Re: [Savonet-users] Liquidsoap on FreeBSD

2009-05-25 Thread Romain Beauxis
Hi !

Thanks for you detailled report. Here are some comments:

Le Tuesday 26 May 2009 05:30:10 Andrew, vous avez écrit :
 In order to get configure to go through cleanly I had to use this
 configure command:
 CFLAGS=-I/usr/local/include LDFLAGS=-L/usr/local/lib CPPFLAGS=-I/
 usr/local/include ./configure

Ok. That is probably normal.

 To compile with OSS support I needed to create /usr/src/include/linux
 and copy the following sound include files there:
 cp /usr/src/sys/dev/sound/pcm/sound.h /usr/src/include/linux/sound.h
 cp /usr/src/sys/sys/soundcard.h /usr/src/include/linux/soundcard.h

Could you check with the attached patch if it fixes oss compilation for you 
without this hack ?

 With the LDFLAGS on the command line I was able to make it go through
 the configure process successfully, however make would continually fail.

 To fix the issues with make I had to symlink the following libraries
 from /usr/local/lib to /usr/lib

 ln -sf /usr/local/lib/libmp3lame.so /usr/lib/libmp3lame.so
 ln -s /usr/local/lib/libmad.so /usr/lib/libmad.so
 ln -sf /usr/local/lib/libogg.so /usr/lib/libogg.so
 ln -sf /usr/local/lib/libvorbisenc.so /usr/lib/libvorbisenc.so
 ln -sf /usr/local/lib/libvorbisfile.so /usr/lib/libvorbisfile.so
 ln -sf /usr/local/lib/libshout.so /usr/lib/libshout.so
 ln -sf /usr/local/lib/libfaac.so /usr/lib/libfaac.so
 ln -sf /usr/local/lib/libfaad.so /usr/lib/libfaad.so
 ln -sf /usr/local/lib/libsamplerate.so /usr/lib/libsamplerate.so
 ln -sf /usr/local/lib/libtag.so /usr/lib/libtag.so
 ln -sf /usr/local/lib/libtag_c.so /usr/lib/libtag_c.so

Ok. We had found the issue for this some times ago.
I just made a big commit on trunk/ in order to fix this. 
Could you try with the latest SVN trunk if you can compile without this hack ?

 This fixed the problems of finding codecs and such, however despite
 specifically compiling using gmake it would try and use the FreeBSD
 make instead. This forced me to temporarily rename the BSD version and
 symlink gmake:

 mv /usr/bin/make /usr/bin/make.good
 ln -sf /usr/local/bin/gmake /usr/bin/make
 make

Hmm..
We have tried to remove all hardcoded make calls. Apparently there still are 
some of them..
Could you send us the compilation logs where this fails so that we can fix them 
?

 And lastly in order for 'make install' to work I had to add a user and
 group named liquidsoap using adduser:

 adduser

 make install

Hmmm..
So far, this part is assumed to be done by the user/administrator:
# User ${user} and group ${group} are expected to exist.
# They are defined in Makefile.defs, written by configure.

However, since you are not the first to complain, I wonder whether this could 
be done in a better way..

 This produced a working /usr/local/bin/liquidsoap binary that worked
 with OSS audio input.

Good ! You seemed quite patient :)

 For some reason though I had to remove the /usr/local/lib/liquidsoap/
 0.9.0/utils.liq file as the program would just hang there with it in
 place.

That is really strange.
Using liquidsoap without utils.liq will be mush less usefull. A lot of useful 
functions are defined there.. It would be good to debug and fix this issue..
Do you have more informations ?


Romain

Index: configure.ac
===
--- configure.ac	(révision 6589)
+++ configure.ac	(copie de travail)
@@ -1166,7 +1166,7 @@
 AC_ARG_ENABLE([oss],AC_HELP_STRING([--disable-oss],[don't use OSS]))
 
 if test x$enable_oss != xno ; then
-  AC_CHECK_HEADERS([linux/sound.h],[W_OSS=yes])
+  AC_CHECK_HEADERS([sys/soundcard.h],[W_OSS=yes])
 fi
 
 AC_SUBST(W_OSS)
Index: src/io/oss_io_c.c
===
--- src/io/oss_io_c.c	(révision 6589)
+++ src/io/oss_io_c.c	(copie de travail)
@@ -5,7 +5,7 @@
 #include caml/misc.h
 #include caml/mlvalues.h
 
-#include linux/soundcard.h
+#include sys/soundcard.h
 #include sys/ioctl.h
 #include assert.h
 
--
Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT
is a gathering of tech-side developers  brand creativity professionals. Meet
the minds behind Google Creative Lab, Visual Complexity, Processing,  
iPhoneDevCamp asthey present alongside digital heavyweights like Barbarian
Group, R/GA,  Big Spaceship. http://www.creativitycat.com ___
Savonet-users mailing list
Savonet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/savonet-users


Re: [Savonet-users] Liquidsoap on FreeBSD

2009-05-25 Thread Romain Beauxis
Le Tuesday 26 May 2009 13:36:27, vous avez écrit :
  Could you check with the attached patch if it fixes oss compilation  
  for you
  without this hack ?

 Unfortunately it did not. I got this:

 checking linux/sound.h usability... no
 checking linux/sound.h presence... no

 And at the end:
 - OSS   : no

 The patch did apply cleanly in the savonet/liquidsoap directory.

I forgot to mention that you need to run ./bootstrap again for this patch to 
work..

  Ok. We had found the issue for this some times ago.
  I just made a big commit on trunk/ in order to fix this.
  Could you try with the latest SVN trunk if you can compile without  
  this hack ?

 I just tried SVN revision 6591 and that seems to have fixed things.  
 However I did need to run gmake with the LDFLAGS:

 CFLAGS=-I/usr/local/include LDFLAGS=-L/usr/local/lib CPPFLAGS=-I/
 usr/local/include gmake

 after that it compiled cleanly without having to make any symlinks to  
 libraries or gmake.

Great !
It seems strange that you still need to pass the compilation variables. 
However, this is far much better..

  Hmmm..
  So far, this part is assumed to be done by the user/administrator:
  # User ${user} and group ${group} are expected to exist.
  # They are defined in Makefile.defs, written by configure.
 
  However, since you are not the first to complain, I wonder whether  
  this could
  be done in a better way..

 I think if there was some information in INSTALL that would be  
 suitable. It just threw me off because I wasn't expecting that. 

Yes. I think we should detect when the user/group does not exist and issue an 
error in this case..

 The LDFLAGS on the command line was also a bit non-intuitive as I'm used  
 to libraries being automatically discovered when running ./configure --
 prefix=/usr/local. It's not a big deal though.

Ok, I will look at that.

  That is really strange.
  Using liquidsoap without utils.liq will be mush less usefull. A lot  
  of useful
  functions are defined there.. It would be good to debug and fix this  
  issue..
  Do you have more informations ?

 The SVN releases seems to have fixed this issue as well.

Good !

 One note. With ocaml-faac and ocaml-faad enabled in PACKAGES it would  
 not compile cleanly. I don't particularly need regular AAC so that is  
 ok for my purposes, but I thought I would report it:

Ok, will look at that.


Romain

--
Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT
is a gathering of tech-side developers  brand creativity professionals. Meet
the minds behind Google Creative Lab, Visual Complexity, Processing,  
iPhoneDevCamp asthey present alongside digital heavyweights like Barbarian
Group, R/GA,  Big Spaceship. http://www.creativitycat.com 
___
Savonet-users mailing list
Savonet-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/savonet-users


Re: [Savonet-users] Liquidsoap on FreeBSD

2009-05-25 Thread Andrew


 Could you check with the attached patch if it fixes oss compilation  
 for you
 without this hack ?

Unfortunately it did not. I got this:

checking linux/sound.h usability... no
checking linux/sound.h presence... no

And at the end:
- OSS   : no

The patch did apply cleanly in the savonet/liquidsoap directory.


 Ok. We had found the issue for this some times ago.
 I just made a big commit on trunk/ in order to fix this.
 Could you try with the latest SVN trunk if you can compile without  
 this hack ?

I just tried SVN revision 6591 and that seems to have fixed things.  
However I did need to run gmake with the LDFLAGS:

CFLAGS=-I/usr/local/include LDFLAGS=-L/usr/local/lib CPPFLAGS=-I/ 
usr/local/include gmake

after that it compiled cleanly without having to make any symlinks to  
libraries or gmake.


 Hmmm..
 So far, this part is assumed to be done by the user/administrator:
 # User ${user} and group ${group} are expected to exist.
 # They are defined in Makefile.defs, written by configure.

 However, since you are not the first to complain, I wonder whether  
 this could
 be done in a better way..


I think if there was some information in INSTALL that would be  
suitable. It just threw me off because I wasn't expecting that. The  
LDFLAGS on the command line was also a bit non-intuitive as I'm used  
to libraries being automatically discovered when running ./configure -- 
prefix=/usr/local. It's not a big deal though.


 That is really strange.
 Using liquidsoap without utils.liq will be mush less usefull. A lot  
 of useful
 functions are defined there.. It would be good to debug and fix this  
 issue..
 Do you have more informations ?
The SVN releases seems to have fixed this issue as well.



One note. With ocaml-faac and ocaml-faad enabled in PACKAGES it would  
not compile cleanly. I don't particularly need regular AAC so that is  
ok for my purposes, but I thought I would report it:

gmake[3]: Entering directory `/root/downloads/liqsvn/savonet/ocaml- 
faad/src'
/usr/local/bin/ocamlc.opt -c -cc gcc -ccopt -fPIC -I/usr/local/ 
include  -Wall -DCAML_NAME_SPACE \
 -DPIC -I/usr/local/include   \
 -o faad_stubs.o  faad_stubs.c
faad_stubs.c:43:19: error: mp4ff.h: No such file or directory
faad_stubs.c: In function 'ocaml_faad_init':
faad_stubs.c:78: warning: passing argument 4 of 'NeAACDecInit' from  
incompatible pointer type
faad_stubs.c: In function 'ocaml_faad_init2':
faad_stubs.c:95: warning: passing argument 4 of 'NeAACDecInit2' from  
incompatible pointer type
faad_stubs.c: At top level:
faad_stubs.c:150: error: expected specifier-qualifier-list before  
'mp4ff_t'
faad_stubs.c: In function 'finalize_mp4':
faad_stubs.c:165: error: 'mp4_t' has no member named 'ff'
faad_stubs.c:166: warning: implicit declaration of function  
'mp4ff_close'
faad_stubs.c:166: error: 'mp4_t' has no member named 'ff'
faad_stubs.c:167: error: 'mp4_t' has no member named 'read_cb'
faad_stubs.c:168: error: 'mp4_t' has no member named 'read_cb'
faad_stubs.c:169: error: 'mp4_t' has no member named 'write_cb'
faad_stubs.c:170: error: 'mp4_t' has no member named 'write_cb'
faad_stubs.c:171: error: 'mp4_t' has no member named 'seek_cb'
faad_stubs.c:172: error: 'mp4_t' has no member named 'seek_cb'
faad_stubs.c:173: error: 'mp4_t' has no member named 'trunc_cb'
faad_stubs.c:174: error: 'mp4_t' has no member named 'trunc_cb'
faad_stubs.c: In function 'read_cb':
faad_stubs.c:195: error: 'mp4_t' has no member named 'fd'
faad_stubs.c:196: error: 'mp4_t' has no member named 'fd'
faad_stubs.c:200: error: 'mp4_t' has no member named 'read_cb'
faad_stubs.c:210: warning: control reaches end of non-void function
faad_stubs.c: In function 'write_cb':
faad_stubs.c:216: error: 'mp4_t' has no member named 'fd'
faad_stubs.c:217: error: 'mp4_t' has no member named 'fd'
faad_stubs.c:220: warning: control reaches end of non-void function
faad_stubs.c: In function 'seek_cb':
faad_stubs.c:228: error: 'mp4_t' has no member named 'fd'
faad_stubs.c:229: error: 'mp4_t' has no member named 'fd'
faad_stubs.c:233: error: 'mp4_t' has no member named 'seek_cb'
faad_stubs.c:240: warning: control reaches end of non-void function
faad_stubs.c: In function 'ocaml_faad_mp4_open_read':
faad_stubs.c:255: error: 'mp4_t' has no member named 'fd'
faad_stubs.c:256: error: 'mp4_t' has no member named 'ff_cb'
faad_stubs.c:257: error: 'mp4_t' has no member named 'read_cb'
faad_stubs.c:258: error: 'mp4_t' has no member named 'read_cb'
faad_stubs.c:261: error: 'mp4_t' has no member named 'ff_cb'
faad_stubs.c:262: error: 'mp4_t' has no member named 'write_cb'
faad_stubs.c:263: error: 'mp4_t' has no member named 'write_cb'
faad_stubs.c:267: error: 'mp4_t' has no member named 'ff_cb'
faad_stubs.c:268: error: 'mp4_t' has no member named 'write_cb'
faad_stubs.c:272: error: 'mp4_t' has no member named 'ff_cb'
faad_stubs.c:273: error: 'mp4_t' has no member named 'seek_cb'
faad_stubs.c:274: