1. 2.5 alpha ecu problems
Rob,
It seems that gcc (v3.2 at least) doesn't like the new check that you do
for a call_proc/func() for calling convention. I fixed all the errors
that occurred by commenting out the _stdcall parts (which was what was
being complained about). It then compiled and linked OK, although I haven't
been able to get the .so to work with open_dll(). I ran 'nm' on it,
and it reported out a bunch of symbols. I was able to get a trivial so
to work. Not sure yet what the deal is.
I was trying to compile ooeu_trace.e into a shared lib.
Matt Lewis
2. Re: 2.5 alpha ecu problems
Matt Lewis wrote:
>
>
> Rob,
>
> It seems that gcc (v3.2 at least) doesn't like the new check that you do
> for a call_proc/func() for calling convention. I fixed all the errors
> that occurred by commenting out the _stdcall parts (which was what was
> being complained about). It then compiled and linked OK, although I haven't
> been able to get the .so to work with open_dll(). I ran 'nm' on it,
> and it reported out a bunch of symbols. I was able to get a trivial so
Ok, tried again, and after the manual edits, it linked just fine, and
works with open_dll() and everything.
Matt Lewis
3. Re: 2.5 alpha ecu problems
Matt Lewis wrote:
> It seems that gcc (v3.2 at least) doesn't like the new check that you do
> for a call_proc/func() for calling convention.
Thanks. On Linux/FreeBSD there is only one calling convention,
so that bug fix code shouldn't have been emitted. I'll take it out.
It was intended to fix a problem with call_proc/func in a
Windows .dll, that might call an exported (_stdcall) routine.
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
