The announcement of the 1.1.5 release candidates asked the packagers to check
for packaging issues. Here are my notes:
Regarding the bugs that I had submitted earlier to the tracker:
1471 (exit code on slon man page) is fixed but I think the code is actually
wrong. Calling exit() with negative status is unportable.
1472 (bison/flex) seems OK.
1473, 1474 (files needlessly deleted) is not fixed.
Now additional notes:
- The directory autom4te.cache/ should be removed before packaging. It should
probably be cleaned in the distclean target or similar.
- Finding pg_config doesn't work quite as I'd wish. If I specify
--with-pgconfigdir, then configure should fail if it's not there, and it
should not search other locations. (With multiple PG versions installed,
this can easily lead to tears.) I have attached a simple patch for this,
which you may want to rewrite to be more portable shell-wise.
- configure looks for plpgsql.so during the build but does not actually need
it during the build. This is a bit annoying as it requires you to install
the server package when you don't need it.
- There are a number of compiler warnings (GCC 4.0.3) that might be nice to
clean up.
- Slony-I-concept.pdf does not build. The file is truncated after a few
pages. Does one need special versions of the tools here?
ps2pdf Slony-I-concept.ps
Error: /undefined in 0,9
Operand stack:
Execution stack:
%interp_exit .runexec2 --nostringval-- --nostringval--
--nostringval-- 2 %stopped_push --nostringval-- --nostringval--
--nostringval-- false 1 %stopped_push 1 3 %oparray_pop 1 3
%oparray_pop 1 3 %oparray_pop 1 3 %oparray_pop .runexec2
--nostringval-- --nostringval-- --nostringval-- 2 %stopped_push
--nostringval--
Dictionary stack:
--dict:1121/1686(ro)(G)-- --dict:0/20(G)-- --dict:81/200(L)--
--dict:62/120(L)--
Current allocation mode is local
Last OS error: 2
Current file position is 259628
ESP Ghostscript 815.01: Unrecoverable error, exit code 1
make[3]: *** [Slony-I-concept.pdf] Fehler 1
- Building Slony-I-implementation gives the following messages. The result
looks OK, but it's a bit scary anyway.
groff -Tlatin1 -U -t -p -ms -mpspic Slony-I-implementation.nr
>Slony-I-implementation.txt
Slony-I-implementation.nr:31: warning: can't find font `TR'
grotty:Slony-I-implementation.nr:7: character above first line discarded
grotty:Slony-I-implementation.nr:15: X command without `tty:' tag ignored
grotty:Slony-I-implementation.nr:18: X command without `tty:' tag ignored
grotty:Slony-I-implementation.nr:20: X command without `tty:' tag ignored
grotty:figure-2.pic:4354: Unknown color (65535, 65535, 65535) mapped to
default
grotty:figure-2.pic:4367: Unknown color (65535, 65535, 65535) mapped to
default
[dozens of times]
- The file HISTORY-1.1 seems to be outdated. I doubt it's necessary to keep
it.
Other patches that I needed to apply: (I think I mentioned them before, but
just for completeness.)
- Patch to disable the search for postgresql.conf.sample
- Patch to disable rpath
Anyway, I think the two critical issues are finding the right pg_config and
the failing documentation build (which worked a few weeks ago I think...).
diff -Nur slony1-1.1.5-rc2/configure slony1-1.1.5-rc2.new/configure
--- slony1-1.1.5-rc2/configure 2005-12-15 10:23:15.000000000 +0100
+++ slony1-1.1.5-rc2.new/configure 2005-12-15 10:24:15.000000000 +0100
@@ -6065,7 +6065,7 @@
echo "$as_me:$LINENO: checking for pg_config" >&5
echo $ECHO_N "checking for pg_config... $ECHO_C" >&6
-PGCONFIG_POSSIBLE_LOCATIONS="${with_pgconfigdir} /usr/local/pgsql/bin /usr/local/bin /usr/bin /bin /usr/lib/postgresql/bin /opt/local/pgsql/bin /opt/pgsql/bin/"
+PGCONFIG_POSSIBLE_LOCATIONS=${with_pgconfigdir:-'/usr/local/pgsql/bin /usr/local/bin /usr/bin /bin /usr/lib/postgresql/bin /opt/local/pgsql/bin /opt/pgsql/bin/'}
for i in $PGCONFIG_POSSIBLE_LOCATIONS; do
if test -x $i/pg_config; then
PG_CONFIG_LOCATION="$i/pg_config"
_______________________________________________
Slony1-general mailing list
[email protected]
http://gborg.postgresql.org/mailman/listinfo/slony1-general
