• last updated 16 hours ago
Constraints
Constraints: committers
 
Constraints: files
Constraints: dates
set bt folder context to the bt package id instead of defaulting to -100 so that perms on package trickle to bugs

make sure notification fields are localized

upgrade 4.1 to 4.1.1

new files for handling distributions

  1. … 18 more files in changeset.
typo in bug_tacker, should be bug_tracker

fix typos

  1. … 1 more file in changeset.
Adding tables for autosubmition functionality.

adding scheduled procs in order to close bugs automatically ( Controlled by a parameter ).

    • -0
    • +15
    ./bug-tracker-scheduled-init.tcl
    • -0
    • +24
    ./bug-tracker-scheduled-procs-postgresql.xql
    • -0
    • +74
    ./bug-tracker-scheduled-procs.tcl
    • -0
    • +16
    ./bug-tracker-scheduled-procs.xql
  1. … 8 more files in changeset.
commit to HEAD of merge of changes on oacs-5-1 branch between tags jcd-merge-5-1-20050111 and jcd-merge-5-1-20050224

  1. … 314 more files in changeset.
Fixing broken query project_new.instance_info

added assignee and fix for version to the bug add script; added appropriate upgrade scripts for postgres and oracle; fixed an oracle installation bug

  1. … 6 more files in changeset.
Added a whole shitpile of caching to speed things up

removed internationalization

We can further simplify the display query when using the paginator because

we don't need to include the filter clauses that are used to restrict the

rows returned. The paginator returns a list of ids that have already been

so restricted so the filter clauses are redundant in the display query.

Switched views for filtered queries that return a bug assigned to a

particular user. The workflow views invoked in these queries simply

do not scale well. This needs to be addressed fundamentally in workwflow

at some point but hacking bug-tracker seems easiest for now.

It looks like things have sped up to the point where it's worth trying

to pagination 10 pages of 25 entries a page at a time for now.

Finally figured out the obnoxious main query and rewrote the pagination

query to match my new understanding ...

Added per-instance list builder caching ...

  1. … 1 more file in changeset.
Sigh Oracle has really weird rules about "*" and duplicate column ids

that cause it to fail if you wrap "select *" as "select p.* from (select *) p"

Yet another fix to pagination ... getting there...

Ahem. Pagination might work now ... it didn't really before.

First pagination code ... more coming (caching in particular)

commit to HEAD of merge of changes on oacs-5-1 branch between tags jcd-merge-5-1-20040724 and jcd-merge-5-1-20050111

  1. … 2268 more files in changeset.
Added code to properly upgrade for workflow, by adding enabled actions to

each bug (which results in the "edit", "resolve" etc buttons appearing on

the bug page).

  1. … 1 more file in changeset.
oops last commit accidently screwed up the xql file

Tch tch someone put some SQL code inline rather than in an XQL file 'cause

they thought it was DB independent and it was really PG dependent. I made

it DB independent and stuck it in the XQL file where it belongs anyway.

Error in my test case, you have to flush the components cache if you add

one...

New instantiation of Bug-Tracker was broken due to old usage of the CR. Fixed in an ugly, hasty hack for postgresql, as I assume Don will work on it once he comes around to it

  1. … 1 more file in changeset.
Finished up some more basic testing

There was a missing package_id parameter in part of the Tcl API. HOWEVER,

don't believe the bug-tracker Tcl API that misleads you into thinking that

you can explicitly pass package_id as a parameter to various procs. The

vile bug_tracker::conn proc guarantees this does not work. I rewrote my

test case to save/set/restore ad_conn package_id ...

Added test subdirectory and first cut at the first test (which doesn't work

at the moment, but never fear). Decided to write some simple tests for this

package while working out how to write populate scripts to test scalability

(fix scalability, we already know it sucks on PG!) ...