RE: zlib issue

FYI here is more detail on the work around.

The data that comes in from google is in gzip format and contains the gzip
header.  htzip.c seems to choke on the header piece, so I borrowed a piece
of code from zlib itself that reads past the header before calling
'inflate'.  With that fix in libwww seems to work fine.  I feel I am
definitely missing something as I can't believe others are not retrieving
gzip encoded data.

Thanks in advance for any I/O on this topic.

Fred Covely
BCF Technology
fcovely@bcftech.com
(B)760-631-8157
(C)760-717-9689

-----Original Message-----
From: www-lib-request@w3.org [mailto:www-lib-request@w3.org]On Behalf Of
Fred Covely
Sent: Friday, March 08, 2002 8:49 AM
To: www-lib@w3.org
Subject: zlib issue


Is anyone else have problems with sites that send back gzip encoded data?

For example, if I do a simple HTTP Get on www.google.com accepting gzip, I
get no output.  I have traced this to htzip.c where the call to the zlib
function 'inflate' returns -3 (bad gzip format).  By various manueverings I
can force it to accept the google zip data, but I have been unable to make
it work with 5.3.2.

Any ideas?  With PK, doing a release shortly, I'd like to get my proposed
fix in if it is in fact a real bug and not some screw up on my part.

Fred Covely
BCF Technology
fcovely@bcftech.com
(B)760-631-8157
(C)760-717-9689

Received on Friday, 8 March 2002 12:54:46 UTC