Unresolved bugs

I’ve been catching up on the list and noticed a number of bug reports
that have not been resolved. Is it time for a bug tracker? I went ahead
and compiled a few examples here:

http://wiki.codemongers.com/NginxBugs

Feel free to organize, append to, or fix these bugs.

Evan

On Dec 29, 2007 6:40 PM, Evan M. [email protected] wrote:

I’ve been catching up on the list and noticed a number of bug reports
that have not been resolved. Is it time for a bug tracker? I went ahead
and compiled a few examples here:

http://wiki.codemongers.com/NginxBugs

Feel free to organize, append to, or fix these bugs.

Isn’t that SSL bug fixed in 0.6.24?

-bob

Bob Ippolito wrote:

-bob

Ah yes, that was fixed. I’ll scratch it off.

Evan

On Sat, 29 Dec 2007 18:50:09 -0500
“Bob Ippolito” [email protected] wrote:

Isn’t that SSL bug fixed in 0.6.24?
The 0.6.22-24 SSL stuff was to fix a memory leak IIRC. I recently
migrated a lighttpd installation to nginx 0.6.22 recently due to the
an occurence of this same bug with SSL connects in a single thread in
lighttpd. My tests with ab-2.0.40 in SSL mode worked with large
concurrent and aborted partial requests on nginx-0.6.22 for me.

On Sun, Dec 30, 2007 at 01:53:30AM +0100, Bryon Roche wrote:

Feel free to organize, append to, or fix these bugs.

Isn’t that SSL bug fixed in 0.6.24?
The 0.6.22-24 SSL stuff was to fix a memory leak IIRC. I recently
migrated a lighttpd installation to nginx 0.6.22 recently due to the
an occurence of this same bug with SSL connects in a single thread in
lighttpd. My tests with ab-2.0.40 in SSL mode worked with large
concurrent and aborted partial requests on nginx-0.6.22 for me.

There was no memory leak. It was memory framgentation in OpenSSL
built-in
session cache. 0.6.23 disables the cache by default. Also you might use
shared session cache.

Also, 0.6.23 has fixed SSL “bad write retry” errors,
and 0.6.24 has decreased SSL memory leaved in keep-alive connection.