On 12/20/2013 08:55 PM, Jack Howarth wrote:
Taj,
The new cloog 0.18.2 release fails when a configure build tries to install...
make install DESTDIR=/sw/src/fink.build/root-cloog-org2-0.18.2-1 VERBOSE=1
Making install in .
test -z "/sw/lib" || autoconf/install-sh -c -d "/sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/lib"
/bin/sh ./libtool --mode=install /usr/bin/install -c libcloog-isl.la '/sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/lib'
libtool: install: /usr/bin/install -c .libs/libcloog-isl.4.dylib /sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/lib/libcloog-isl.4.dylib
libtool: install: (cd /sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/lib && { ln -s -f libcloog-isl.4.dylib libcloog-isl.dylib || { rm -f libcloog-isl.dylib && ln -s libcloog-isl.4.dylib libcloog-isl.dylib; }; })
libtool: install: /usr/bin/install -c .libs/libcloog-isl.lai /sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/lib/libcloog-isl.la
libtool: install: /usr/bin/install -c .libs/libcloog-isl.a /sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/lib/libcloog-isl.a
libtool: install: chmod 644 /sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/lib/libcloog-isl.a
libtool: install: ranlib /sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/lib/libcloog-isl.a
libtool: warning: remember to run 'libtool --finish /sw/lib'
test -z "/sw/bin" || autoconf/install-sh -c -d "/sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/bin"
/bin/sh ./libtool --mode=install /usr/bin/install -c cloog '/sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/bin'
libtool: install: /usr/bin/install -c .libs/cloog /sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/bin/cloog
/usr/bin/install -c -m 644 cloog-isl.pc "/sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/lib/pkgconfig/cloog-isl.pc"
/usr/bin/install -c -m 644 "./cmake/isl-config.cmake" "/sw/src/fink.build/root-cloog-org2-0.18.2-1/sw/lib/isl/"
install: ./cmake/isl-config.cmake: No such file or directory
make[2]: *** [install-data-local] Error 71
make[1]: *** [install-am] Error 2
Apparently the new release has dropped the cmake subdirectory from the toplevel of the source distribution.
Thanks Jack for noticing so quickly. It seems the release was not tested
in that way.
Taj Khan, it seems we need a subsequent release to fix this. (Do not
recreate the tar ball with different content, as this will break all MD5
hashes of distributions).
Cheers,
Tobias
On 12/20/2013 08:55 PM, Jack Howarth wrote:
Thanks Jack for noticing so quickly. It seems the release was not tested
in that way.
Taj Khan, it seems we need a subsequent release to fix this. (Do not
recreate the tar ball with different content, as this will break all MD5
hashes of distributions).
Cheers,
Tobias