Forum OpenACS Q&A: Bug Reports & Patches

Collapse
1: Bug Reports & Patches
Posted by Robin Felix on 05/21/01 05:43 PM
I've submitted several bug reports with proposed fixes, and I want to ensure that I'm doing it in a way that's easiest for the maintainers. Would you prefer submitted patches? Do patches go live immediately, or only after evaluation by a module owner?

I imagine there's a confidence level or level of trust that can be assigned to code appearing on the site here that is driven by your processes and others' understanding of the proper place to post fixes. For example, from highest confidence to lowest:

  • Releases x.x.x (production)
  • Releases x.x.xb (beta)
  • Releases x.x.xa (alpha)
  • CVS tree commitments
  • Patches (assuming they're evaluated before publishing)
  • Fixes accompanying bug reports in the SDM
  • Code snippets in the forums
A description of your evaluation and publishing process would be helpful, and it would help me assign a confidence level to the various forms of updates and patches that appear from others.
Collapse
2: Response to Bug Reports & Patches (response to 1)
Posted by Don Baccus on 05/21/01 06:56 PM
We've been very informal in the past, as we've been a very small and intimate project in the past.  We're starting to get more formal.  Submitting patches via the SDM is definitely the right way to go.  For  the OpenACS 3x series we'll just be doing maintenance releases, so can probably live with a beta then production series (i.e. don't really need alpha).

We're careful about CVS commits so in the past, at least, the latest CVS tree has typically been more reliable than the latest production release.

I expect us to be a bit more formal about OpenACS 4 rollouts when we get there.  For one thing, OpenACS 4 is much more complex than OpenACS 3.2, and testing is intrinsically more complex due to our support of (currently) two rather than just one RDBMS.  The project's also larger and growing in terms of people doing work, and the larger a project gets the less likely that an informal approach will work well.

Collapse
3: Response to Bug Reports & Patches (response to 1)
Posted by Robin Felix on 05/21/01 07:20 PM
Tks, Don. Two questions:
  1. Are patches verified before making them available publically?
  2. Would it be helpful for you if I turned my previous bug submissions into patches?