rdmd and D equivalent for PYTHONPATH?

[Chris Piker]

On Wednesday, 17 March 2021 at 09:34:21 UTC, Mike Parker wrote:
> On Wednesday, 17 March 2021 at 07:13:31 UTC, Chris Piker wrote:
>> Very handy example.  Unfortunately it means that paths are 
>> embedded
>> in scripts, which is usually a bad idea.
>
> The ability to use D source modules “script style” is something 
> that has grown organically over time largely as a convenience. 
> I doubt anyone is using it heavily enough to have tested the 
> boundaries, In your exploration of those boundaries, please 
> take note of what you discover so they can potentially be 
> expanded where possible.

Sure will, thanks for the invite to contribute in a specific way.

D looks to be a good replacement for split Python/C development 
though I don't want to drag all my python baggage in here.  I'm 
trying to understand the D way of doing things before suggesting 
changes.

Interpreted languages like Python MATLAB, IDL are the norm in my 
field.  So anything that makes D easier to use for 
"quick-and-dirty" data analysis tasks would make it more 
palatable to the casual programmers I interact with.

I general dub seems fantastic!  I was stunned yesterday by a 
three-line vibe.d test script I ran that produced a compiled 
running web-server.

So, if I could do the equivalent of:

   dub add-path

via an environment variable (not a permanent change under 
~/.dub), or have some environment variable that tells dub where 
to read a "system-level" local-packages.json file and merge it's 
paths in with any personal settings, that would likely handle our 
internal code sharing needs.