How could we cross compile ck?

[zetalog]

It looks the "configure" is a hand-writing script not a script generated by autoconf, and contains the following code:

cat << EOF > .1.c
#include <stdio.h>
int main(void) {
#if defined(_WIN32)
#if defined(__MINGW64__)
	puts("mingw64");
	return (0);
#elif defined(__MINGW32__) && (__MINGW32_MAJOR_VERSION >= 3)
	puts("mingw32");
	return (0);
#else
	return (1);
#endif /* __MINGW32__ && __MINGW32_MAJOR_VERSION >= 3 */
#elif defined(__clang__) && (__clang_major__ >= 3)
	puts("clang");
	return (0);
#elif defined(__SUNPRO_C) && (__SUNPRO_C >= 0x5110)
	puts("suncc");
	return (0);
#elif defined(__GNUC__) && (__GNUC__ >= 4)
	puts("gcc");
	return (0);
#else
	return (1);
#endif
}
EOF

$CC -o .1 .1.c
COMPILER=`./.1 2> /dev/null`
r=$?
rm -f .1.c .1

The block is trying to execute a target binary via the following line and definitely will fail:

COMPILER=`./.1 2> /dev/null`

It thus looks like that the "ck" doesn't support cross compile.
Am I right? If not, please help to tell how could we cross compile ck.

Thanks in advance.

[zetalog]

Other than this, it looks ck can perfectly support cross compilation.
I confirmed that by forcing a COMPILER name here to bypass the detection code.

So please help to fix the issue in the right way in ck upstream.
Thanks in advance.

[zetalog]

Related PR:
#194

[zetalog]

Below is the patch to bypass COMPILER detection:

Index: ck/configure
===================================================================
--- ck.orig/configure
+++ ck/configure
@@ -657,7 +669,8 @@ if test ! -x "${CC}"; then
 fi
 assert "$CC" "not found"

-cat << EOF > .1.c
+if test "x$COMPILER" = "x"; then
+       cat << EOF > .1.c
 #include <stdio.h>
 int main(void) {
 #if defined(_WIN32)
@@ -685,10 +698,11 @@ int main(void) {
 }
 EOF

-$CC -o .1 .1.c
-COMPILER=`./.1 2> /dev/null`
-r=$?
-rm -f .1.c .1
+       $CC -o .1 .1.c
+       COMPILER=`./.1 2> /dev/null`
+       r=$?
+       rm -f .1.c .1
+fi

 if test "$r" -ne 0; then
        assert "" "update compiler"

This may not be required by upstream if the issue can be properly handled.
But if you want, I can also push it to the upstream.

[sbahra]

Thanks for patch!

[jscheid-ventana]

@sbahra, I don't see #193 (comment) making it into the master branch. #194 did not include it.