December 02

Greetings,

I'm still waiting for Razvan to finish reviewing last week's PR before doing anything else in that area.

In the meantime, I have started hacking around trying to enable full semantic analysis in D-Scanner. There is one impediment I've discovered - before running full semantic, DMD requires a list of versions to compile.

Since users can define custom versions, and some versions can be mutually exclusive, I was thinking with Razvan that we should let the user specify the version list(s).

I'm curious to see if there are any thoughts about this.

December 02
On 02/12/2024 2:28 PM, Vladiwostok wrote:
> Greetings,
> 
> I'm still waiting for Razvan to finish reviewing last week's PR before doing anything else in that area.
> 
> In the meantime, I have started hacking around trying to enable full semantic analysis in D-Scanner. There is one impediment I've discovered - before running full semantic, DMD requires a list of versions to compile.
> 
> Since users can define custom versions, and some versions can be mutually exclusive, I was thinking with Razvan that we should let the user specify the version list(s).
> 
> I'm curious to see if there are any thoughts about this.

Because any flag can set versions, especially targets, your best bet is to accept all CLI flags that (l/g)dmd does.

That way you can throw your compile command directly at it, and not worry about errors.