[Twisted-Python] Trial: The Volunteering
Hello Everyone. I've been off-list for a while, but I am back. As the original author of Trial, I still feel partly responsible for it. So, I would like to become the official maintainer of Trial again. I am _not_ planning on re-writing Trial. Rather, I'm going to restrict myself to fixing bugs, doing incremental refactorings to clean-up the cruft and rejecting ideas for new features. I've asked Jonathan Simms (slyphon) whether he would mind. He is more than happy for me to assume the mantle of Trial maintainer. Let me know if there are any objections. Otherwise, start assigning trial bugs to me by default. cheers, jml
I've been off-list for a while, but I am back. As the original author of Trial, I still feel partly responsible for it. So, I would like to become the official maintainer of Trial again.
Would you like to describe what Trial is? I have no clue and don't know where to find those clues either. Tim
On Mon, Aug 22, 2005, Tim Terlegård wrote:
Would you like to describe what Trial is? I have no clue and don't know where to find those clues either.
It's the Twisted unit testing framework. API documentation is available at [1] What other documentation there is seems to be in the Testing Standard HOWTO at [2] -Mary [1] http://twistedmatrix.com/documents/current/api/twisted.trial.html [2] http://twistedmatrix.com/projects/core/documentation/howto/policy/test-stand...
On Sun, 21 Aug 2005 19:47:58 -0700, Jonathan Lange <jonathan.lange@gmail.com> wrote:
I would like to become the official maintainer of Trial again.
Excellent. Please allow me to officially present you with the ceremonial lead albatross ;-). I believe the main direction that Trial has to go in, to wit, removing all reactor-spinning calls, requires first that all Twisted unit tests which call wait() or deferredResult() or the like be converted to instead return Deferreds. This is probably a pretty big effort so feel free to harrass individual project maintainers; but please consider coordinating that effort.
I believe the main direction that Trial has to go in, to wit, removing all reactor-spinning calls, requires first that all Twisted unit tests which call wait() or deferredResult() or the like be converted to instead return Deferreds.
There's a couple of files in twisted.trial, doctest.py and tdoctest.py . The first looks like a patched version of the stdlib one. What about present and/or future support for writing tests in doctest format? -- Nicola Larosa - nico@tekNico.net PHP is such a load of crap, right down to the standard library, that it creates a culture where it's acceptable to write horrible code. [...] Maybe with PHP 5 they are trying to clean up the neighborhood, but that doesn't change the fact when you program in PHP you are programming in a dump. -- Ian Bicking, July 2005
On Mon, 22 Aug 2005 18:59:48 +0200, Nicola Larosa <nico@teknico.net> wrote:
I believe the main direction that Trial has to go in, to wit, removing all reactor-spinning calls, requires first that all Twisted unit tests which call wait() or deferredResult() or the like be converted to instead return Deferreds.
There's a couple of files in twisted.trial, doctest.py and tdoctest.py . The first looks like a patched version of the stdlib one. What about present and/or future support for writing tests in doctest format?
It's a mess. I think it should be removed. A future version of trial might have support added in a more maintainable fashion. Jp
On Sun, 21 Aug 2005 19:47:58 -0700, Jonathan Lange <jonathan.lange@gmail.com> wrote:
Hello Everyone.
I've been off-list for a while, but I am back. As the original author of Trial, I still feel partly responsible for it. So, I would like to become the official maintainer of Trial again.
I am _not_ planning on re-writing Trial. Rather, I'm going to restrict myself to fixing bugs, doing incremental refactorings to clean-up the cruft and rejecting ideas for new features.
I've asked Jonathan Simms (slyphon) whether he would mind. He is more than happy for me to assume the mantle of Trial maintainer.
Let me know if there are any objections. Otherwise, start assigning trial bugs to me by default.
I'd like to talk about any big changes you'd like to make before you make them. Trial has recently almost become just barely usable, and I'd like to keep it from slipping back at all. Also, there is some major refactoring/rewriting that is definitely required. If you don't want to tackle these, that's fine, but someone's going to have to. Jp
On 22/08/05, Jp Calderone <exarkun@divmod.com> wrote:
On Sun, 21 Aug 2005 19:47:58 -0700, Jonathan Lange <jonathan.lange@gmail.com> wrote:
Hello Everyone.
I am _not_ planning on re-writing Trial. Rather, I'm going to restrict myself to fixing bugs, doing incremental refactorings to clean-up the cruft and rejecting ideas for new features.
Let me know if there are any objections. Otherwise, start assigning trial bugs to me by default.
I'd like to talk about any big changes you'd like to make before you make them. Trial has recently almost become just barely usable, and I'd like to keep it from slipping back at all.
Also, there is some major refactoring/rewriting that is definitely required. If you don't want to tackle these, that's fine, but someone's going to have to.
I do want to tackle them, but carefully and slowly. Like you said, Trial has reached a local optimum of utility. I'm more than happy to discuss big changes. I have a largish patch that Robert Collins and I worked on at the .au sprint. The changes separate dry run from the main tree, and are a step towards separating the code that runs tests from the code that finds tests. Tell me where you want it mailed. jml
On Tue, Aug 23, 2005, Jonathan Lange wrote:
I'm more than happy to discuss big changes. I have a largish patch that Robert Collins and I worked on at the .au sprint. The changes separate dry run from the main tree, and are a step towards separating the code that runs tests from the code that finds tests. Tell me where you want it mailed.
Not that it's likely that I personally will be reviewing it in any shape or form, but why not move it onto a subversion branch? I gather people are slowly moving in that direction, and it does at least potentially allow people to call "halt this madness!" when they see your commits rather than when they see a big change land in trunk. (It might also be useful eventually to be able to specify that the buildbot tests run over at least some branches?) -Mary
On Tue, Aug 23, 2005, Mary Gardiner wrote:
Not that it's likely that I personally will be reviewing it in any shape or form, but why not move it onto a subversion branch?
This does have the disadvantage that you'd need to find a good way to commit Robert's changes to the branch if he does any more work on it. Andrew had at least some kind of way to make Subversion and bzr sandboxes co-exist, it might have been more or less as simple as having .bzr and .svn directories in the same sandbox. -Mary
participants (6)
-
glyph＠divmod.com -
Jonathan Lange -
Jp Calderone -
Mary Gardiner -
Nicola Larosa -
Tim Terlegård