Lots of changes, got the first complete(-ish) build with the new extended

build system.  I left as much of the old functionality in place as possible,
definitely were it was being used.  But there are some changes that break
backwards compatibility, thought they should be noticed by few, and be easy to
fix going forward.

Some highlights:

 - centralized patch system (packages/patches with targets patch_pd and
   unpatch_pd)

 - easily redirected builds, using INSTALL_PREFIX and all of the *_DEST
   variables.  This makes building packages like Pd.app, .deb, .rpm,
   etc. much easier.

 - libdir format: basically a libdir is a directory that has both the objects
   and the help files together in one folder.  It can be added using -lib or
   the StartUp preferences, or you can access them via geiger namespaces,
   i.e. [mylibrary/myobject].

 - special characters allow in setup function/file names for objects.  This
   makes objects like [||~] possible without having to be in a library.  Now
   they can be either .pd files or individual .pd_darwin files (thanks
   IOhannes for the patch).

svn path=/trunk/; revision=3993

1 files changed, 22 insertions, 21 deletions

diff --git a/packages/TODO b/packages/TODO
index 45d83e5d..7525eefc 100644
--- a/packages/TODO
+++ b/packages/TODO
@@ -1,12 +1,26 @@
 
-- make abstraction -lib loader.  when a -lib mylibrary name is tried and fails
-  in the path, then fails with IO's altname, then try to load as a path on top
-  of extra: i.e. extra/mylibrary.  If found, add to path, and add
-  doc/5.reference/mylibrary to helppath.  
+- global versioning system
+  		the package releases should have versions of there own which apply to
+  		all packages, regardless of format or OS.  This will make for easy and
+  		seemless cross-platform work.
+
+- global optimization flags
+  		ideally, you would use a version of Pd that is fully optimized for your
+  		CPU.  In order to do this, a system of optimization flags needs to be
+  		developed so that you can specific the target CPU in one place and have
+  		everything compile for that CPU.
+
+- libdir
+
+	 - document libdir format
+
+	 - write .sh script that generates the mylibrary/mylibrary.pd file with the
+  	 	meta data.  It should take command line arguments, and put them into the
+  	 	meta file.
 
-    - call it sys_load_lib_folder();, 
+- rename "applications" to "examples"
 
-	 - insert it after  if (!sys_load_lib()) in s_main.c
+- downcase *_DEST and *_SRC to be more readable
 
 - make pd section since I can't easily modify the pd tree
 
@@ -18,24 +32,11 @@
 
 - merge in tigital's Makefile++
 
-- move purepd to abstractions 
-
 - add GRIPD for Anders
 
 - add reg keys for Martin Dupras
 
 - mDNS objects!
 
-- import nqpoly4 and write helpfiles based on website: 
-  			http://pix.test.at/pd/nqpoly/nqpoly4.html
-
-- global versioning system
-  		the package releases should have versions of there own which apply to
-  		all packages, regardless of format or OS.  This will make for easy and
-  		seemless cross-platform work.
-
-- global optimization flags
-  		ideally, you would use a version of Pd that is fully optimized for your
-  		CPU.  In order to do this, a system of optimization flags needs to be
-  		developed so that you can specific the target CPU in one place and have
-  		everything compile for that CPU.
+- make standard targets: distclean (*~, etc.), installdirs (make just the
+  dirs), uninstall instead of install_clean