BOOL APIENTRY DllEntryPoint?

[Joerg Wunsch]

Hello,

I tried porting gpib-tcl to FreeBSD, and basically succeeded, but only

by #ifdef'ing this function:

BOOL APIENTRY DllEntryPoint(HINSTANCE hInstance, DWORD reason, LPVOID
lpCmdLine)

{

if (reason == DLL_PROCESS_DETACH)

GPIBDevice::CloseAllDevices();

return TRUE;

}

out. How would that possibly work on other Posix-style operating

systems?

Also, I needed a lot of hacks (like adding /usr/local/include and

/usr/local/lib to the Makefile) that would normally be better done by

an automatic system. I'm not a real fan of autoconf/automake, but

IMHO they would be appropriate here... Same for deciding the Tcl

version installed.

Finally, one inconsistency occurred to me in the code:

GPIBDevice::GetInterfaceSRQ uses TestSRQ() while the remaining stuff

uses all the ibfoo() functions. Shouldn't this be using iblines()

instead? (FreeBSD only supports the ibfoo() function interface, it

doesn't support the multidevice API functions.)

[Ian Dees]

Hi, Joerg.

How would that possibly work on other Posix-style operating

systems?

As you correctly point out, that function needs to be #ifdef'ed out for non-
Win32 systems.

I'm not a real fan of autoconf/automake, but

IMHO they would be appropriate here...

I agree; a distant horizon on the to-do list is to try to get the project
built under CMake.

GPIBDevice::GetInterfaceSRQ uses TestSRQ() while the remaining stuff

uses all the ibfoo() functions. Shouldn't this be using iblines()

instead?

Probably so. There may have been a reason to use TestSRQ() instead, but I
suspect it was just a lack of knowledge about iblines(). Is iblines() working
well on your BSD machine?

--Ian

[Joerg Wunsch]

As you correctly point out, that function needs to be #ifdef'ed

out for non-Win32 systems.

OK. I've only been surprised since I know this package is already

distributed as a binary for Linux, how that would have possibly

worked at all...

Shouldn't this be using iblines()

instead?

Probably so. There may have been a reason to use TestSRQ() instead, but I

suspect it was just a lack of knowledge about iblines().

I'll try filing a patch for this then.

Is iblines() working well on your BSD machine?

No, it's only implemented as a stub :), but at least, I can get it to compile
then.

As I'm also a FreeBSD developer, I could then always go ahead, and fill in

the stub when needed, but I first have to find an application for SRQs, i.e.

an instrument that would actually assert SRQ in a meaningful context.