Re: FastCGI and NT

Benjamin Gavin (gavinb@supranet.net)
Mon, 25 Aug 1997 13:48:56 -0500

Message-Id: <3.0.3.32.19970825134856.00950450@mail.supranet.net>
Date: Mon, 25 Aug 1997 13:48:56 -0500
To: fastcgi-developers@OpenMarket.com
From: Benjamin Gavin <gavinb@supranet.net>
Subject: Re: FastCGI and NT
In-Reply-To: <3.0.3.32.19970821124319.0096de90@mail.supranet.net>

-----BEGIN PGP SIGNED MESSAGE-----

Hi,
  I have looked into it further, and it seems that no matter what 
environment variables you change, the FastCGI program runs as if it 
had only the original set.  i.e. I submit a form with GET, then I 
submit the form as POST with different info, and the FastCGI program 
still only sees GET and the original Query String.  Also, I do GET, 
then GET again with different data, I still see the original GET and 
the same data.  Why is the FastCGI process not seeing the changes in 
the environment variables??  I have browsed through the source, and 
is the getenv function macro'ed out to use the FCGI_PARAMS array??  
If not, should it be??  It needs something to get those new 
environment variables.

Ben Gavin

At 12:43 PM 8/21/97 -0500, Benjamin Gavin wrote:
>-----BEGIN PGP SIGNED MESSAGE-----
>
>Hi,
>  I have just recently gotten FastCGI up and Running under NT.  The 
>programs seem to work fine, at least the example ones.  I do have 
one 
>small problem though.  I have written my own FastCGI program, runs 
>fine under regular CGI, and as a CGI under FastCGI.  However when I 
>try to run it as a FastCGI program there seems to be a bug in the 
>FastCGI libraries.  It seems that no matter how many times I call 
the 
>FastCGI program, I always get the same CONTENT_LENGTH value.  No 
>matter how long the content really is.
>  This seems to be set the first time the program is called.  It 
>first starts the program and passes it the environ variables.  Then 
>my program reads those variables and reads from the socket the 
amount 
>given by that CONTENT_LENGTH variable.  Well it works fine for the 
>first time, and for any subsequent time after that where the content-

>length is IDENTICAL, however, if I try to send too little, or too 
>much, it reads in garbage, or truncates the data respectively.
>  It seems that the cgi-fcgi shim application is not functioning 
>properly.  Has anyone seen this before, or is there a patch for this 

>problem??  I really need the performance gain of FastCGI, or I'll 
>have to convert to ISAPI.... YUCK!!
>
>Thanks in advance,
>Ben
>-----BEGIN PGP SIGNATURE-----
>Version: PGP for Personal Privacy 5.0
>Charset: noconv
>
>iQCVAwUBM/x+N0tkpl99enKxAQHTbwP/V6GVHVgCvZXJ8U6v8yxQEq0ZnYCZrYCw
>NLnOyqsI9XqqxS34gbWC4m9BGj+ue0vqw5a8y67/O5dAxaTDRYX6wl2hV+9fvTnt
>i/S1cTitqUzGcrJcDsGkv8Es/xAzhaf3DIPkQqX/Vowpj6ZoDA2BNlEw+hBJIMPq
>6ZyyrDHQKdY=
>=SbWS
>-----END PGP SIGNATURE-----
>
>
>
-----BEGIN PGP SIGNATURE-----
Version: PGP for Personal Privacy 5.0
Charset: noconv

iQCVAwUBNAHTl0tkpl99enKxAQFQkAQAgJTuPMGp5latf1tykAop3rUxpJVT4Cen
pP/po0ZfEiSgREFTf9K3h+3b9LphuUR0ON8OjiUAyf4KYe4vgGWB8bKN9JSoXY7v
NTLwBbcVgfWdfPpmIQDR1K1N7u6MLEuyClp6syLpCS1D9PGOfzfRsRv+GqOGhrZf
cqux6L9CXyg=
=3/Ug
-----END PGP SIGNATURE-----