China Sahana/VM Deployment–Day 3


The Monday night conference call consisted of the same cast of players: the project leader from IBM-China and his assistant, three members of the Sahana core team, and the Trinity VM team.

One problem has been finding a suitable way to work collaboratively on a common code base–i.e., CVS control. The Sahana team has been committing bug fixes and customizations to rel_0_6 on their CVS repository. But the China team is making direct changes to the codebase and doesn’t have (easy?) access to Sahana’s CVS. For now, the solution has been the China team develops “fix lists” and transmits them to the Sahana team, which commits them to the repository–not the most efficient operation.

A more efficient solution would be to create a separate branch in the repository for the China deployment. Unfortunately, only Chamindra has the ability to do that and he is still enroute to Colombo.

The China team has been sending bug reports and feature requests via .doc files. Today they sent files on the missing person registry, organization registry, and the inventory module. The Colombo team will work on those and send back fixes and explanations.

The Trinity VM team has been working on translation fixes–some sub-menus and error messages were missed in the initial set. These were packaged up and sent directly to the China team.

The China team introduced us to a IBM-NYC volunteer, a native of Chengdu and an expert in PHP+MySQL. He will be helping us modify and test the VM module.

The IBM-China team leader left the call at 11:31 PM EDT, urging us to try to find more volunteers to help out and thanking everybody for their efforts. The IBM-China assistant leader asked for documentation–specifically ER diagrams. Apparently, there are no ER diagrams for the majority of the Sahana code base, although the Sahana development site has useful documentation. Happily, there is one for the VM module. This type of exercise underscores the importance of good documentation. We will improve our VM documentation over the summer!

The Trinity team remained on the call until 1:30 AM, working in a parallel Skype voice session on some of the translation problems that were discovered and other problems–all rather small in scope. (Antonio fell asleep at (on?) his keyboard–we could occasionally here him grunting and moaning.) One problem involved the skills tree on the Volunteer page–the Javascript code that Giovanni developed to display a hierarchy of skills and occupations. Giovanni fixed that problem.

Another involved the fact that despite having tagged English phrases for internationalization in the code, they were still being missed by Sahana. It was determined that this was due to the way the VM templates generate temporary inc files in the cache. But the Sahana xgettext scan does not search the cache. Therefore these instances had to be processed by hand.

To help mitigate the ongoing mis-communication problems Trishan and Giovanni added a version number to the VM module.  This way when the China team reports missing translations, we can tell whether they are running the latest update.

One of the tasks added to the VM team is to incorporate paging into the VM module–this is currently not implemented. Having added that to our “to do” list, we turned around 2:00 AM.

  1. No comments yet.
(will not be published)