One of the key signs of a mature product is that regressions are treated with top priority.  With that in mind, I'd like to see us acting similarly.  I know some of you have been doing so already for the parts you've been working in, and thanks for that.

Here's the list of open dmd regression bugs filed since the release of 1.070/2.055:

  http://d.puremagic.com/issues/show_bug.cgi?id=6656
  static alias this broken in 2.055

  http://d.puremagic.com/issues/show_bug.cgi?id=6665
  Regression(2.055) ICE(cg87.c): static double inside closure

  http://d.puremagic.com/issues/show_bug.cgi?id=6669
  Compiler seg fault when using square brackets in inline assembly

  http://d.puremagic.com/issues/show_bug.cgi?id=6670
  Compiler seg fault using std.concurrency.atomicOp

  http://d.puremagic.com/issues/show_bug.cgi?id=6675
  Regression(2.054) ICE(glue.c) template parameter deduction with errors gagged

  http://d.puremagic.com/issues/show_bug.cgi?id=6681
  bogus duplicate union initialization or overlapping initialization errors

None of them have pull requests yet.  It's also entirely possible that some of these aren't actually regressions in the 2.055 release.  Either way, I'll start looking at 6669.  Anyone want to volunteer to own any of the others?

If you've already started working on any of them, please shout.  Or, even better, change the ownership in bugzilla to assigned to you.  That's probably worth doing in general to avoid duplicate efforts.

Thanks,
Brad

On 18 September 2011 02:14, Brad Roberts <braddr at puremagic.com> wrote:
> One of the key signs of a mature product is that regressions are treated with top priority. ?With that in mind, I'd like to see us acting similarly. ?I know some of you have been doing so already for the parts you've been working in, and thanks for that.
>
> Here's the list of open dmd regression bugs filed since the release of 1.070/2.055:

> ?http://d.puremagic.com/issues/show_bug.cgi?id=6675
> ?Regression(2.054) ICE(glue.c) template parameter deduction with errors gagged

I've got a fix for this one already.