configure of cairo 1.8.8 ends with: checking whether float word ordering is bigendian... unknown configure: error: Unknown float word ordering. You need to manually preset ax_cv_c_float_words_bigendian=no (or yes) according to your system. workaround (on x86): "export ax_cv_c_float_words_bigendian=yes" indeed fixes things Reproducible: Always Steps to Reproduce: 1. emerge x11-libs/cairo-1.8.8 Actual Results: fails to configure Expected Results: it should indeed configure and compile out of the box * * ERROR: x11-libs/cairo-1.8.8 failed. * Call stack: * ebuild.sh, line 49: Called src_configure * environment, line 2797: Called econf '--enable-xlib' '--disable-gtk-doc' '--enable-directfb' '--enable-xcb' '--enable-svg' '--enable-glitz' '--enable-xlib-xrender' '--disable-test-surfaces' '--enable-pdf' '--enable-png' '--enable-ft' '--enable-ps' * ebuild.sh, line 534: Called die * The specific snippet of code: * die "econf failed" * The die message: * econf failed *
uh, sorry, the workaround on x86 should be: "export ax_cv_c_float_words_bigendian=no"
Start with attaching config.log.
Created attachment 196095 [details] config.log
Something doesn't seem right: you have '-march=nocona' but claim to be on x86. On x86 it works fine. Try compiling that test manually and see what result you'll get. It's simply a specially crafted double - see build/aclocal.float.m4.
Yeah that is probably it! I had (in /etc/make.conf): CHOST="i686-pc-linux-gnu" CFLAGS="-march=nocona -O2 -pipe" changing "-march=nocona" to "-march=core2" fixes things. I want a 32-bit system, so "-march=nocona" was/is probably a nonsense. Thank you! (In reply to comment #4) > Something doesn't seem right: > you have '-march=nocona' but claim to be on x86. > On x86 it works fine. > Try compiling that test manually and see what result you'll get. > It's simply a specially crafted double - see > build/aclocal.float.m4. >
Please reopen the bug, the problem does still exist on arm. My /etc/make.conf: CFLAGS="-O2 -pipe -fomit-frame-pointer -mcpu=cortex-a8 -mfpu=neon -ftree-vectorize -mfloat-abi=softfp -ffast-math -fsingle-precision-constant" CXXFLAGS="${CFLAGS}" CHOST="armv7a-unknown-linux-gnueabi"
@comment 6: did you try with saner CFLAGS ? As the test uses a specifically crafted *double*, it might be that '-fsingle-precision-constant' is breaking that test.