Ryan Lane wrote:
I don't think we should encourage people to run trunk in production. We should encourage people to run release candidates in production, and possibly betas for those that know the software *really* well. We should likely encourage people to run trunk on their live testing environments, though.
Occasionally security issues pop up in trunk that get caught in code review. People who run trunk are much more likely to have security problems, so on a production site, it's a problem. Similarly, it's possible that commits may come in that can cause data loss, which will later get caught in code review.
Our trunk usually stays stable enough that developers can have a usable working environment, but I don't think it's stable enough to run on any site where you care about security or your data.
- Ryan Lane
It is the aim of the MediaWiki community to have an always-working trunk. It /should/ be possible to run a small wiki from trunk without much more than verifying before deployment that the tests passes and a few look out for fixmes. We should add tests for any -serious- issue fixed. They should also be able to cover the whole site (eg. add a test for Special:Cotnributions stub threshold [1]). Obviously, running trunk needs requires to have watch for possible errors and has some expectation of possible rendering problems (eg. the templates and tables stuff) or some corner case leading to a blank page.
I was going to say something like "it is very very strange to have commits leading to data loss" but we too had amazing failures, with the tests added backdoors or overwriting the wiki. :(
1- This should be possible with Selenium. How does that effort go?