So being a noob and all I've gotten stuck on something stupid...
Whenever I try to compile using DMD, and try to link .lib files I
always get errors like:

"file4.d(2): Error: module file3 is in file 'file3lib\file3.d'
which cannot be read
import path[0] = %cd%\file3
import path[1] = C:\dmd2\Windows\bin\..\..\src\phobos
import path[2] = C:\dmd2\Windows\bin\..\..\src\druntime\import
"

For reference on this I've been using Adam's book and this page:
http://wiki.dlang.org/Compiling_and_linking_with_DMD_on_Windows


I also tried using COFF2OMF and COFFIMPLIB (everytime I used
coffimplib it just said "Error: Missing archive signature")

Here's the code I was working with (plz excuse my lazy
variable/file names, wasn't feeling creative this morning :P):

File3.d which I turned into a .lib
---------------------------------
module file3lib.file3;;;
import std.stdio;

class shit
{
	int lead;
	this(int y){lead += y;}
	int fuck(int x)
	{
		lead += x;
		return lead;
	}
}

/*
void main()
{
	auto okay = new shit(10);
	writeln(okay.fuck(2));
}
*/
----------------------------------

file4.d
----------------------------------
module file4;
import file3lib.file3;
import std.stdio;

void main()
{
	auto thing = new shit(2);
	writeln(thing.fuck(2));
}
----------------------------------



I tried using different names, and making extra directories but
that didn't help much ofcourse..
+I know there's that weird thing I did with file3lib & file3, but
I swapped that around a bunch of times with no success.
Everything's just been getting messier as I go on.

I've sortof just been going in circles so if anyone could point
out what I'm doing wrong it'd be a huge help! :)

When you compile the final program, the library .d file needs to be available too, either in the folder based on its name or passed straight to dmd explicitly.

Despite the presence of the .lib file, the .d file is still needed so it can get code prototypes and type names, etc., out of it.

(If you have the .d though, the lib is rarely needed. Indeed, the way I do most D libraries is to just pass all the .d files to the compiler at once and forget about lib files.)

On Friday, 6 June 2014 at 16:33:27 UTC, Adam D. Ruppe wrote:
> When you compile the final program, the library .d file needs to be available too, either in the folder based on its name or passed straight to dmd explicitly.
>
> Despite the presence of the .lib file, the .d file is still needed so it can get code prototypes and type names, etc., out of it.
>
> (If you have the .d though, the lib is rarely needed. Indeed, the way I do most D libraries is to just pass all the .d files to the compiler at once and forget about lib files.)

Hey Adam!
I have kept all the files in one folder. So if I don't need to
explicitly call the .lib in DMD, does that mean the .lib is just
a passive object? Should I make libs in place of object files?

On Friday, 6 June 2014 at 16:41:24 UTC, K.K. wrote:
> I have kept all the files in one folder. So if I don't need to
> explicitly call the .lib in DMD, does that mean the .lib is just
> a passive object? Should I make libs in place of object files?

If you pass all the .d files at once, you don't need the .lib at all and the compiler will also just do one object file.

Simplest (and usually fastest!) way to build:

dmd yourfile.d otherfile.d lib/file.d lib/otherfile.d otherlib/file.d

and so on - then don't worry about object or lib files. Of course if you need a dll or something that's different...

On Friday, 6 June 2014 at 16:52:24 UTC, Adam D. Ruppe wrote:
> On Friday, 6 June 2014 at 16:41:24 UTC, K.K. wrote:
>> I have kept all the files in one folder. So if I don't need to
>> explicitly call the .lib in DMD, does that mean the .lib is just
>> a passive object? Should I make libs in place of object files?
>
> If you pass all the .d files at once, you don't need the .lib at all and the compiler will also just do one object file.
>
> Simplest (and usually fastest!) way to build:
>
> dmd yourfile.d otherfile.d lib/file.d lib/otherfile.d otherlib/file.d
>
> and so on - then don't worry about object or lib files. Of course if you need a dll or something that's different...

Oh okay, I get what you mean. I guess I was really over
complicating it then? xD

Thanks for the help Adam! -I'm totally digging your book btw!

(I'll save the trouble of figuring out dll's for another day!)

On Friday, 6 June 2014 at 17:02:33 UTC, K.K. wrote:
> Oh okay, I get what you mean. I guess I was really over
> complicating it then? xD

Yea, I think a lot of people do: building C++ takes some extra steps out of necessity that you can just ignore in D for the most part :)

> Thanks for the help Adam! -I'm totally digging your book btw!

Cool, write amazon reviews when you're done too if you feel like it :P

BTW also feel free to ask here or to even email me if you see anything in the book that you want clarification or more info on.

On Saturday, 7 June 2014 at 16:06:39 UTC, Adam D. Ruppe wrote:
> Yea, I think a lot of people do: building C++ takes some extra steps out of necessity that you can just ignore in D for the most part :)
>

I still don't think I've fully figured out compiling in C++! For
compiling the way you showed me, it almost reminds of running
and/or compiling python. :3


> Cool, write amazon reviews when you're done too if you feel like it :P
>

I'm nowhere near done with the book, but I happily wrote you an
Amazon review already :)

(I think it said they have to review my review first or
something? Idk, despite all the shopping on amazon that I do, I
rarely write reviews)

> BTW also feel free to ask here or to even email me if you see anything in the book that you want clarification or more info on.

Thanks I'll keep it in mind!

On Friday, 6 June 2014 at 16:33:27 UTC, Adam D. Ruppe wrote:
> When you compile the final program, the library .d file needs to be available too, either in the folder based on its name or passed straight to dmd explicitly.
>
> Despite the presence of the .lib file, the .d file is still needed so it can get code prototypes and type names, etc., out of it.
>
> (If you have the .d though, the lib is rarely needed. Indeed, the way I do most D libraries is to just pass all the .d files to the compiler at once and forget about lib files.)

You are not ought to have .d files to link with library, because to link you need only interfaces, not implementation (except for templates, since they are not included in object files), otherwise there would be impossible to create closed-source developer libraries. You can generate .di files with dmd -c -o- -H option and use them.