Hey ho,
On Wed, 12 Feb 2003, Adam Lindsay wrote:
Looks like it. Did you see my email to Michael McGonagle 2 weeks ago? The linux-ish flags, for example: LINUXLDFLAGS = -export_dynamic -shared
ah yes, sorry, did see that, just didn't register when I was trying this
should pretty much be converted into:
MACOSXLDFLAGS = -bundle -bundle_loader /path/to/pd/bin/pd -flat_namespace
ok tried this, but now getting (from trying to compile cxc/counter):
ld -bundle -bundle_loader /usr/local/pd/bin/pd -flat_namespace -o cxc.pd_darwin *.o -lc -lm ld: Undefined symbols: dyld_stub_binding_helper saveFP make: *** [darwin] Error 1
Which is just (on the topic of usability) making me nuts. I am now so &^%@£ed off that I'm just pining for Debian. But, this is a departmental computer, and making PD work on OSX is something I said I'd do, precisely because a lot of MAX/MSP users here would like to/might try PD instead...
But I'm just not getting very far with this. Will try again later if I feel less depressed ;)
Also, if you're modifying a makefile, best to do a bit of modified cut- and-paste and create a pd_darwin target alongside the pd_linux target. I notice from the error that your makefile is trying to make 'rhythm_ioi_histogram.pd_linux', which can create problems on the Mac, even if you succeed. ;)
Ah yes, quite. But of course, when you get a makefile like rhythm_estimator's (which to the ignorant eye looks hugely more complicated than, say, chaos's, which already has multiple targets), I just don't even know where to begin separating the bits and declaring additional targets...
Is there -any- chance that people who made 'cxc' and 'rhythm_estimator' might take a look at their makefiles to make them more osx friendly (yup, I realise people have lives)?
Meantime, any comment on error above would help, and be appreciated...
In general, should I just trash makefiles in there (for whatever external does not have a darwin target) and just slink a darwin target in myself? Because it seems to me that there is -large- differences in makefile styles across the externals...which makes it rather confusing.
Is it possible that we (you, the actual developers) might agree on guidelines for external makefiles, so little bugs like me can better study/learn the practice and start to understand?
d
Cheers, adam
and it does the same for other externals, too. of course, taking those flags out screws things up.
Muh?
hum.
david casal said this at Wed, 12 Feb 2003 13:58:16 +0000:
Hi all,
I keep getting:
In file included from /usr/include/math.h:24, from bangfilt.c:5: /usr/include/architecture/ppc/math.h:228: warning: declaration of `exp' shadows a global declaration
and:
In file included from /usr/include/math.h:24, from rhythm_ioi_histogram.c:22: /usr/include/architecture/ppc/math.h:228: warning: declaration of `exp' shadows a global declaration /usr/include/architecture/ppc/math.h:214: warning: shadowed declaration is here
the two above are (respectively) from trying to compile all externals from a tweaked makefile in the build/darwin dir, and from trying to compile rhythm_estimator.
since I had success in compiling ann.pd_darwin using chaos's makefile, I thought I'd venture into more dangerous territory. Obviously foolish of me.
(all I wanted was to compile cxc/counter, but it -would- indeed be nice to crack compiling all externals on darwin from cvs)
Any thoughts, anyone?
d
david casal --0+
d.casal@uea.ac.uk --9+
ecdc.dyndns.org/dc --)+
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
--
Adam T. Lindsay atl@comp.lancs.ac.uk Computing Dept, Lancaster University +44(0)1524/594.537 Lancaster, LA1 4YR, UK Fax:+44(0)1524/593.608 -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
david casal --0+
d.casal@uea.ac.uk --9+
ecdc.dyndns.org/dc --)+
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
--
Adam T. Lindsay atl@comp.lancs.ac.uk Computing Dept, Lancaster University +44(0)1524/594.537 Lancaster, LA1 4YR, UK Fax:+44(0)1524/593.608 -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
david casal --0+
d.casal@uea.ac.uk --9+
ecdc.dyndns.org/dc --)+
Hallo, david casal hat gesagt: // david casal wrote:
Is there -any- chance that people who made 'cxc' and 'rhythm_estimator' might take a look at their makefiles to make them more osx friendly (yup, I realise people have lives)?
I didn't write these, but I wrote some other externals that don't have OS-X makefiles and not even W32-makefiles. To me, the problem is simply access to those machines. I could get a hand at Win 2000 and OS-X machines at work, but there I don't have the time to install a development environment.
This is why my externals come with Linux-only makefiles. There's not much I can do about it if I don't want to include totally untested makefiles written by guessing. But I welcome anyone to commit their W32 and OS-X makefiles into my CVS-directory on sf.net.
ciao
Frank Barknecht _ ______footils.org__
Hello,
On Thu, 13 Feb 2003, Frank Barknecht wrote:
This is why my externals come with Linux-only makefiles. There's not much I can do about it if I don't want to include totally untested makefiles written by guessing. But I welcome anyone to commit their W32 and OS-X makefiles into my CVS-directory on sf.net.
I understand, of course. Well, if I get further with this (though I suspect that Adam will be of much more use in this respect), I will definitely commit makefiles I know work. As Adam suggested, anybody trying to compile stuff in OSX should -add- targets instead of making new makefiles, so I'll try to do that where I can, otherwise I'll commit 'makefile_darwin's...
I'd like to begin by putting the ann external makefile I made in there...can I have permission to do that? WHo's 'forum::für::umläute' again? Is it Iohannes?
d
david casal --0+
d.casal@uea.ac.uk --9+
ecdc.dyndns.org/dc --)+
I think what Frank says applies to most of the pd externals on CVS. There is a directory /build/src which is a collection of links to externals that can be built from these single files.
I compile them with a "simple" makefile from /build/linux
There is build/darwin, which is in need of something similar (or better, considering that not all externals can be built from one file, and some of them need additional libraries).
But for now, probably you can start from the /build/linux/makefile and do something similar. Would be great.
Guenter
On Thu, 13 Feb 2003, Frank Barknecht wrote:
Hallo, david casal hat gesagt: // david casal wrote:
Is there -any- chance that people who made 'cxc' and 'rhythm_estimator' might take a look at their makefiles to make them more osx friendly (yup, I realise people have lives)?
I didn't write these, but I wrote some other externals that don't have OS-X makefiles and not even W32-makefiles. To me, the problem is simply access to those machines. I could get a hand at Win 2000 and OS-X machines at work, but there I don't have the time to install a development environment.
This is why my externals come with Linux-only makefiles. There's not much I can do about it if I don't want to include totally untested makefiles written by guessing. But I welcome anyone to commit their W32 and OS-X makefiles into my CVS-directory on sf.net.
ciao
Frank Barknecht _ ______footils.org__
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
Hey,
On Thu, 13 Feb 2003, guenter geiger wrote:
I compile them with a "simple" makefile from /build/linux
as do I, when I am at home in the blissful arms of debian 2.4.18. unfortunately, I work in a department full of macs running osx (which I am learning to love/hate). Try running Cocoon/Jboss in osx 10.2 (sorry OT)
There is build/darwin, which is in need of something similar (or better, considering that not all externals can be built from one file, and some of them need additional libraries).
yup, and build/darwin contains a suspiciously linux-looking makefile ;)
But for now, probably you can start from the /build/linux/makefile and do something similar. Would be great.
Well following Adam's advice, I did start with that one and tried, but there's a lot to do. I suspect that Adam would do better than I will. Maybe, if I just go about compiling everything and finding bits, he could help in integrating those into the general makefile? Adam?
So far, I've managed to produce a makefile which works for ann, maxlib had n osx target already (just did -Wno-shadow and -Wno-error), and I'm now trawling through the rest.
d
On Thu, 13 Feb 2003, Frank Barknecht wrote:
Hallo, david casal hat gesagt: // david casal wrote:
Is there -any- chance that people who made 'cxc' and 'rhythm_estimator' might take a look at their makefiles to make them more osx friendly (yup, I realise people have lives)?
I didn't write these, but I wrote some other externals that don't have OS-X makefiles and not even W32-makefiles. To me, the problem is simply access to those machines. I could get a hand at Win 2000 and OS-X machines at work, but there I don't have the time to install a development environment.
This is why my externals come with Linux-only makefiles. There's not much I can do about it if I don't want to include totally untested makefiles written by guessing. But I welcome anyone to commit their W32 and OS-X makefiles into my CVS-directory on sf.net.
ciao
Frank Barknecht _ ______footils.org__
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
david casal --0+
d.casal@uea.ac.uk --9+
ecdc.dyndns.org/dc --)+
FYI: I've almost completed a proper OS X makefile for ext13 and have started on ggee.
h.c
On Friday, Feb 14, 2003, at 03:38 America/New_York, david casal wrote:
Hey,
On Thu, 13 Feb 2003, guenter geiger wrote:
I compile them with a "simple" makefile from /build/linux
as do I, when I am at home in the blissful arms of debian 2.4.18. unfortunately, I work in a department full of macs running osx (which I am learning to love/hate). Try running Cocoon/Jboss in osx 10.2 (sorry OT)
There is build/darwin, which is in need of something similar (or better, considering that not all externals can be built from one file, and some of them need additional libraries).
yup, and build/darwin contains a suspiciously linux-looking makefile ;)
But for now, probably you can start from the /build/linux/makefile and do something similar. Would be great.
Well following Adam's advice, I did start with that one and tried, but there's a lot to do. I suspect that Adam would do better than I will. Maybe, if I just go about compiling everything and finding bits, he could help in integrating those into the general makefile? Adam?
So far, I've managed to produce a makefile which works for ann, maxlib had n osx target already (just did -Wno-shadow and -Wno-error), and I'm now trawling through the rest.
d
On Thu, 13 Feb 2003, Frank Barknecht wrote:
Hallo, david casal hat gesagt: // david casal wrote:
Is there -any- chance that people who made 'cxc' and 'rhythm_estimator' might take a look at their makefiles to make them more osx friendly (yup, I realise people have lives)?
I didn't write these, but I wrote some other externals that don't have OS-X makefiles and not even W32-makefiles. To me, the problem is simply access to those machines. I could get a hand at Win 2000 and OS-X machines at work, but there I don't have the time to install a development environment.
This is why my externals come with Linux-only makefiles. There's not much I can do about it if I don't want to include totally untested makefiles written by guessing. But I welcome anyone to commit their W32 and OS-X makefiles into my CVS-directory on sf.net.
ciao
Frank Barknecht _ ______footils.org__
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
david casal --0+
d.casal@uea.ac.uk --9+
ecdc.dyndns.org/dc --)+
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
FYI: I've hacked together a makefile for externals/build/darwin, but am having troubles committing to CVS.
I believe Hans-Christoph Steiner said this around Sun, 16 Feb 2003:
FYI: I've almost completed a proper OS X makefile for ext13 and have started on ggee.
h.c
On Friday, Feb 14, 2003, at 03:38 America/New_York, david casal wrote:
Hey,
On Thu, 13 Feb 2003, guenter geiger wrote:
I compile them with a "simple" makefile from /build/linux
as do I, when I am at home in the blissful arms of debian 2.4.18. unfortunately, I work in a department full of macs running osx (which I am learning to love/hate). Try running Cocoon/Jboss in osx 10.2 (sorry OT)
There is build/darwin, which is in need of something similar (or better, considering that not all externals can be built from one file, and some of them need additional libraries).
yup, and build/darwin contains a suspiciously linux-looking makefile ;)
But for now, probably you can start from the /build/linux/makefile and do something similar. Would be great.
Well following Adam's advice, I did start with that one and tried, but there's a lot to do. I suspect that Adam would do better than I will. Maybe, if I just go about compiling everything and finding bits, he could help in integrating those into the general makefile? Adam?
So far, I've managed to produce a makefile which works for ann, maxlib had n osx target already (just did -Wno-shadow and -Wno-error), and I'm now trawling through the rest.
d
On Thu, 13 Feb 2003, Frank Barknecht wrote:
Hallo, david casal hat gesagt: // david casal wrote:
Is there -any- chance that people who made 'cxc' and 'rhythm_estimator' might take a look at their makefiles to make them more osx friendly (yup, I realise people have lives)?
I didn't write these, but I wrote some other externals that don't have OS-X makefiles and not even W32-makefiles. To me, the problem is simply access to those machines. I could get a hand at Win 2000 and OS-X machines at work, but there I don't have the time to install a development environment.
This is why my externals come with Linux-only makefiles. There's not much I can do about it if I don't want to include totally untested makefiles written by guessing. But I welcome anyone to commit their W32 and OS-X makefiles into my CVS-directory on sf.net.
ciao
Frank Barknecht _ ______footils.org__
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
david casal --0+
d.casal@uea.ac.uk --9+
ecdc.dyndns.org/dc --)+
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
PD-list mailing list PD-list@iem.kug.ac.at http://iem.kug.ac.at/cgi-bin/mailman/listinfo/pd-list
--
Adam Lindsay +44(0)1524 594 537 atl@comp.lancs.ac.uk http://www.comp.lancs.ac.uk/computing/users/atl/ -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Hi, guenter geiger schrieb:
I think what Frank says applies to most of the pd externals on CVS. There is a directory /build/src which is a collection of links to externals that can be built from these single files.
I compile them with a "simple" makefile from /build/linux
Currently I'm using a makefile-system 'borrowed' from Thomas Grill's makefiles, so his, tested, Darwin-files can be used for my externals as well, with little changes, I suppose.
But in the end, I'd like to use autoconf, it's only, that my first trys resulted in externals, that couldn't be loaded. When I'm home again, I'll send this to pd-devel, maybe some configure-wizard can help me with that. Having an autoconf'ed build system *should* be much easier to port to other systems like OS-X, shouldn't it?
ciao
Frank Barknecht
On Fri, 14 Feb 2003, Frank Barknecht wrote:
Hi, guenter geiger schrieb:
I think what Frank says applies to most of the pd externals on CVS. There is a directory /build/src which is a collection of links to externals that can be built from these single files.
I compile them with a "simple" makefile from /build/linux
Currently I'm using a makefile-system 'borrowed' from Thomas Grill's makefiles, so his, tested, Darwin-files can be used for my externals as well, with little changes, I suppose.
But in the end, I'd like to use autoconf, it's only, that my first trys resulted in externals, that couldn't be loaded. When I'm home again, I'll send this to pd-devel, maybe some configure-wizard can help me with that. Having an autoconf'ed build system *should* be much easier to port to other systems like OS-X, shouldn't it?
:) depends.
I have not used autoconf, because I want to put some restrictions on the externals. If possible, they should not depend on external libraries. This should make them more portable.
If there are no external dependencies, there is no need to use autoconf, because the libraries do not have to be searched for.
I think the same thing applies for flext externals, unless you want to test for flext version, etc....
Guenter
-
Adam Lindsay -
david casal -
Frank Barknecht -
guenter geiger -
Hans-Christoph Steiner