1. 32-bit Ints in Eu2C dlls
Ok, I've uploaded the code. It should be on the Recent Contributions page
shortly, and it's up on my page now (only 4K download):
http://www14.brinkster.com/matthewlewis/projects.html
Sorry, Juergen, I guess I was only scanning that Newbie thread. I'd done
this back in October, but I guess I got distracted by other stuff, plus I
think I wanted to wait until 2.5 came out, so I could make sure it would
work there, too.
Matt Lewis
2. Re: 32-bit Ints in Eu2C dlls
Matt Lewis wrote:
> Ok, I've uploaded the code. It should be on the Recent Contributions page
> shortly, and it's up on my page now (only 4K download):
>
> http://www14.brinkster.com/matthewlewis/projects.html
>
> Sorry, Juergen, I guess I was only scanning that Newbie thread. I'd done
> this back in October, but I guess I got distracted by other stuff, plus I
> think I wanted to wait until 2.5 came out, so I could make sure it would
> work there, too.
Not a problem at all.
Again you "did the impossible". Well, I probably shouldn't be surprised
by that any more. 
Thanks a lot!!
I think I was able to "dot an i". I sent you the code privately.
After my amateurish understanding, it should also be possible to build
some similar code directly into the translator, shouldn't it?
Regards,
Juergen
3. Re: 32-bit Ints in Eu2C dlls
Juergen Luethje wrote:
>
> Not a problem at all.
> Again you "did the impossible". Well, I probably shouldn't be surprised
> by that any more. Thanks a lot!!
> I think I was able to "dot an i". I sent you the code privately.
Got it. Thanks. It was something I had sort of thought about doing,
but was too lazy to think too hard about it, since it didn't really do
anything, but I think I probably will add it in.
> After my amateurish understanding, it should also be possible to build
> some similar code directly into the translator, shouldn't it?
>
In theory, yes. The problem, I believe comes if you try to call the routine
from another Eu routine within the dll. I suppose that Rob could check to
see if that happens, and only do it with routines that don't have that
issue (it would be simlar to his handling of routine_id, I suppose).
Matt Lewis
4. Re: 32-bit Ints in Eu2C dlls
Matt Lewis wrote:
> Juergen Luethje wrote:
>>
>> Not a problem at all.
>> Again you "did the impossible". Well, I probably shouldn't be surprised
>> by that any more. Thanks a lot!!
>> I think I was able to "dot an i". I sent you the code privately.
>
> Got it. Thanks. It was something I had sort of thought about doing,
> but was too lazy to think too hard about it, since it didn't really do
> anything, but I think I probably will add it in.
Of course it's a minor issue. I just thought people who have a close
look at a DLL might wonder what those routines are doing, and why I
didn't document them. I'm currently working on the next version of a
DLL, that I distribute on my website, so your tool comes at the right
time for me.
Yesterday I just used your small file 'test.ew' to test your program,
and then I had that idea, better not to export 'make_atom()' and
'make_atomu()'.
Today I built the current version of my DLL, using "make_atom". It
worked fine! After translating the code, I realized that the .def file
containd 114 routine names. But I only want 11 routines to be exported!
So it doesn't matter, whether or not there are 2 more names in the .def
file, it isn't "clean" anyway. Sorry, I didn't think of that yesterday.
So until I have a better idea I'll do the same as before, concerning the
.def file. For each project, I'll write the desired .def file by hand,
and after translation I'll replace the automatically produced .def file
with my hand-written .def file. The replacement can be done automatically
by a batch file.
The cause of this clutter is, that there is no way to tell the translator
explicitely, which routines I want to export, and that the translator
automatically exports *all* global routines -- even those, which are in
included files. Again that scope issue concerning "include" ... 
<snip>
Regards,
Juergen
5. Re: 32-bit Ints in Eu2C dlls
Me wrote:
<snip>
> Today I built the current version of my DLL, using "make_atom". It
> worked fine! After translating the code, I realized that the .def file
> containd 114 routine names. But I only want 11 routines to be exported!
> So it doesn't matter, whether or not there are 2 more names in the .def
> file, it isn't "clean" anyway. Sorry, I didn't think of that yesterday.
>
> So until I have a better idea I'll do the same as before, concerning the
> .def file. For each project, I'll write the desired .def file by hand,
> and after translation I'll replace the automatically produced .def file
> with my hand-written .def file. The replacement can be done automatically
> by a batch file.
>
> The cause of this clutter is, that there is no way to tell the translator
> explicitely, which routines I want to export, and that the translator
> automatically exports *all* global routines -- even those, which are in
> included files. Again that scope issue concerning "include" ...
Some hours ago, I uploaded the new version of my wrapper for the Eu2C
Windows translator to RDS' User Contributions page. It has the
additional option to export only global routines that are defined in the
main module of a DLL.
<shameless double-plug for the record>
Using your (Matt's) "32bits.zip" together with my "ecwr.zip" makes it
_easy_ now to translate Euphoria (at least) 2.4 and 2.5 programs, and
passing 32-bit integers to exported routines in DLLs and SOs, as well
as controlling which routines are exported, and including resources in
DLLs and EXE files on Windows. 
</shameless double-plug for the record>
Regards,
Juergen
