[dmd-beta] Time for a new beta?

[Jonathan M Davis]

On Friday, September 02, 2011 09:00:43 Don Clugston wrote:
> On 1 September 2011 08:06, Walter Bright <walter at digitalmars.com> wrote:
> > I'm game.
> > 
> > On 8/31/2011 9:37 PM, Brad Roberts wrote:
> >> Well, there seemed to be some limited consensus 2 weeks ago that it
> >> might be time for a beta, then the period of
> >> extended brokenness.  Now that we're back in a happy state, shall we?
> >> 
> >> Anyone have a 'really must be fixed before the next release' list?
> >> 
> >> As always, mine consists of the open regression bugs, but since I'm
> >> not
> >> personally investing in fixing any of them, it's
> >> poor form to gripe too much about them.
> 
> At least, I don't think we should do a release when there are open
> regression bugs that we have valid pull requests for.
> If the patches are rejected, we shouldn't block the release, but they
> should at least be reviewed.
> There are four in that category:
> 
> IMHO the most important one is bug 3797 (Regression 2.038). It
> generates bad code and has been reported 8 times. And it's been
> waiting a very long time for review.
> https://github.com/D-Programming-Language/dmd/pull/96
> 
> This one, which is a collaboration involving yebblies and me (mostly
> yebblies), fixes regression 5373 (Regression 2.051) and an ancient
> bug:
> https://github.com/D-Programming-Language/dmd/pull/362
> 
> This one fixes regression 4953 (Regression 2.031)
> https://github.com/D-Programming-Language/dmd/pull/269
> 
> This one fixes regression 6398 (Regression 2.054)
> https://github.com/D-Programming-Language/dmd/pull/279
> 
> The only other concern I have is, is the "disable default
> constructor"/"nonNull" implementation stable enough for release?

It's not a regression, but if

https://github.com/D-Programming-Language/dmd/pull/345

is acceptable, then it would definitely improve Phobos' ability to deal with 
deprecated symbols and symbols which are scheduled for deprecation, which 
should fix some of the complaints from the last release (about messages 
informing users of impending deprecation). So, it would help make for a much 
more pleasant release, but it's not exactly necessary - particularly if we're 
in a hurry to release. It would be a definite improvement though.

- Jonathan M Davis