• last updated 14 hours ago
Constraints
Constraints: committers
 
Constraints: files
Constraints: dates
1. The externals group was pretty-named "Externals" rather than "External".

2. Date arithmetic issue for PG 7.3, switched to PG interval type

I've changed "timestamp" to "timestamptz" and have begun fixing some

code that used to be legal but as of PG 7.3 is not. You can no longer

subtract an integer from a timestamp but must cast to date, "timestamp"

disappeared (must use "interval" now), "datetime" disappeared ("timestamptz"),

and suchwhat.

Not done but closer than it was before ...

I've changed "timestamp" to "timestamptz" and have begun fixing some

code that used to be legal but as of PG 7.3 is not. You can no longer

subtract an integer from a timestamp but must cast to date, "timestamp"

disappeared (must use "interval" now), "datetime" disappeared ("timestamptz"),

and suchwhat.

Not done but closer than it was before ...

  1. … 25 more files in changeset.
A handful of various fixes, mostly for PG 7.3. Unfortunately I copied

over the fixed directory tree from my test instance here by mistake so

everything looks changed.

    • -2
    • +0
    /openacs-4/packages/dotlrn/dotlrn.info
I've changed "timestamp" to "timestamptz" and have begun fixing some

code that used to be legal but as of PG 7.3 is not. You can no longer

subtract an integer from a timestamp but must cast to date, "timestamp"

disappeared (must use "interval" now), "datetime" disappeared ("timestamptz"),

and suchwhat.

Not done but closer than it was before ...

  1. … 39 more files in changeset.
Made changes so we can install under PG 7.3.

Worked on fixing cms-news-demo in order to give Greenpeace (and others) an

example CR/CMS app to play with. It's not finished but partway there. On

the way found a couple of CR bugs and fixed them.

Partial conversion from timestamp to timestamptz for PG 7.3 (need to keep

time zone timestamps so dumps from PG 7.2 installs will still work)

  1. … 4 more files in changeset.
Added 'groupnum_last_p' pseudocolumn to group tag, and made the 'groupnum' pseudocolumn work correctly with nested group tags.

When searching for an enclosing tag, start with the parent tag, not the current tag. Otherwise it won't work if you have to identical tags inside each other, for example a <group> tag inside a <group> tag.

Only ns_register_adptag if we haven't already registered it. This makes debugging template tags much easier, because you don't have to restart the server each time.

regenerated html to accomodate new sections and css/xsl

  1. … 57 more files in changeset.
Placeholders for tutorial and submissions sections; integrated upgrade, tutorial, submissions, and i18n; tweaked css and xsl

return a more meaningful error when a package is not mounted

fix for bug 274 where node(name) sometimes doesnt exist (it should always exist though)

Merged from HEAD: Fixed typo (extra parenthesis) in content_keyword__is_assigned

Fixed typo (extra parenthesis) in content_keyword__is_assigned

register_child_type would only register one type per parent since it

ignored relation tag when checking if the relation existed.

is_valid_type failed with more than one type since it wanted to count

max_n but did not sum.

needs an upgrade script.

missing query in oracle, remove dups in allowed children

Workaround for new and old nsxml modules.

add email and url as acs_datatypes

add more template::widget @see stuff

add some param docs

Initial commit

    • -0
    • +56
    /openacs-4/packages/ref-itu/sql/oracle/ref-itu-create.sql
    • -0
    • +36
    /openacs-4/packages/ref-itu/sql/oracle/ref-itu-drop.sql
Initial commit

    • -0
    • +36
    /openacs-4/packages/ref-itu/ref-itu.info
    • -0
    • +655
    /openacs-4/packages/ref-itu/sql/common/ref-itu-data.sql
removing package_key_select queries no longer used

making catalog messages cascade when packages are deleted

chaing catalog import proc so that import for uninstalled packages is not an upgrade, also if the version number in the catalog file is lower than that of the package it's not an upgrade

index creates moved to 4.6.2 upgrade to be consistent with the oracle side

Most of the RI indexes were duplicated between the 4.6-4.6.1 and the 4.6.1-4.6.2 upgrade scripts