

    ASSERT(BufferLength(reqPtr->reqOutbufPtr) == 0);

    now = time(NULL);
    SendHeader(reqPtr, SERVER_PROTOCOL, " ", msg, NULL);
    SendHeader(reqPtr, "Date: ", HTTPTime(gmtime(&now)), NULL);
    SendHeader(reqPtr, "Server: ", SERVER_VERSION, NULL);
    SendHeader(reqPtr, "MIME-version: 1.0", NULL);

    /*
     * We shouldn't have run out of space in the output buffer.
     */
    ASSERT(BufferFree(reqPtr->reqOutbufPtr) > 0);
}



On Wed, 9 Oct 1996, Mark Brown wrote:

> I suspect that the problem is that mod_fastcgi is generating these
> headers itself (see the function BeginHeader) rather than using
> Apache facilities.  I'll put this on the list to look at.
> If you are interested you could try to patch it yourself
> by studying how mod_cgi generates response headers and making
> mod_fastcgi as similar as possible.
> 
>     --mark
>