• ryuukk_ (7/7) May 14 I just saw this yesterday, and i haven't see anybody talk about
• ryuukk_ (12/12) May 14 It'll be interesting to know what the experience was for the
• evilrat (5/15) May 14 Yep, i think it was there long enough already so rvalue ref
• Guillaume Piolat (3/4) May 14 Complaints from C and C++ programmers could mention that at least
• ryuukk_ (11/11) May 14 LDC yet again is proving how essential it is for D, now that more
• Hipreme (17/29) May 14 - On my work with redub, I've used dmd as a default. If it is not

ryuukk_ <ryuukk.dev gmail.com> writes:

I just saw this yesterday, and i haven't see anybody talk about 
it here, so i decided to share this great news for D here

Sokol is a popular gamedev library https://github.com/floooh/sokol

It now officially supports D, thanks to the work led by 
https://github.com/kassane

https://github.com/floooh/sokol/blob/master/CHANGELOG.md#13-may-2024

https://twitter.com/FlohOfWoe/status/1790043272385114421

ryuukk_ <ryuukk.dev gmail.com> writes:

It'll be interesting to know what the experience was for the 
maintainer to play around with D (for the first time?)

 From what i could gather, problems encountered:

- rvalue ref params (wich led to someone telling him to use 
-preview=all wich led to other issues)

- attributes soup

- float nan default

- no warning on use of uninitialized floats


I think it is time to decide what to do with all these -previews, 
rvalue ref params should be a feature already imo

For float nan default, i personally think everything should be 
consistent therefore 0 initialized

evilrat <evilrat666 gmail.com> writes:

On Tuesday, 14 May 2024 at 12:36:27 UTC, ryuukk_ wrote:
 It'll be interesting to know what the experience was for the 
 maintainer to play around with D (for the first time?)

 From what i could gather, problems encountered:

 - rvalue ref params (wich led to someone telling him to use 
 -preview=all wich led to other issues)

 - attributes soup

 - float nan default

 - no warning on use of uninitialized floats


 I think it is time to decide what to do with all these 
 -previews, rvalue ref params should be a feature already imo

Yep, i think it was there long enough already so rvalue ref 
parameters should already be in.
Also what's with this `in` const parameters transition, it just 
like umm hanging in there for years?

Guillaume Piolat <first.name gmail.com> writes:

On Tuesday, 14 May 2024 at 12:36:27 UTC, ryuukk_ wrote:
 - float nan default

Complaints from C and C++ programmers could mention that at least 
stack variables gets initialized to something.

ryuukk_ <ryuukk.dev gmail.com> writes:

LDC yet again is proving how essential it is for D, now that more 
OS are embracing ARM, having LDC available and maintained to 
support latest version of LLVM and all kind of platforms is a 
blessing

The maintainer is using macOS, and his library targets many 
platforms including the web with WASM, both supported by LDC out 
of the box

DMD for fast iteration time (on x86, hopefully ARM in the future)

And LDC for everything else

It's a deadly combo, thanks to everyone working hard on 
maintaining both projects

Hipreme <msnmancini hotmail.com> writes:

On Tuesday, 14 May 2024 at 12:46:13 UTC, ryuukk_ wrote:
 LDC yet again is proving how essential it is for D, now that 
 more OS are embracing ARM, having LDC available and maintained 
 to support latest version of LLVM and all kind of platforms is 
 a blessing

 The maintainer is using macOS, and his library targets many 
 platforms including the web with WASM, both supported by LDC 
 out of the box

 DMD for fast iteration time (on x86, hopefully ARM in the 
 future)

 And LDC for everything else

 It's a deadly combo, thanks to everyone working hard on 
 maintaining both projects

- On my work with redub, I've used dmd as a default. If it is not 
being able to find it, it will give an warning and try LDC 
instead.
- No need to change float default from nan to 0, excluding my own 
opinion, Walter concretely believes that nan is better than 0. I 
would like to at least get warnings on uninitialized floats, so 
it becomes 50-50 and no one will ever initialize float to nan 
again.
- Could we get official LDC full D support to WASM? I haven't 
been having much time to work with D. When I do have time, the 
last thing I want to do is updating my custom runtime again. 
Beyond that, people always get the wrong impression from the WASM 
status quo. I know we won't have a garbage collection, but 90% of 
the work will be done. Just send a warning for anyone using D 
wasm runtime that support is incomplete. D always done that 
nevertheless.