Forum OpenACS Q&A: Any problems found with Postgres 7.2?

Has anyone installed OpenACS4 with a new Postgres 7.2 database server
running? I was having some strange problems with the bootstrap
where the schema gets loaded. Just wanted to find out if anyone had
succeeded with 7.2 yet. I am probably having trouble with something
else (i.e., I'm using a new Slackware install that I have never
tried before)
Collapse
Posted by Vinod Kurup on
I installed OpenACS4 on PG 7.2rc2. There was a minor problem in the query-dispatcher that is now fixed, so if you're using a recent (< 2 wks old) CVS snapshot, it should work.
Collapse
Posted by Bart Teeuwisse on
Henry, the current CVS version does work with Postgres 7.2 with one exception: the intarray from the 7.2 RPMS doesn't. Intarray is needed only for the openfts driver and an easy fix is available. Check out the OpenACS 4.x and Postgres 7.2 compatability thread for more information.
Collapse
Posted by Jon Griffin on
I just installed on a completly fresh 7.2 and APM is broken on package loading. It doesn't like clob_dml.
Collapse
Posted by Vinod Kurup on
Actually, the APM is broken in 7.1 as well. Don fixed a race condition in acs-tcl/tcl/security-procs.tcl yesterday, but changed clob_dml to clob_insert_dml in one spot. Changing that back to clob_dml fixes part of the problem, but then a query is missing (prop_update_dml). Adding this to acs-tcl/tcl/security-procs.xql seems to fix that problem as well:
<fullquery name="ad_set_client_property.prop_update_dml">
      <querytext>

                update sec_session_properties
                set     property_value = :value,
                        last_hit = :last_hit,
                        secure_p = :secure
                where  session_id = :session_id
                and    module = :module
                and    property_name = :name

      </querytext>
</fullquery>
I haven't made these changes in CVS in case I interpreted something wrong.
Collapse
Posted by Don Baccus on
Sorry, I haven't had time to test this last change in PostgreSQL yet.  I wanted to get the change into the tree for Oracle while I was sitting there staring at it.

I'll clean up the PG version ASAP.

BTW this bug has been lurking in the ACS since the client property stuff was introduced in one of the ACS 3 versions.  Ugh.  It explains some of the sporadic and random-seeming bugs in this area folks have reported occassionally.  There may be similar errors elsewhere in the security token handling, I haven't had time to look further.

Fortunately the bug could be triggered reliably by pages which contain frames, as the browser then fires off two parallel requests that end up entering that code simultaneously, triggering the race condition.

Collapse
Posted by Don Baccus on
Oh ... I commited the .tcl file but not the query files ... this is easily fixed!
Collapse
Posted by Don Baccus on
OK, it's working again now, sorry for any inconvenience!