Re: Bugs in OpenLDAP2.0 beta

["Kurt D. Zeilenga" <Kurt@OpenLDAP.org>]

A few additional comments:

At 02:37 PM 8/15/00 +0200, André Brands wrote:
>I am testing OpenLDAP 2.0 beta on our alpha machines running Unix Tru64.
>I use gcc in combination with the native linker.
>
>1. During make I had to supply an extra linker option, because of the
>non-default search behaviour of the linker.
>Configure generated a -pthread in the makefiles where it should be
>-lpthread.

Configure choose -pthread because it worked (including tests which
requiring linking).  Did you run configure with the same environment
settings that you used at make time?  What does your settings
(CC,CFLAGS,CPPFLAGS,LD,LDFLAGS, etc.)?

You should NOT screw with X* flags.  If you want to use -lpthread
instead of -pthread, force the pthread test to no as follows:
  env ol_cv_pthread_pthread=no ./configure
  

>The slow read performance was due to a problem with the socket option
>TCP_NODELAY which was not inherited by the socket created with an
>accept() call.

Please report a bug to your vendor.  The Tru64 accept(2) man page
reads:
  The accept() function extracts the first connection on the queue of pending
  connections, creates a new socket with the same properties as the specified
  socket, and allocates a new file descriptor for that socket.

which is contrary to your experience.