Jacques Caron <jc at oxado.com> writes:
> At 17:15 07/07/2008, Cyril SCETBON wrote:
>> it seems that it was not working without doing this action (adding
>> the v) even if our primary key is the first column.
>
> Yes, the current implementation of Slony needs a letter for each
> column telling it whether it's a key or a value. If you don't have
> enough letters in the argument, at best it will be missing columns in
> the update, at worst it will segfault. I sent a patch a while ago that
> allows you to add columns without having to update the trigger
> arguments, you should be able to find in the archives.

Jacques, can you take a look at the bug report in Bugzilla?  I have
attached a patch based on present CVS HEAD which also adds in a test
to exercise the code (which worked fine for me).

It appears that the changes to src/backend/slony1_funcs.sql have
already made it into HEAD, so all that needs to go in is the change to
the function that generates the logtrigger.  I changed it to use rtrim
rather than the full regular expression "change"; rtrim's a tad
simpler :-).

There's actually a little further change; the function rtrim needs,
for safety, to be referenced as pg_catalog.rtrim(), not just rtrim().
-- 
select 'cbbrowne' || '@' || 'linuxfinances.info';
http://cbbrowne.com/info/lsf.html
Rules  of the  Evil Overlord  #145. "My  dungeon cell  decor  will not
feature exposed pipes.  While they add to the  gloomy atmosphere, they
are good  conductors of vibrations and  a lot of  prisoners know Morse
code." <http://www.eviloverlord.com/>