Makefile does not account for cross-compiling

[rkeene]

tcl-lmdb/Makefile.in

Lines 152 to 161 in 992c681

	# Workaround for glibc pthread robust mutex support (glibc < 2.12) fix
	ifneq ("$(OS)","Windows_NT")
	UNAME_S := $(shell uname -s)
	ifeq ($(UNAME_S),Linux)
	GLIBCCHECK := $(shell expr `ldd --version | grep ^ldd | sed 's/^.* //g'` \< 2.12)
	ifeq "$(GLIBCCHECK)" "1"
	PKG_CFLAGS += -DMDB_USE_ROBUST=0
	endif
	endif
	endif

This assumes that because one is building ON Linux that one is building FOR Linux, this does not hold true when cross-compiling.

[ray2501]

Yes, it is try to fix glibc pthread robust mutex support (glibc < 2.12) case on and for Linux build for myself. So I do not account for cross-compiling. Any idea for this?

[rkeene]

The LMDB source itself seems to deal with it:

tcl-lmdb/generic/mdb.c

Lines 309 to 328 in 992c681

	#ifdef __GLIBC__
	#define GLIBC_VER ((__GLIBC__ << 16 )| __GLIBC_MINOR__)
	#endif
	/** Some platforms define the EOWNERDEAD error code
	* even though they don't support Robust Mutexes.
	* Compile with -DMDB_USE_ROBUST=0, or use some other
	* mechanism like -DMDB_USE_SYSV_SEM instead of
	* -DMDB_USE_POSIX_MUTEX. (SysV semaphores are
	* also Robust, but some systems don't support them
	* either.)
	*/
	#ifndef MDB_USE_ROBUST
	/* Android currently lacks Robust Mutex support. So does glibc < 2.4. */
	# if defined(MDB_USE_POSIX_MUTEX) && (defined(__ANDROID__) || \
	(defined(__GLIBC__) && GLIBC_VER < 0x020004))
	# define MDB_USE_ROBUST 0
	# else
	# define MDB_USE_ROBUST 1
	# endif
	#endif /* !MDB_USE_ROBUST */

[ray2501]

I add this because it results in build failed in the Ubuntu 10.04 environment (glibc 2.11, pthread function is missing or other things).

However, I do not know how to check this in this moment because I updated my environment to newer Ubuntu.

[rkeene]

Maybe changing the glibc check to be 2.11 instead of 2.4 is sufficient ?

[ray2501]

LMDB check glibc < 2.10 when I write code but current LMDB code base check is correct.

I think remove this workaround or instead of check glibc 2.4 is OK. Do you suggest remove the workaround or instead of check glibc 2.4?

[rkeene]

Compiling it for Android still fails with the robust check failing, so maybe something better needs to be done as part of autoconf. I'll submit a patch.