dmd 1.074 and 2.059 release

[Jonathan M Davis]

On Saturday, April 14, 2012 17:01:56 Adam D. Ruppe wrote:
> On Saturday, 14 April 2012 at 02:52:59 UTC, Walter Bright wrote:
> > Sorry about that, but the beta has been out for a week and a
> > half, and we fixed every reported regression since 2.058.
> 
> I'm sure I'm not the only one who doesn't use the betas though!

Sure, regressions can happen, but they only get found if there are already 
tests which catch them (in which case, they generally don't get checked in) or 
if people compile their code with the latest head or with the betas and then 
report them. The primary purpose of the betas is to get the newest compiler 
and library implementations tested on 3rd party projects in order to catch 
regressions. If not enough people pay attention to the betas, then they won't 
be tested on much code, and we won't catch the regressions.

The only way that a regression that your code is going to be hit by is going 
to be caught in beta if you don't participate is if someone else who _does_ 
participate happens to hit it in their code. If you want to ensure that all of 
your code always compiles with each new release, then participate in the 
betas. We try and catch all of the regressions, but even with solid test 
suites and what beta participation we do get, we can't always catch 
everything.

- Jonathan M Davis