• mike (58/58) Sep 07 2006 Hi!
• Walter Bright (2/6) Sep 07 2006 True, it's not very nice. (C++ has the exact same issue.)
• Max Samuha (5/51) Sep 14 2006 I can't beleive that anybody except me is trying to write a VST host
• mike (17/21) Sep 14 2006 Hehe ... and there I thought I was the only one that had that idea :-)
• Max Samuha (8/16) Sep 19 2006 i was not planning to convert anything other than aeffect.h and
• mike (15/22) Sep 20 2006 Yeah, you're right. And after seeing how much stuff is there in the =
• Max Samuha (4/21) Sep 21 2006 If you think MIDI is dead, you might be interested to know that VST is
• mike (34/37) Sep 22 2006 Thanks!

mike <vertex gmx.at> writes:

Hi!

Maybe you read my last posts about VST plugins ... just to clear that up=
,  =

I don't want to leave wrong info here: They've got a C interface, so all=
  =

my writing about linking C++ was bs. Just in case anybody wants to do so=
me  =

VST development in D. Go ahead, no problem with D either way (host or  =

plugin - both works when one converts the SDK files to D). I'm already  =

happily loading VST plugins and toying aroung with them.

Anyway, I've found something annoying, nothing really bad, but just  =

terribly annoying for a medium-level coder like me.

 From my code:

' void *proc =3D GetProcAddress(hlib, toStringz("main"));
' if(proc is null)
' {
'     [snip]
' }
'
' auto getNewPluginInstance =3D cast(mainFunc)proc;
'
' // get new instance
' AEffect *newInstance =3D null;
' newInstance =3D getNewPluginInstance(&audioMaster);

That works. Why? Because:

' extern (C)
' {
'     typedef AEffect *function(audioMasterCallback) mainFunc;
'     [snip]
' }

There's something that didn't work (Access violation when calling  =

getNewPluginInstance) and it cost me hours of debugging and trying out  =

insanely unlogical things, until I had the idea to define mainFunc as a =
 =

type. Good thing is I found a lot of bugs in advance. Well, that's how i=
t  =

originally was:

' auto getNewPluginInstance =3D cast(AEffect  =

*function(audioMasterCallback))proc;

Without the typdef. Access violation, because getNewPluginInstance has D=
  =

linkage. And there's no way to cast that void * pointer to a C function.=
  =

Defining the mainFunc type solves this.

Just wanted to point that out. After half a year with D - this was the  =

first time I ever got angry at this language.

' auto getNewPluginInstance =3D cast(extern (C) AEffect  =

*function(audioMasterCallback))proc;

That would be a logical and intuitive thing - at least for me. If you ha=
ve  =

a call-once-forget-afterwards function with C linkage you shouldn't be  =

forced to define a type. I even had the whole module extern (C). Didn't =
 =

work.

-Mike

-- =

Erstellt mit Operas revolution=E4rem E-Mail-Modul: http://www.opera.com/=
mail/

Walter Bright <newshound digitalmars.com> writes:

mike wrote:
 That would be a logical and intuitive thing - at least for me. If you 
 have a call-once-forget-afterwards function with C linkage you shouldn't 
 be forced to define a type. I even had the whole module extern (C). 
 Didn't work.

True, it's not very nice. (C++ has the exact same issue.)

Max Samuha <maxter i.com.ua> writes:

On Fri, 08 Sep 2006 03:03:40 +0200, mike <vertex gmx.at> wrote:

Hi!

Maybe you read my last posts about VST plugins ... just to clear that up,  
I don't want to leave wrong info here: They've got a C interface, so all  
my writing about linking C++ was bs. Just in case anybody wants to do some  
VST development in D. Go ahead, no problem with D either way (host or  
plugin - both works when one converts the SDK files to D). I'm already  
happily loading VST plugins and toying aroung with them.

Anyway, I've found something annoying, nothing really bad, but just  
terribly annoying for a medium-level coder like me.

 From my code:

' void *proc = GetProcAddress(hlib, toStringz("main"));
' if(proc is null)
' {
'     [snip]
' }
'
' auto getNewPluginInstance = cast(mainFunc)proc;
'
' // get new instance
' AEffect *newInstance = null;
' newInstance = getNewPluginInstance(&audioMaster);

That works. Why? Because:

' extern (C)
' {
'     typedef AEffect *function(audioMasterCallback) mainFunc;
'     [snip]
' }

There's something that didn't work (Access violation when calling  
getNewPluginInstance) and it cost me hours of debugging and trying out  
insanely unlogical things, until I had the idea to define mainFunc as a  
type. Good thing is I found a lot of bugs in advance. Well, that's how it  
originally was:

' auto getNewPluginInstance = cast(AEffect  
*function(audioMasterCallback))proc;

Without the typdef. Access violation, because getNewPluginInstance has D  
linkage. And there's no way to cast that void * pointer to a C function.  
Defining the mainFunc type solves this.

Just wanted to point that out. After half a year with D - this was the  
first time I ever got angry at this language.

' auto getNewPluginInstance = cast(extern (C) AEffect  
*function(audioMasterCallback))proc;

That would be a logical and intuitive thing - at least for me. If you have  
a call-once-forget-afterwards function with C linkage you shouldn't be  
forced to define a type. I even had the whole module extern (C). Didn't  
work.

-Mike

I can't beleive that anybody except me is trying to write a VST host
in D! :). I was thinking of a way to call the VST plugin from D
through the C function pointer and was going to ask the NG members for
help but mike was nice enough to do that for me! Isn't it a miracle?

mike <vertex gmx.at> writes:

Am 14.09.2006, 14:58 Uhr, schrieb Max Samuha <maxter i.com.ua>:

 I can't beleive that anybody except me is trying to write a VST host
 in D! :). I was thinking of a way to call the VST plugin from D
 through the C function pointer and was going to ask the NG members for=

 help but mike was nice enough to do that for me! Isn't it a miracle?

Hehe ... and there I thought I was the only one that had that idea :-)

When you're working with VST too ... any interest in setting up a VST AP=
I  =

project on dsource.org for VST plugin developers? I was thinking of doin=
g  =

that since I've converted most of the header files anyway, but I'm stuck=
  =

with some PHP and VBA development at the moment. I think that could be a=
  =

good thing for D, getting all those freeware VST developers interested i=
n  =

the language.

-Mike

-- =

Erstellt mit Operas revolution=E4rem E-Mail-Modul: http://www.opera.com/=
mail/

Max Samuha <maxter i.com.ua> writes:

On Fri, 15 Sep 2006 00:59:20 +0200, mike <vertex gmx.at> wrote:

Hehe ... and there I thought I was the only one that had that idea :-)

When you're working with VST too ... any interest in setting up a VST API  
project on dsource.org for VST plugin developers? I was thinking of doing  
that since I've converted most of the header files anyway, but I'm stuck  
with some PHP and VBA development at the moment. I think that could be a  
good thing for D, getting all those freeware VST developers interested in  
the language.

-Mike

i was not planning to convert anything other than aeffect.h and
aeffectx.h. IMO, this is quite a trivial task and apparently doesn't
require a separate project at dsource. I'm developing a lightweight
real-time VST host with freely routable signal paths. The whole thing
is totally experimental and there is still not much to share with the
community. Actually, nothing but scanning for plugins and extracting
information about them for now.

mike <vertex gmx.at> writes:

Am 19.09.2006, 09:00 Uhr, schrieb Max Samuha <maxter i.com.ua>:

 i was not planning to convert anything other than aeffect.h and
 aeffectx.h. IMO, this is quite a trivial task and apparently doesn't
 require a separate project at dsource. I'm developing a lightweight
 real-time VST host with freely routable signal paths. The whole thing
 is totally experimental and there is still not much to share with the
 community. Actually, nothing but scanning for plugins and extracting
 information about them for now.

Yeah, you're right. And after seeing how much stuff is there in the  =

C++ side of the SDK ... no, better invest time in the project itself.
I'm also currently only extracting info, working on the UI for now, unti=
l  =

I can at least build some simple setups for testing. I'm writing a host =
 =

oriented for realtime composing and live improvisation, with OSC  =

integration (MIDI should be dead by now, really). Got nothing fancy  =

either, just a window with some basic controls and a console full of  =

debugging messages :)

-mike

-- =

Erstellt mit Operas revolution=E4rem E-Mail-Modul: http://www.opera.com/=
mail/

Max Samuha <maxter i.com.ua> writes:

On Wed, 20 Sep 2006 17:47:12 +0200, mike <vertex gmx.at> wrote:

Am 19.09.2006, 09:00 Uhr, schrieb Max Samuha <maxter i.com.ua>:

 i was not planning to convert anything other than aeffect.h and
 aeffectx.h. IMO, this is quite a trivial task and apparently doesn't
 require a separate project at dsource. I'm developing a lightweight
 real-time VST host with freely routable signal paths. The whole thing
 is totally experimental and there is still not much to share with the
 community. Actually, nothing but scanning for plugins and extracting
 information about them for now.

Yeah, you're right. And after seeing how much stuff is there in the  
C++ side of the SDK ... no, better invest time in the project itself.
I'm also currently only extracting info, working on the UI for now, until  
I can at least build some simple setups for testing. I'm writing a host  
oriented for realtime composing and live improvisation, with OSC  
integration (MIDI should be dead by now, really). Got nothing fancy  
either, just a window with some basic controls and a console full of  
debugging messages :)

-mike


If you think MIDI is dead, you might be interested to know that VST is
almost dead, too:) At least this guy claims it is dying:

mike <vertex gmx.at> writes:

Am 21.09.2006, 13:59 Uhr, schrieb Max Samuha <maxter i.com.ua>:

 If you think MIDI is dead, you might be interested to know that VST is=

 almost dead, too:) At least this guy claims it is dying:


Thanks!

Now that was a great read, although he could have calmed down a bit. He =
 =

really must hate VST a lot. I especially liked: "Yeah, VST may have been=
  =

around for 5 years or so before AU came along, yet it still doesn't have=
  =

any host documentation, example code, or anything!!! Every developer who=
  =

wants to host AUs [sic! That should read: VSTs] faces a constant uphill =
 =

battle, trying to figure out how to host VSTs via reverse psychological =
 =

analysis of the plugin SDK, which is pretty damn pathetic itself already=
."

My only two reference sources available (probably the two everybody find=
s  =

and uses) are one app written in delphi and another one written in C++, =
 =

with an indentation level of 1 (one) space per level and a note from the=
  =

author complaining that everybody steals his code so he won't release ne=
w  =

versions as source code anymore. And the SDK doesn't help you in any cas=
e,  =

just remember my first topic, I was so confused how to even load a VST. =
 =

Now if only AU would work on Windows, everything would be perfect :)

-Mike

(I fear that's way off topic, maybe we should discuss via email?)

-- =

Erstellt mit Operas revolution=E4rem E-Mail-Modul: http://www.opera.com/=
mail/