Re: Tablespace for temporary objects and sort files

Hi,
here's a new version of the patch against HEAD with both, table and sort
files working correctly for me. Regression tests work too.
I'd like to ask again the question I made on the first post as no answer was
given at that time:

"The GetTempTablespace function correctly returns a different tablespace
each time is called, but I store the position of the last tablespace used
with an integer and iterate through the list of tablespaces each time. I
tried to keep the iterator from call to call but I got a segfault, I
imagine due to the memory context. Should I try to keep the iterator? How
can I do it?"

Now I'm working on some regression tests that could be added to
tablespace.source using something like:

SET temp_tablespaces='testspace';

CREATE TEMP TABLE temp_foo(a VARCHAR);

SELECT COUNT(pg_ls_dir('pg_tblspc/' || (SELECT oid FROM
pg_catalog.pg_tablespace WHERE spcname='testspace' .... )

Do you think it's a good idea to list the files in the directory of the
tablespace to ensure temporary table files are created where they should? Do
you think there is a smart way to ensure the same with sort files?

Any feedback welcome!

A Dimecres 25 Octubre 2006 00:45, Albert Cervera Areny va escriure:
> Hi,
>
> I'm trying to introduce myself into postgresql development and I'm working
> on the "tablespace for temporary objects and sort files" TODO item. The
> attached patch shows what I've already done. The GUC is currently
> called "temp_tablespaces".
>
> The tablespace changes correctly for me when creating temporary tables.
> I've got some questions though:
>
> How can I test that the tablespace is correctly used for sort files? Is
> there an easy way? Or should I reduce work_mem to a minimum, populate the
> database with data and try an "ORDER BY"?
>
> The GetTempTablespace function correctly returns a different tablespace
> each time is called, but I store the position of the last tablespace used
> with an integer and iterate through the list of tablespaces each time. I
> tried to keep the iterator from call to call but I got a segfault, I
> imagine due to the memory context. Should I try to keep the iterator? How
> can I do it?
>
> Hope the diff and idents are ok. Please let me know if there's something
> wrong with them.
>
> Thanks!