don't ever call me nicholas

[Nick Sabalausky (Abscissa) via Digitalmars-d-announce]

don't ever call me nicholas

Re: Introducing alid

[Ali Çehreli via Digitalmars-d-announce]

On 9/14/22 12:08, Ali Çehreli wrote:

>- Subpackage design (thinking about it...)

Ok, I think I fixed that one as well.

I think my main problem was trying to import 'alid':

  import alid;  // WRONG - Could not make it work
  import alid.alid; // Worked with package.d file

I think subpackages work as well:

  import alid.cached;
  // etc.

(Yes, I used many "I think"s. :) )

Ali

Re: Introducing alid

[Ali Çehreli via Digitalmars-d-announce]

On 9/12/22 02:15, Ali Çehreli wrote:
> I am happy to publish on code.dlang.org for the first time:
>
>https://code.dlang.org/packages/alid

I've fixed 2 of the 3 issues reported here:

  - Removed gratuitous function attributes

  - Fixed Salih's .cycle use case

  - Subpackage design (thinking about it...)

Let's continue with the issues at a more appropriate place:

  https://github.com/acehreli/alid/issues

Thanks! :)

Ali

Re: Introducing alid

[rikki cattermole via Digitalmars-d-announce]

On 14/09/2022 8:44 PM, Ali Çehreli wrote:

On 9/12/22 09:34, rikki cattermole wrote:

 > dub.json
 > errornogc/alid/errornogc.d
 > circularblocks/alid/circularblocks.d

Considering I may want to let the users import the entire package as 
well with


   import alid;

how can I achieve my goal of subpackages? Telling me to forget about 
subpackages altogether :) is an option because I've already spent hours 
trying to find my way through different directory structures, random 
dub.json edits, experimenting with ways of stopping dub from fetching 
and using the bad version of the repo repeatedly, and many other random 
things...


In your root package you can still have the package.d file.

You would use the version added by Dub to detect if you should public 
import the other modules.


> DUB provides version identifier of dependencies for conditional 
compilation with version conditions. Have_ version 
identifier can be used for conditional compilation.


https://dub.pm/advanced_usage

Re: Introducing alid

[Adam D Ruppe via Digitalmars-d-announce]

On Wednesday, 14 September 2022 at 08:44:48 UTC, Ali Çehreli 
wrote:
Considering I may want to let the users import the entire 
package as well with


dub is poorly designed and doesn't cooperate well with D 
features. There's no good solution - hence the hundreds of lines 
long dub.json trying to hack it in. Like you could have the 
top-level package depend on all of its own subpackages and then 
if you depend on it, you can import alid. Lots of spam for 
something that just works out of the box with stock dmd!

Re: Introducing alid

[Ali Çehreli via Digitalmars-d-announce]

On 9/12/22 09:34, rikki cattermole wrote:

> dub.json
> errornogc/alid/errornogc.d
> circularblocks/alid/circularblocks.d

Considering I may want to let the users import the entire package as 
well with


  import alid;

how can I achieve my goal of subpackages? Telling me to forget about 
subpackages altogether :) is an option because I've already spent hours 
trying to find my way through different directory structures, random 
dub.json edits, experimenting with ways of stopping dub from fetching 
and using the bad version of the repo repeatedly, and many other random 
things...


The main problem is likely I don't have an accurate mental model of the 
whole thing. For example, I don't know whether dub will take the state 
of my local directory as-is or will it copy from the committed state of 
the local repo? I ask because the add-local command puts a commit hash 
as well. If that's the case, I wouldn't want to commit every little 
change before seeing it's effect.


Yes, I did the path override thing as well...

Ali