Hello all,
I'm trying to compile the 2.6.5 tarball on a non-developer system
(a freshly installed OpenIndiana zone). Right away the make failed
due to revision checks (fixed below) due to missing git and svn
software (and for make, a failed step fails everything). The fix
is to catch the non-zero return codes and just successfully set
an empty string.
Barring this, the default compilation (so far without CGI and
some drivers) seems to succeed.
###################
--- include/Makefile.orig Mon Nov 4 13:12:39 2013
+++ include/Makefile Mon Nov 4 13:23:02 2013
@@ -473,9 +473,11 @@
# magic to include SVN revision number in NUT version string
nut_version.h: FORCE
- @GITREV=`git describe --tags 2>/dev/null | sed
's/^v\([0-9]\)/\1/' `; \
+ @GITREV=`git describe --tags 2>/dev/null | sed
's/^v\([0-9]\)/\1/' ` || \
+ GITREV=""; \
if [ -z "$$GITREV" ]; \
- then SVNREV=`LANG=C svnversion -n $(top_srcdir)
2>/dev/null`; \
+ then SVNREV=`LANG=C svnversion -n $(top_srcdir)
2>/dev/null` || \
+ SVNREV=""; \
if [ -z "$$SVNREV" -o "$$SVNREV" =
"exported" ]; \
then NUT_VERSION="$(PACKAGE_VERSION)"; \
else
NUT_VERSION="$(PACKAGE_VERSION)-$$SVNREV"; \
On Nov 4, 2013, at 4:37 AM, Jim Klimov wrote:> Hello all, > > I'm trying to compile the 2.6.5 tarball on a non-developer system > (a freshly installed OpenIndiana zone). Right away the make failed > due to revision checks (fixed below) due to missing git and svn > software (and for make, a failed step fails everything). The fix > is to catch the non-zero return codes and just successfully set > an empty string.Thanks. I remember checking to make sure it worked if one or the other of git and svnversion failed, but apparently not both. -- Charles Lepple clepple at gmail
Seemingly Similar Threads
- eliminating GNU Make features in include/Makefile.am
- [nut-commits] svn commit r1207 - in trunk: . include
- Patch -- SVN revision in the version string
- SUMMARY: mounting datasets from a read-only pool with aid of tmpfs
- [LLVMdev] [PATCH] llvm-config: add svn revision to --version