[csw-devel] proposed alteration to git patching mechanism

Ben Walton bwalton at opencsw.org
Sun Mar 18 18:13:59 CET 2012


Hi All,

Maciej found a problem with the makepatch workflow he used and it
boils down to:

mgar configure
vi ...
mgar makepatch
add PATCHFILES += ... to Makefile
mgar reconfigure build

The patch targets would then see an un-applied patch and trigger a
chain of events that culminated in post-patch-gitsnap failing as the
branches/tags it works on are already setup.

The following patch fixes up some cookie files after creating the new
patch file.  It's sort of a cheap hack, but I don't see a nicer way to
do this right now.  (Suggestions welcomed.)

Does anyone see a problem with committing this?

Thanks
-Ben

Index: gar.mk
===================================================================
--- gar.mk	(revision 17428)
+++ gar.mk	(working copy)
@@ -558,7 +558,14 @@
 			echo; \
 			echo "	svn add" $$FILES_PATCHES; \
 			echo; \
-			mv $$NEWPATCHES $(abspath $(FILEDIR)); ) \
+			mv $$NEWPATCHES $(abspath $(FILEDIR)); \
+			for p in $$NEWPATCHES; do \
+				touch $(abspath $(COOKIEDIR))/normal-patch-$$p; \
+				touch $(abspath $(COOKIEDIR))/patch-extract-$$p; \
+			done; \
+			touch $(abspath $(COOKIEDIR))/extract-modulated; \
+			touch $(abspath $(COOKIEDIR))/patch-modulated; \
+			touch $(abspath $(COOKIEDIR))/post-patch-gitsnap; ) \
 		fi; \
 	    else \
 		echo "No extracted sources so we can't create patches..."; \
--
Ben Walton
Systems Programmer - CHASS
University of Toronto
C:416.407.5610 | W:416.978.4302



More information about the devel mailing list