dmd 1.045 / 2.030 release

[Don]

Steven Schveighoffer wrote:
> On Tue, 12 May 2009 12:40:11 -0400, Walter Bright 
> <newshound1 at digitalmars.com> wrote:
> 
>> Tomas Lindquist Olsen wrote:
>>> Is there a reason for the missing announcement ?
>>
>> Yes, I sent it to people who'd asked for a prerelease so they could 
>> check their builds against it.
> 
> This should be a private release then.
> 
> If 1.045 is to be a "pre-release", then it should not be on the web 
> site.  If 1.045 has some critical bug that you decide needs to be fixed, 
> then do you release 1.045 again with the fix, or 1.046?  The only sane 
> choice is to not rerelease a version that people already have, that 
> would make bug reporting a constant struggle (oh, you have 1.045?  Is 
> that the broken 1.045 or the good 1.045?) not only for DMD but for other 
> projects where compiler version is important in the bug reports.  A 
> prerelease version of DMD should be marked as such (i.e. 1.045rc as 
> people have suggested).

Agreed. This is why recently DMD2.028 was released, rather than a 
re-release of DMD2.027 with bug 2812 fixed.

> 
> I'm not trying to cause trouble, I just don't want to see something like 
> this (2 releases with the same identifier).  I know in the past I've 
> advocated for "prerelease" versions, but I meant private pre-release.  

Private prereleases are happening. This was actually DMD2.030 RC3.

> The response was always, "how do you know who should get the 
> prerelease?"  It's a fair point.  The counter argument is, libraries 
> like Tango have to blanketly disregard all bugs reported with a broken 
> compiler, which sometimes isn't included on the bug report.
> 
> -Steve