Attached is a patch with some fixes that (I think that) should go into

6.4.1. Here is the list:

- The type int8 now works. In fact, the bug(s) were in
src/backend/port/snprintf.c, so int8 is probably broken in every platform
that hasn't a native snprintf/vsnprintf. The type itself worked as
expected, only the output was wrong. Anyway, this patch should be checked
in other platforms.

- The regression tests for int2 and int4, which were broken due to
differences in the error messages, are fixed.

- The regression test for float8, which was broken in the reference
platform, is also fixed. I don't know if the new file (float8-OSF1.out)
will work on other platforms, but it might be worth to try it.

- Two new template files are provided (alpha_cc, which includes
optimization, and alpha_gcc), and src/templates/.similar is updated
accordingly. src/templates/alpha should be removed from the distribution.
*IMPORTANT NOTE*: I don't know if you can use gcc to compile postgres;
I've written the alpha_gcc file because alpha_cc has some flags that are
specific to DEC C.

- There is a (very basic) Digital Unix specific FAQ in
doc/FAQ_DigitalUnix.

--
-------------------------------------------------------------------
Pedro José Lobo Perea                   Tel:    +34 91 336 78 19

doc/FAQ_DigitalUnix

@@ -0,0 +1,103 @@
+=======================================================
+Frequently Asked Questions (FAQ) for PostgreSQL  V6.4.1
+FreeBSD Specific
+TO BE READ IN CONJUNCTION WITH THE NORMAL FAQ
+=======================================================
+last updated:           Mon Dec 14 17:26:03 CET 1998
+
+current maintainer:     Pedro J. Lobo (pjlobo@euitt.upm.es)
+original author:        Pedro J. Lobo (pjlobo@euitt.upm.es)
+
+This FAQ covers issues that are specific for PostgreSQL running on Digital
+Unix (formerly known as DEC OSF/1). Please check the global FAQ for questions
+not specific to this platform.
+
+Changes in this version (* = modified, + = new, - = removed):
+
+This file is divided approximately as follows:
+
+1.*)	Installing PostgreSQL
+
+1.1.*)		Compiling PostgreSQL
+
+1.2.*)		Running the regression tests
+
+
+Questions answered:
+1.1.1)	I can't compile PostgreSQL with gcc.
+
+1.1.2)	DEC C dies with an internal error when optimization is
+		enabled (-O flag).
+
+1.2.1)	The regression tests fail for char, varchar, select_implicit,
+		select_having and rules.
+
+1.2.2)	The regression tests fail for abstime, tinterval and horology.
+
+1.2.3)	The regression tests fail for geometry.
+
+1.2.4)	The regression tests fail for inet.
+
+----------------------------------------------------------------------
+Section 1:      Installing PostgreSQL
+----------------------------------------------------------------------
+
+Section 1.1:	Compiling PostgreSQL
+------------------------------------
+
+1.1.1)	I can't compile PostgreSQL with gcc.
+
+		Me too O:-) I tried to compile it with gcc 2.7.2.1 without
+		success. However, I've found that DEC C works much better than
+		gcc for alphas, and so I didn't make any effort to find out
+		why gcc fails. If you succedd on this task, please take the
+		time to update this section (see http://www.postgresql.org for
+		the details on how to do this).
+
+1.1.2)	The DEC C compiler dies with an internal error when optimization is
+		enabled (-O flag).
+
+		I've seen this happening on 3.2c and 3.2g systems, and only on
+		one of the source files (can't remember exactly which one).
+		The workaround was to compile the offending module without
+		optimization by hand, and re-running gmake to let the process
+		continue.
+
+		It doesn't happen on 4.0d systems (mine at least). Don't know
+		about earlier 4.0 versions.
+
+Section 1.2:	Running the regression tests
+--------------------------------------------
+
+1.2.1)	The regression tests fail for char, varchar, select_implicit,
+		select_having and rules.
+
+		This only happens when you enable locale support, and is due
+		to a bug in the system's locale libraries. The problem is
+		that strcoll("Axxx", "axxx") returns a value greater than
+		zero instead of lower, which is how it should be.
+
+		This isn't a great problem unless you (or your application)
+		relies on upper-case letters being considered "smaller" than
+		lower-case for ordering purposes. Otherwise, you will only
+		notice that the ordering between upper-case and lower-case
+		letters is reversed.
+
+1.2.2)	The regression tests fail for abstime, tinterval and horology.
+
+		I think that these failures are due to some inconsistencies
+		in time zone handling in some years near 1950, in which case
+		the problem would be in the operating system's time zone
+		libraries. Recent dates seem to work as expected, but I can't
+		assure it.
+
+1.2.3)	The regression tests fail for geometry.
+
+		These are (small) rounding errors that shouldn't affect any
+		application (but could do, who knows).
+
+1.2.4)	The regression tests fail for inet.
+
+		Yes, they do. You must consider the inet type broken for
+		Digital Unix.
+