On 2 June 2015 at 04:14, Andrei Alexandrescu via Digitalmars-d <[email protected]> wrote: > On 6/1/15 11:11 AM, Andrei Alexandrescu wrote: > [...] > > Meant to also add > > 4. We need to marshal our efforts behind 2.068, and clarify the big ticket > items accomplished. I'm thinking rangification of Phobos - GC no longer > needed for most primitives, and documented where needed. As much as I want > it, ddmd seems to not be happening for 2.068 because of, simply put, > insufficient resources. > > What other large topics for 2.068?
Based on other conversations, perhaps a good concept to get behind would be: LDC+GDC are first-class toolchains, and DMD devs will work together with the other compilers to release in parallel, making sure distributions for all compilers are available at the same time from the same place. I think it's been a big problem forever now that nobody can really have any confidence that the language version is matching between available compilers. You've said elsewhere, this is a major issue inhibiting ddmd? Perhaps a focus for 2.068 is to organise and address this general problem by developing better systems of coordination and release? If this can be achieved, perhaps we could promote LDC as the default go-to compiler, such that newcomers doing experiments and benchmarks will experience realistic results against competition? Constantly having to correct benchmarkers to use LDC and offer the proper flags is a bit embarrassing. I wonder how many times this has happened and it's not posted on the internet such that we are able to correct them?
