linuxcnc - linuxcnc - enhanced machine controller (EMC)

Age	Commit message (Collapse)	Author	Files	Lines
2013-06-17	Add license notices where they were missing	Jeff Epler	11	-0/+165

2009-06-20	clean up ignores some more	Jeff Epler	6	-24/+0

2009-06-20	convert cvs ignores to git ignores	Jeff Epler	6	-0/+0

2009-06-20	Get rid of RTAPI_SUCCESS, use 0 instead	Jeff Epler	8	-28/+28
	The test "retval < 0" should feel familiar; it's the same kind of test you use in userspace (returns -1 for error) and in kernel space (returns -ERRNO for error) Signed-off-by: Jeff Epler <jepler@unpythonic.net>
2006-03-27	remove unused files	Jeff Epler	5	-254/+0

2006-02-10	Revamp the build system of emc2.	Jeff Epler	1	-79/+0
	The make is "almost nonrecursive": All userspace targets are built from a single invocation of make, and then all realtime targets are built from a second invocation of make. For more information about why this is a good thing, read http://www.pcug.org.au/~millerp/rmch/recu-make-cons-harm.html On one test system, the time for a complete rebuild dropped 20% from 3:17 to 2:37, and the time for the subsequent "make" (which, in the new build system, built no targets) decreased 90% from 16 seconds to less than 2 seconds. A subsequent "make userspace" takes less than 0.3 seconds. Unlike the old build, "-j" can be used to improve build speed when multiple CPUs are available. Also, all dependencies among userspace targets should be correct and automatically generated, so the need to 'make clean' after a CVS update should go away, and there is no separate 'make dep' step. ('make modclean' may still be necessary if the kernel makefiles do not properly track dependencies) 'make clean' does not remove all files in ../bin, which means it is no longer necessary to reinstall axis after cleaning. librs274 is now built as a shared library. The new makefiles do not * have the 'install' targets * have support for building modules against kernel 2.4 * build the items from rtapi/examples None of these should represent unsurmountable limitations of the new system.
2006-02-08	use '$(MAKE) -Cxxx yyy' instead of '(cd xxx; make yyy)'. Using $(MAKE) ↵	Jeff Epler	1	-11/+11
	instead of make has various advantages as described in the 'make' info pages, and using -Cxxx instead of '(cd xxx; ...) means that directories that don't exist will trigger an error instead of a make that eternally re-invokes itself in the same directory.
2006-01-24	complete cvsignores for the whole tree. I don't get any "?"s on update anymore	Jeff Epler	6	-0/+24

2005-07-23	removed -fstrict_aliasing and -Wno-strict-aliasing... the latter caused ↵	John Kasunich	1	-2/+0
	compile errors on 3 of 4 BDI systems in the compile farm, and the former may not be safe for kernel modules and related code
2005-07-22	fix make	Paul Corner	2	-1/+3

2005-07-16	minor tweak to get things to work on BDI-2.xx	John Kasunich	1	-1/+1

2005-07-14	modified the RTAPI examples Makefiles so that they build on 2.6 and other ↵	Alex Joni	6	-12/+116
	kbuild systems
2005-04-27	removed a layer of '../' from a few more lower level makefiles	John Kasunich	5	-10/+10

2004-01-28	Reverted a couple of files that were changed in error	Paul Corner	2	-20/+13

2004-01-25	Base set of files committed so that work can progress on emc2	Paul Corner	2	-13/+20

2003-12-23	Minor oops....	Paul Corner	1	-1/+1

2003-12-23	Range checking on msg_level in the procfs	Paul Corner	1	-1/+1

2003-11-17	restructured halmeter, tweaked makefiles, ran indent on all source files	John Kasunich	10	-711/+731

2003-11-15	added last two rtapi examples	John Kasunich	7	-0/+633

2003-11-15	added semaphore RTAPI example	John Kasunich	6	-0/+314

2003-11-07	added first two rtapi examples	John Kasunich	7	-0/+385