I have the following project structure.

my_proj/
    script_1/runme_1.d
    script_2/runme_2.d
    evaluator.d

The project has a library dependency:

"dependencies": {
    "mir": "~>3.2.0",
},

The "evaluator.d" script contains some general functions that convert dataset and pass it to "runme_1.d" and "runme_2.d" scripts the following way:

import mir.ndslice;

void main() {
    import script_1.runme_1 : runTask1;
    import script_2.runme_2 : runTask2;

    mir_slice!(real*, 1LU, cast(mir_slice_kind)2)[]) dataset = convertDataset("data.txt");
    runTask1(dataset);
    //runTask2(dataset); //disable for now
}

So, basically what I need is to be able to run "./evaluator" and get the results by running the corresponding scripts.

I am trying to build the project the following way:

dub build --compiler=ldc2 --single evaluator.d

I get the following error:

Performing "debug" build using ldc2 for x86_64.
mir-core 1.0.2: target for configuration "library" is up to date.
mir-algorithm 3.7.2: target for configuration "default" is up to date.
mir-linux-kernel 1.0.1: target for configuration "library" is up to date.
mir-random 2.2.8: target for configuration "extended" is up to date.
survey ~master: building configuration "application"...
lld-link: error: undefined symbol: _D5runme_114runTask1FS3mir7ndslice5slice__T9mir_sliceTPeVmi2VEQBoQBnQBi14mir_slice_kindi2ZQBvZv
>>> referenced by C:\Users\user\my_proj\evaluator.d:89
>>>               .dub\obj\evaluator.obj:(_Dmain)
Error: linking with LLD failed
ldc2 failed with exit code 1.

How should I organize my project for this kind of setup?

(I am sorry for the absence of code formatting. I could not find any quick howto here)

On 12/11/19 2:18 PM, Pavel Shkadzko wrote:
> I have the following project structure.
> 
> my_proj/
>      script_1/runme_1.d
>      script_2/runme_2.d
>      evaluator.d
> 
> The project has a library dependency:
> 
> "dependencies": {
>      "mir": "~>3.2.0",
> },
> 
> The "evaluator.d" script contains some general functions that convert dataset and pass it to "runme_1.d" and "runme_2.d" scripts the following way:
> 
> import mir.ndslice;
> 
> void main() {
>      import script_1.runme_1 : runTask1;
>      import script_2.runme_2 : runTask2;
> 
>      mir_slice!(real*, 1LU, cast(mir_slice_kind)2)[]) dataset = convertDataset("data.txt");
>      runTask1(dataset);
>      //runTask2(dataset); //disable for now
> }
> 
> So, basically what I need is to be able to run "./evaluator" and get the results by running the corresponding scripts.
> 
> I am trying to build the project the following way:
> 
> dub build --compiler=ldc2 --single evaluator.d
> 
> I get the following error:
> 
> Performing "debug" build using ldc2 for x86_64.
> mir-core 1.0.2: target for configuration "library" is up to date.
> mir-algorithm 3.7.2: target for configuration "default" is up to date.
> mir-linux-kernel 1.0.1: target for configuration "library" is up to date.
> mir-random 2.2.8: target for configuration "extended" is up to date.
> survey ~master: building configuration "application"...
> lld-link: error: undefined symbol: _D5runme_114runTask1FS3mir7ndslice5slice__T9mir_sliceTPeVmi2VEQBoQBnQBi14mir_slice_kindi2ZQBvZv 
> 
>>>> referenced by C:\Users\user\my_proj\evaluator.d:89
>>>>               .dub\obj\evaluator.obj:(_Dmain)
> Error: linking with LLD failed
> ldc2 failed with exit code 1.
> 
> How should I organize my project for this kind of setup?
> 
> (I am sorry for the absence of code formatting. I could not find any quick howto here)
> 
> 

I would restructure your
folders this way:
my_proj/
    source/
        script_1/runme_1.d
        script_2/runme_2.d
        evaluator.d

then using `dub --compiler=ldc2`

On Wednesday, 11 December 2019 at 13:00:32 UTC, drug wrote:
>
> I would restructure your
> folders this way:
> my_proj/
>     source/
>         script_1/runme_1.d
>         script_2/runme_2.d
>         evaluator.d
>
> then using `dub --compiler=ldc2`

It works like this, yes. However, by default it builds a "library" while I want to run "evaluator" as a script which uses other scripts by calling their functions.

Doing:
dub --compiler=ldc2 --single evaluator.d // throws an error

source\evaluator.d(93,12): Error: module runme_1 is in file 'runme_1.d' which cannot be read

On 12/11/19 6:52 PM, Pavel Shkadzko wrote:
> 
> It works like this, yes. However, by default it builds a "library" while I want to run "evaluator" as a script which uses other scripts by calling their functions.
> 
> Doing:
> dub --compiler=ldc2 --single evaluator.d // throws an error
> 
> source\evaluator.d(93,12): Error: module runme_1 is in file 'runme_1.d' which cannot be read
> 
Add to dub.json:
"targetType": "executable"

and it will build executable. Or rename evaluator.d to app.d
It is confused that there is no app.d so it thinks that you build a library.

On Wednesday, 11 December 2019 at 16:06:30 UTC, drug wrote:

> Add to dub.json:
> "targetType": "executable"
>
> and it will build executable. Or rename evaluator.d to app.d
> It is confused that there is no app.d so it thinks that you build a library.

I added "targetType" and now everything works without renaming to the script to "app.d". Thank you!