Summary: | dev-libs/crypto++-5.6.1 ignores LDFLAGS (including its own) on HPPA | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Jeroen Roovers (RETIRED) <jer> |
Component: | [OLD] Library | Assignee: | Crypto team [DISABLED] <crypto+disabled> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | hppa |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 343373 | ||
Attachments: | dev-libs:crypto++-5.6.1:20101102-212701.log [hppa,fails tests] |
Description
Jeroen Roovers (RETIRED)
2010-11-02 22:14:29 UTC
Created attachment 252967 [details]
dev-libs:crypto++-5.6.1:20101102-212701.log [hppa,fails tests]
It ignores my LDFLAGS as well as its own (notably -pthread, which causes the test suite to fail. The only place in GNUMakefile where LDFLAGS is being /unset/ is in: ifneq ($(IS_SUN_CC),0) # override flags for CC Sun C++ compiler CXXFLAGS = -DNDEBUG -O -g0 -native -template=no%extdef $(M32OR64) LDFLAGS = AR = $(CXX) But the equivalent of: IS_SUN_CC = $(shell $(CXX) -V 2>&1 | $(EGREP) -c "CC: Sun") returns this: elmer ~ # hppa2.0-unknown-linux-gnu-g++ -V 2>&1 | egrep -c "CC: Sun" 0 so that can't be it. I guess libtool is to blame somehow, i.e. the 5.6.1 patch. I've committed a patch to 5.6.1 that fixes this: --- GNUmakefile.orig 2010-11-04 18:18:03.000000000 +0100 +++ GNUmakefile 2010-11-04 19:21:29.000000000 +0100 @@ -102,8 +102,6 @@ ifeq ($(UNAME),SunOS) LDLIBS += -lnsl -lsocket M32OR64 = -m$(shell isainfo -b) -endif - ifneq ($(IS_SUN_CC),0) # override flags for CC Sun C++ compiler CXXFLAGS = -DNDEBUG -O -g0 -native -template=no%extdef $(M32OR64) LDFLAGS = @@ -117,6 +115,7 @@ CXXFLAGS += -DCRYPTOPP_INCLUDE_VECTOR_CC endif endif +endif SRCS = $(wildcard *.cpp) ifeq ($(SRCS),) # workaround wildcard function bug in GNU Make 3.77 |