Hi,

in my attempt to fix #265 (http://github.com/ldc-developers/ldc/pull/275), I transitively had to fix other things related to static initialization of structs and arrays. Unfortunately, I hit a wall today when I thought I just had it. Consider:

struct S { int array[4]; }

void main()
{
    S local = S(67); // local.array = [67, 67, 67, 67]
    static S static_ = S(67); // static_.array = [67, 0, 0, 0]
    assert(local == static_);
}

This passes in my current ldc commit, but fails for the current releases of dmd and ldc2, with values as commented. I could have lived with a bug and filed a bug report http://d.puremagic.com/issues/show_bug.cgi?id=9425 , but was baffled again when the ldc build test suite failed at http://github.com/D-Programming-Language/dmd/blob/e7eff4deef0fb5ec9a89cee4716785832fc41fb9/test/runnable/test42.d#L3223. How should I proceed? Just try to unfix it as expected in that test case, wait for a fix or try another issue instead?

On Wednesday, 30 January 2013 at 21:21:37 UTC, Sebastian Graf wrote:
> http://d.puremagic.com/issues/show_bug.cgi?id=9425
> [...]
> How should I proceed? Just try to unfix it as expected in that test case, wait for a fix or try another issue instead?

You might want to send a message to dmd-internals, asking for clarification on the intended behavior and noting that the issue blocks LDC development.

This way, you'll hopefully get an answer quicker than by just hoping that the right people stumble over the issue soon.

David