On 2011-01-06 07:28, Walter Bright wrote:
Nick Sabalausky wrote:
Automatically accepting all submissions immediately into the main line
with no review isn't a good thing either. In that article he's
complaining about MS, but MS is notorious for ignoring all non-MS
input, period. D's already light-years ahead of that. Since D's purely
volunteer effort, and with a lot of things to be done, sometimes
things *are* going to tale a while to get in. But there's just no way
around that without major risks to quality. And yea Walter could grant
main-line DMD commit access to others, but then we'd be left with a
situation where no single lead dev understands the whole program
inside and out - and when that happens to projects, that's inevitably
the point where it starts to go downhill.
That's pretty much what I'm afraid of, losing my grip on how the whole
thing works if there are multiple dmd committers.
That is very understandable.
Maybe we can have a look at the linux kernel development process:
http://ldn.linuxfoundation.org/book/how-participate-linux-community
As how I understands it, Linus Torvalds day to day work on the linux
kerenl mostly consist of merging changes made in developer branches into
the main branch.
On the bright (!) side, Brad Roberts has gotten the test suite in shape
so that anyone developing a patch can run it through the full test
suite, which is a prerequisite to getting it folded in.
Has this been announced (somewhere else than the DMD mailing list)?
Where can one get the test suite? It should be available and easy to
find and with instructions how to run it. Somewhere on the Digitalmars
site or/and perhaps released with the DMD source code?
In the last release, most of the patches in the changelog were done by
people other than myself, although yes, I vet and double check them all
before committing them.
--
/Jacob Carlborg
