Problem with secondary expansion and target specific variables
The following makefile.1 produces correct results: --- target1 : var := one target2 : var := two targets := target1 target2 all : $(targets) .SECONDEXPANSION: $(targets) : foo/$$(var)/bar foo/%/bar : mkdir -p $(dir $@) touch $@ .PHONY : target1 target2 - > make -f makefile.1 mkdir -p foo/one/ touch foo/one/bar mkdir -p foo/two/ touch foo/two/bar However, makefile.2 does not pick up var: - target1 : var := one target2 : var := two targets := target1 target2 all : $(targets) .SECONDEXPANSION: $(targets) : test/$$@ $(targets:%=test/%) : foo/$$(var)/bar foo/%/bar : mkdir -p $(dir $@) touch $@ .PHONY : $(targets) $(targets:%=test/%) > make -f makefile.2 make: *** No rule to make target `foo//bar', needed by `test/target1'. Stop. According to make documentation: "when you define a target-specific variable that variable value is also in effect for all prerequisites of this target, and all their prerequisites, etc." Since test/target1 is a prerequisite for target1 and var is defined for target1, it should be available also for test/target1. Secondary expansion works when expanding 'var' for target1 and target2 in makefile.1, but it doesn't when expanding it in makefile.2 Am I doing something incorrect here or is it an unimplemented functionality, specifically that target specific variables are not inherited in prerequisites of prerequisite? The behavior is the same in 3.81 and 3.82. ___ Bug-make mailing list Bug-make@gnu.org https://lists.gnu.org/mailman/listinfo/bug-make
Dependency violation in parallel build (3.82, regression)
Using the attached Makefile, I get the following result: % rm -f foo; make -j foo1 bar1 qux start foo bar foo So the qux recipe is run before foo is finished (or even started), although qux depends on foo. The $(shell) calls as well as .SECONDARY and the seemingly unrelated bar rules all seem to be necessary to trigger the bug. This is reproducible with 3.82. It does not seem to occur with 3.81. BTW, I wanted to check against the current repository version, but when I tried to fetch it as described on https://savannah.gnu.org/git/?group=make, I got: % git clone git://git.savannah.gnu.org/make.git Cloning into make... warning: You appear to have cloned an empty repository. Makefile Description: Binary data ___ Bug-make mailing list Bug-make@gnu.org https://lists.gnu.org/mailman/listinfo/bug-make
Re: Dependency violation in parallel build (3.82, regression)
> Date: Sun, 30 Dec 2012 23:38:44 +0100 > From: Frank Heckenbach > > BTW, I wanted to check against the current repository version, but > when I tried to fetch it as described on > https://savannah.gnu.org/git/?group=make, I got: > > % git clone git://git.savannah.gnu.org/make.git > Cloning into make... > warning: You appear to have cloned an empty repository. Use CVS, not git, to fetch the latest development sources of Make. ___ Bug-make mailing list Bug-make@gnu.org https://lists.gnu.org/mailman/listinfo/bug-make
