let's roll up rc2 ..

> * Increase identifier length (NAMEDATALEN) if small performance hit,
>   perhaps to standard length of 128;  change struct pgNotify to use pid
>   first, breaks notify API;

> * Increase identifier length (NAMEDATALEN) if small performance hit;
>   change struct pgNotify to use pid first, breaks notify API

> * Evaluate AIX cs() spinlock macro for performance optimizations (Tatsuo)

cturner@redhat.com.

It would try to call the input conversion routines for them anyway.
So, a valid input string for the datatype had to be supplied.

allows plperl routines to return NULL as intended.

removes any empty chunks, the chunk previously added won't be there
anymore, so it's possible there is zero free space in the rel's page list
afterwards.  Must loop back and rerun the part that adds a chunk to
the list.

value of TCL_SHLIB_LD in the shell script that creates Makefile.tcldefs,
rather than trying to persude gmake to do it.

> * Reorder postgresql.conf WAL items in order of importance
> * Remove wal_files postgresql.conf option because WAL files are now recycled
> * Find proper defaults for postgresql.conf WAL entries
> * Add checkpoint_min_warning postgresql.conf option to warn about checkpoints
>   that are too frequent

<   http://www1.distributed.net/source, in client/common/cpucheck.cpp
---
>   from distributted.net, http://www1.distributed.net/source,
>   in client/common/cpucheck.cpp

> * Add code to detect an SMP machine and handle spinlocks accordingly

compile in client apps that use the standard installed header set.
To allow removing that include, move DLLIMPORT definitions out of c.h
and into the appropriate port-specific header files.

okay, sorry for delay all ... here is the tag for RC1 ...

options.  Document --with-pam and --enable-depend, do some minor
copy-editing.

> * Make IN/NOT IN have similar performance to EXISTS/NOT EXISTS [exists]

choke at 2Gb of output.  From Andrew Sullivan.

> * Use thread-safe crypt() in libpq, if available

tom lane