Recursively expanded vars used as target deps are expanded
early when make is constructing the dep graph. Thus, $(TAR_DEPS)
will be empty when evaluated. For now use the $(DIST) as a quick
We need to cleanup all the deps as we pretty much ignore this
in favor of targets that will always be ran.
Signed-off-by: Spencer Shimko <sshimko(a)tresys.com>
rhel6/src/Makefile | 6 ++----
1 files changed, 2 insertions(+), 4 deletions(-)
diff --git a/rhel6/src/Makefile b/rhel6/src/Makefile
index 27d86af..9d75999 100644
@@ -23,8 +23,6 @@ SRPM_OUTPUT_DIR ?= $(OUTPUT_DIR)
# This file will be created by tar'ing up $(OUT)/
TARBALL = $(SRPM_OUTPUT_DIR)/$(DIST).tar.gz
-# exhaustive list of deps for the RPM, used to determine if RPM needs to be rebuilt
-TAR_DEPS = $(wildcard $(DIST)/*)
RPM_DEPS := $(TARBALL) $(RPM_SPEC) Makefile
RPM_TMPDIR ?= $(ROOT_DIR)/tmp
@@ -126,8 +124,8 @@ dist: content guide tables
-$(TARBALL): dist $(TAR_DEPS)
- tar -czf $@ $(TAR_DEPS)
+$(TARBALL): dist $(DIST)
+ tar -czf $@ $(DIST)
This series of patches modifies the method for generating RPMs
to remove hard-coded directory structure expectations and some other minor clean ups.
Spencer Shimko (4):
Switch rpm gen to use command-line macro defs.
Add generated files to ignore list.
Switch from curly braces to parens in Makefile
Minor cleanups in Makefile
rhel6/src/.gitignore | 2 +
rhel6/src/Makefile | 201 ++++++++++++--------
.../input/rpmbuild/scap-security-guide-alpha.spec | 11 +-
3 files changed, 132 insertions(+), 82 deletions(-)
Some patches have been posted to the list as attachments instead of
in-line. This makes it very difficult to provide feedback on parts of the
patch. I'd like to see everyone submitting patches use "git format-patch"
and "git send-email.. Additionally a [PATCH 0/?] should be generated and
a signed-off-by line added.
Basic workflow to accomplish this:
1. Modify code.
2. Pull from master and rebase your patches on top of any changes pulled
3. $ git format-patch --cover-letter -s HEAD~4 (where 4 equals the number
of commits you submitting to the list)
4. $ git send-email --to "scap-security-guide(a)lists.fedorahosted.org
I've had some discussions with Jeff and we are moving to the ACK/NACK
model. If you have commit access please wait for an ACK from another
maintainer prior to pushing to master. If the person submitting the patch
does not have commit access the maintainer providing the ACK should push
it into master.