Re: RE24 connection code reworking

Aaron Richton wrote:

On Sun, 25 Jan 2009, Howard Chu wrote:

...

socket to become writable. What happened to the other servers at this point?

https://www.nbcs.rutgers.edu/~richton/test050logs.tgz

I think that con2 is trying to replicate back-config poorly, note in particular:

send_ldap_result: conn=-1 op=0 p=3

conn=-1? assuming that hasn't been retooled and still should be a monotonically incrementing counter, that's an interesting state...

conn=-1 is normal for a syncrepl consumer thread.

Based on these traces, I think this has been fixed with my last commit to result.c. Please update and try again...