Towards 0.12.0; release an alpha now?

[Benjamin Thaut]

Am 06.10.2013 21:42, schrieb David Nadlinger:
> Hi all,
>
> The biggest problem we are facing right now is arguably that there
> hasn't been a proper binary release of LDC for almost four months.
>
> The reason for not making any progress on that front has mainly been
> that there are a couple of serious issues that occur only relatively
> rarely, but are hard to fix, such as:
> https://github.com/ldc-developers/ldc/issues/407
>
> But while I am not really comfortable with doing a release while
> issues such as this are still unresolved, I think it is vital that we
> at least have binaries available for people who want to test/use LDC
> in its current state, featuring the 2.063 frontend and a couple of
> critical fixes for MinGW.
>
> As #407 is the only potentially huge regression I'm aware of, (the
> -march/-mcpu issue has been sorted out) I propose to release an alpha
> version based off current Git master. If the remaining issues turn out
> to be not so much of a problem, we can release 0.12.0 (at least one
> beta as usual) shortly after that.
>
> This would also make sense as 2.064 will probably come out fairly
> soon, and people will expect us to support shared libraries (which I
> know how to implement, but we really want the 2.064 druntime changes
> to avoid duplicating work).
>
> What do you think? I'll start work on getting alpha packages out, but
> wait for other opinions before I officially put them up.
>
> Cheers,
> David
>

A binary release would be great. I would also greatly appreciate a new 
windows x64 build even in a unstable form. Issue 407 doesn't look that 
critical to me, because if you really need associative arrays you wrote 
your own by now considering how many issues there are with the druntime 
implementation of AAs.

Has there been any further progress on windows x64 exception handling?

Kind Regards
Benjamin Thaut