Forum OpenACS Q&A: OpenACS 5.1, dotLRN 2.0 branches merged to HEAD

I merged all the work on the dotlrn-2-0 and oacs-5-1 branch to head. The merge point tag is jcd-merge-20040420.

I have not yet tested that HEAD still installs so if there is a problem with installing that is probably the cause.

All the changed files

Collapse
Posted by Ola Hansson on
Hi Jeff,

Is it to late to commit bug fixes to 5.1? You see, I caught a couple of problems with some of the upgrade scripts when I completed the upgrade of infogettable.net from 4.5b to openacs-5-0-compat yesterday:

Several upgrade scripts in acs-kernel drop and recreate the views "registered_users" and "cc_user" ... However, they can't simply be dropped but need to be dopped with CASCADE (I'm on PG 7.3.5 in case that matters) since "registered_users_of_package_id" and "cc_users_of_package_id" are referencing the base views respectively.

ETP defines a new content_type twice in one upgrade script (see the diff below).

If this is not too late and seems OK with Joel and you, which branch should I commit it to?

Index: upgrade-4.6.3-4.6.4.sql
===================================================================
RCS file: /cvsweb/infogettable/packages/acs-kernel/sql/postgresql/upgrade/upgrade-4.6.3-4.6.4.sql,v
retrieving revision 1.1.1.1
retrieving revision 1.2
diff -u -r1.1.1.1 -r1.2
--- upgrade-4.6.3-4.6.4.sql     18 Apr 2004 15:55:09 -0000      1.1.1.1
+++ upgrade-4.6.3-4.6.4.sql     20 Apr 2004 12:51:30 -0000      1.2
@@ -5,7 +5,7 @@
 --
 -- @author Lars Pind (lars@collaboraid.biz)
 -- @creation-date 2003-05-28
--- @cvs-id $Id: upgrade-4.6.3-4.6.4.sql,v 1.1.1.1 2004/04/18 15:55:09 nsadmin Exp $
+-- @cvs-id $Id: upgrade-4.6.3-4.6.4.sql,v 1.2 2004/04/20 12:51:30 nsadmin Exp $
 --


@@ -26,7 +26,7 @@

 -- recreate the registered_users view

-drop view registered_users;
+drop view registered_users CASCADE;

 create view registered_users
 as
@@ -41,10 +41,15 @@
   and mr.member_state = 'approved'
   and u.email_verified_p = 't';

+create view registered_users_of_package_id
+as
+SELECT u.*, au.package_id
+FROM application_users au, registered_users u
+WHERE (au.user_id = u.user_id);

 -- recreate the cc_users view

-drop view cc_users;
+drop view cc_users CASCADE;

 create view cc_users
 as
@@ -59,6 +64,12 @@
   and m.rel_id = mr.rel_id
   and m.container_id = m.group_id;

+create view cc_users_of_package_id
+as
+SELECT u.*, au.package_id
+FROM application_users au, cc_users u
+WHERE (au.user_id = u.user_id);
+
 -- Fixing a really lame = null bug in this proc that would cause default values
 -- of new parameters to not be propagated to package instances
 create or replace function apm__register_parameter (integer,varchar,varchar,varchar,varchar,varchar,varchar,integer,integer)
Index: upgrade-5.0.0b3-5.0.0b4.sql
===================================================================
RCS file: /cvsweb/infogettable/packages/acs-kernel/sql/postgresql/upgrade/upgrade-5.0d2-5.0d3.sql,v
retrieving revision 1.1.1.1
retrieving revision 1.2
diff -u -r1.1.1.1 -r1.2
--- upgrade-5.0d2-5.0d3.sql     18 Apr 2004 15:55:10 -0000      1.1.1.1
+++ upgrade-5.0d2-5.0d3.sql     20 Apr 2004 12:51:30 -0000      1.2
@@ -3,7 +3,7 @@
 --
 -- @author Peter Marklund (peter@collaboraid.biz)
 --
--- @cvs-id $Id: upgrade-5.0d2-5.0d3.sql,v 1.1.1.1 2004/04/18 15:55:10 nsadmin Exp $
+-- @cvs-id $Id: upgrade-5.0d2-5.0d3.sql,v 1.2 2004/04/20 12:51:30 nsadmin Exp $
 --

 -- ****** New authentication datamodel
@@ -237,7 +237,7 @@
 alter table users add constraint users_authority_username_un
                       unique (authority_id, username);

-drop view cc_users;
+drop view cc_users CASCADE;
 create view cc_users
 as
 select o.*, pa.*, pe.*, u.*, mr.member_state, mr.rel_id
@@ -250,3 +250,9 @@
   and m.group_id = amo.object_id
   and m.rel_id = mr.rel_id
   and m.container_id = m.group_id;
+
+create view cc_users_of_package_id
+as
+SELECT u.*, au.package_id
+FROM application_users au, cc_users u
+WHERE (au.user_id = u.user_id);



....



Index: upgrade-1.1d-1.2.sql
===================================================================
RCS file: /cvsweb/infogettable/packages/edit-this-page/sql/postgresql/upgrade/upgrade-1.1d-1.2.sql,v
retrieving revision 1.1.1.1
retrieving revision 1.2
diff -u -r1.1.1.1 -r1.2
--- upgrade-1.1d-1.2.sql        19 Apr 2004 14:30:46 -0000      1.1.1.1
+++ upgrade-1.1d-1.2.sql        19 Apr 2004 16:07:09 -0000      1.2
@@ -185,14 +185,4 @@
 end;
 ' language 'plpgsql';

-select content_type__create_type (
-        'etp_page_revision',        -- content_type
-       'content_revision',         -- supertype
-       'ETP managed page',       -- pretty_name
-       'ETP managed pages',      -- pretty_plural
-       'etp_page_revisions',            -- table_name
-       'etp_page_revision_id',              -- id_column
-       'content_revision__revision_name'  -- name_method
-);
-
 \i ../edit-this-page-sc-create.sql
\ No newline at end of file


PS. The upgrade went without much trouble. Amazingly, considering the 4.5b starting point. Kudos and many thanks to the good folks that carved out the proper upgrade path to 5.0 (Andrei Popov, Joel Aufrecht, and Dave Bauer to name but a few...) Thanks for the *great* documentation!
Collapse
Posted by Joel Aufrecht on
Please go ahead and commit to 5.1.  It should get merged to HEAD within the next few weeks, after we release 5.1 final (next week, at the earliest).
Collapse
Posted by Jeff Davis on
It's not too late to submit patches. The merge was simply to sync up bug fixes/features from the 5.1 branch to HEAD (so people working on HEAD don't see bugs that are already fixed). The work on oacs-5-1 is ongoing and I imagine we will have a 5.1.0 release in the fairly near term (and subsequent 5.1.1,2,... maintenance releases).

Please ticket a bug and attach the patch to it (or if you have commit on acs-kernel you could go ahead and commit those).

Collapse
Posted by Ola Hansson on
I committed these changes to oacs-5-1 ...
Collapse
Posted by Gabriel Burca on
Will the patches that are currently open be applied to 5.1 before it is released?

In general, if one wants to submit a patch for an upcoming release, is there a cutoff date by which the patch needs to be in the bugtracker to be considered for inclusion?

Collapse
Posted by Jeff Davis on
I think we should make every effort to apply the outstanding patches, and we have not been doing a great job of doing that.

There is not really a cutoff, although the biggest issue is simply resources, there are not enough people taking time to test and apply patches.

If you are submitting a lot of patches you should consider asking for commit which generally we are pretty happy granting.