Re: [ecasound] specifics of the FreeBSD port

Subject: Re: [ecasound] specifics of the FreeBSD port
From: The Anarcat (anarcat_AT_anarcat.dyndns.org)
Date: Tue Nov 27 2001 - 03:55:07 EET

• Next message: Jesse Chappell: "[ecasound] ladspa control outputs"
• Previous message: Jeremy Hall: "Re: [ecasound] problems with an old frames_per_cycle worth of audio after unmuting"
• Maybe reply: The Anarcat: "Re: [ecasound] specifics of the FreeBSD port"

On Sun Nov 25, 2001 at 10:46:37PM -0500, k_AT_eca.cx wrote:
> On Tue, 20 Nov 2001, The Anarcat wrote:
>
> > Please note that ecasound now needs 2 new patches to compile properly on
> > FreeBSD, the first:
> [..]
> > +++ libecasound/audioio-wave.h Mon Nov 19 14:02:55 2001
> > +/* #include <stdint.h> */
> > #include <sys/types.h>
> [...]
> > is strange. I do not know what this header (stdint) is, but it's not on
> > FreeBSD. Ecasound compiles fine without it. This should probably be:
>
> stdint.h is part of the new C-language standard (C99). The original C89
> standard specifies 15 standard headers, C95 adds three more and the C99
> yet more six. stdint.h is one of these six. Previously there was no
> platform indepedent way to declare explicit-length variables like
> uint32_t, etc. On most (?) Unix/POSIX systems these are declared in
> <sys/types.h> and that's why the above works on FreeBSD.

Interesting. I'll check on FreeBSD side to see why the standard isn't followed.

> > #ifdef HAVE_STDINT_H
> > #include <stdint.h>
> > #endif
>
> Yup, this is probably the best fix. I'll commit this to CVS.

Good.

> > The second patch is to have ecawave working again on FreeBSD. It makes
> > more debugging info available and removes a check on group ownership:
>
> It seems that I have already made this change in both ecasound-2.0
> and 2.1 CVS-trees. So 'st_gid == getegid()' is not enforced anymore.

Hmm... Well, the port of ecasound on FreeBSD is 2.0.3 right now, and it
seems that the tarball downloadable from the various sites doesn't
contain the fix.

Should I upgrade the port to another version?

> > Note: ecawave needs some configure hacks to detect FreeBSD's Qt2
> > properly and an installed ecasound. In other words:
> [...]
> > CONFIGURE_ENV+= CPPFLAGS="-I${LOCALBASE}/include" \
> > LDFLAGS="-L${LOCALBASE}/lib -lqt2 ${LDFLAGS}" \
> > CFLAGS="-I${LOCALBASE}/include ${CFLAGS}" \
> > CXXFLAGS="-I${LOCALBASE}/include ${CXXFLAGS}"
>
> Is this handled similarly for other freebsd-ports of Qt2 apps (ie. no need
> to change anything in ecawave ;))?

I don't know, but ecawave compiles fine with these. Actually, by a quick
look, I think most do not have that sort of hacks...

I think some of these things should be handled by ecasound-config. For
example, eacsound-config yields:

anarcat_AT_shall[/usr/ports]% ecasound-config --libs
-L/usr/local/lib -lecasound -lkvutils

here. One shouldn't have to specify -L flags then. I might be mistaken,
though. I will need to recompile ecawave without some of these hacks to
be sure they're necessary.

Anyways, ecawave is ported. I guess it just needs time to get into the
tree.

A.

• application/pgp-signature attachment: stored

--
To unsubscribe send message 'unsubscribe' in the body of the
message to <ecasound-list-request_AT_wakkanet.fi>.

• Next message: Jesse Chappell: "[ecasound] ladspa control outputs"
• Previous message: Jeremy Hall: "Re: [ecasound] problems with an old frames_per_cycle worth of audio after unmuting"
• Maybe reply: The Anarcat: "Re: [ecasound] specifics of the FreeBSD port"

This archive was generated by hypermail 2b28 : Tue Nov 27 2001 - 03:47:47 EET