On 16/10/2023 4:31 PM, dan wrote:
> I suppose if i could figure out a way to make all other modules depend on my module this would happen, but the module which uses the variable i want to set is in some already-compiled dynamic library that i would prefer not to touch.

If its in a shared library, then that shared library gets setup prior to your binary.

There is nothing that you can do. You gotta override rather than initialize.

On Monday, 16 October 2023 at 03:33:55 UTC, Richard (Rikki) Andrew Cattermole wrote:
> On 16/10/2023 4:31 PM, dan wrote:
>> I suppose if i could figure out a way to make all other modules depend on my module this would happen, but the module which uses the variable i want to set is in some already-compiled dynamic library that i would prefer not to touch.
>
> If its in a shared library, then that shared library gets setup prior to your binary.
>
> There is nothing that you can do. You gotta override rather than initialize.

Thanks Rikki.

I was wrong in my statement of the problem: it is not a dynamic library, but rather a static library, libgtkd-3.a.

I apologize for being so careless.

libgtkd-3.a throws an exception before main is reached, in the Loader.d file, in a method with signature 'public static void loadLibrary(string library)'.  That method is used trying to load a library which really is dynamic, libatk-1.0.0.dylib.  That library is in my system, among the Mac ports files, in a standard place, namely /opt/local/lib.  I can set the environment variable GTK_BASEPATH to help it out, and that works, but i would like to do all of this inside the executable rather than outside the executable.

The Loader.d file depends on std.process.

So, given that i was wrong and it is not a dynamic library i'm trying to get in ahead of, but a static one, is there a way to execute a small snippet of code after std.process is initialized, but before any other code (such as Loader.d) uses it?

Thanks in advance for any ideas.

dan

Okay, after looking at gtkd, I don't think this can be solved with module constructors or swapping out to lazy initialization.

One way that might work however is to use a crt_constructor as that runs before the D stuff. However it would now hard code your program to your system. Over all my suspicion is that there is something wrong with your system related to library lookup paths and that should be fixed instead.

```d
void main() {
    import std.stdio;
    import std.process;
    writeln(environment["GTK_BASEPATH"]);
}

pragma(crt_constructor) extern(C) void myEnvironmentVarSetter() {
    import core.sys.posix.stdlib : putenv;
    putenv(cast(char*)"GTK_BASEPATH=~/bin/gtk".ptr);
}
```

On Monday, 16 October 2023 at 10:23:54 UTC, Richard (Rikki) Andrew Cattermole wrote:
> Okay, after looking at gtkd, I don't think this can be solved with module constructors or swapping out to lazy initialization.
>
> One way that might work however is to use a crt_constructor as that runs before the D stuff. However it would now hard code your program to your system. Over all my suspicion is that there is something wrong with your system related to library lookup paths and that should be fixed instead.
>
> ```d
> void main() {
>     import std.stdio;
>     import std.process;
>     writeln(environment["GTK_BASEPATH"]);
> }
>
> pragma(crt_constructor) extern(C) void myEnvironmentVarSetter() {
>     import core.sys.posix.stdlib : putenv;
>     putenv(cast(char*)"GTK_BASEPATH=~/bin/gtk".ptr);
> }
> ```

Really awesome, Rikki, your code does the trick.  Thank you so much for your pragma.


Since you are so right about the code, maybe you are also right about my system.  For reference, i'll describe it in case anybody else somehow wanders into the same situation.

My system is a Mac Catalina (OSX 10.15).

I installed the latest dmd i could find, directly from the Digital Mars website.  This was so that i could use any ports system (macports, fink, or brew) and keep the same d compiler.  The dmd compiler is in /usr/local/bin/dmd, and the files are in some standard location /Library/Developer/CommandLineTools/SDKs/MacOSX10.15.sdk/.....

I installed the gtk stuff (but not gtkd) from macports, and that all went into /opt/local, which is where macports puts things.

I believe that gtkd does not have a 'configure' script, so i directly edited the top level GNUMakefile in one spot, to specify a prefix; my changed line was
prefix?=/opt/some-particular-path-distinct-from-local-and-all-others

I wanted the installation of gtkd to be in some parallel location so that there could be multiple versions if needed, or it could be removed if needed without any chance of disturbing macports or anything else.

It built ok, and the demo worked ok (setting GTK_BASEPATH of course, per gtkd's README.md).  Note that i built it on an account for which the PATH has /opt/local/bin and /opt/local/sbin at the start.  So when gtkd was under construction, it knew just where to find all the gtk files.

Thanks again for pointing out the crt_constructor pragma.

(Now, i still think that when module initialization order is not forced, it should be something a programmer or systems integrator can choose, but i don't want to be too greedy.)

Thanks again for your help!!

dan