Bug 101064 – eBay's new url format is to big for webalizer

• Product: Gentoo Linux
• Status: RESOLVED UPSTREAM

[Description William L. Thomson Jr. (RETIRED) 2005-08-02 00:04:44 UTC]

eBay has changed their url format for auctions, stores or etc. It's causing
webalizer to complain about oversized referrer fields and is truncating them.
Causing urls to be omitted from webalizers reports.

Reproducible: Always
Steps to Reproduce:
get Apache to log one of eBay's new long urls in the referrer field. Basically
what happens when you have an image, or etc on eBay and someone navigates around
eBay. The long urls are logged by Apache in the referrer field
Actual Results:  
When webalizer goes to parse apache's log file. Webalizer complains about
oversized referrer fields and is truncating them. In my case avg of like
3000-6000 per day.

Expected Results:  
webalizer should be able to parse the log data despite the size of eBay's new
urls. Without complaining, and truncating the referrer field.

I tried bringing this up to the webalizer community, but they do not believe it
to be a big deal. At least till it effects others, or etc? Not very proactive,
more like reactive. The change is very minimal and should not make much difference.

Basically webalizer.h needs to be modified.
#define MAXREF   1024                  /* Max referrer field size          */
#define MAXREFH  128                   /* Max referrer field size in htab  */

I am thinking the MAXREF should be 2048, and the MAXREFH be like 512. The reason
for the 512 is the old way was 256 for the referrer, and 128 for the rest of it,
vars etc.  So 512 should give it enough room for both?

[Comment 1 Jonathan Smith (RETIRED) 2005-08-23 20:43:29 UTC]

upstream says:

"Oh, as to your listed bug #101064, it isn't a bug at all.  Truncated fields are
 till processed, but with a warning.  No URLs are lost as indicated in the
report.  The correct fix is indeed to increase the size of the internal buffers
as reported, however it does no harm to leave them where they are and just
igonre the warnings."

[Comment 2 Jonathan Smith (RETIRED) 2005-10-29 17:47:45 UTC]

upstream has been notified and no data is actually lost