This is a respin of bug 11592, building in folder path ~/Documents/Riccardo/Prova/poky/build causes the build itself to fail while compiling python3 recipe. Moving to a lowercase only $PWD such as ~ made the same build succeed. For some reasons python3 configure script is bailing-out while checking for gcc and detecting an unexisting icc (and then failing because of unknown $*FLAGS). Please add a QA warning. Attaching pastebin link with output https://pastebin.com/6uYjQt0d
11592 was related to packages with uppercase names, which breaks dpkg. I always do builds in a directory with an uppercase letter in, so that isn't the problem. Something else must have changed, resolving works for me. Please file a new bug if you can reproduce on demand from a pristine configuration.
Hello, after doing some deeper research, I discovered that this isn't an upper case issue, but rather an "Do you have *icc* in your path name issue" Most likely this comes from some weird python3 configure issue, I'll debug it when I get a chance. For now to reproduce just use icc as build directory.
Ok, the problem is this one, and I crafted a quick and dirty patch (this works assuming the compiler ends with gcc$ and not gcc-7$ or similar. the stuff passed to --sysroot in $CC contained the *icc* string, adding the flags for icc compiler CC='arm-poky-linux-gnueabi-gcc -march=armv7ve -marm -mfpu=neon-vfpv4 -mfloat-abi=hard -mcpu=cortex-a7 -DNDEBUG -fno-inline --sysroot=/home/test/poky/icc-build-directory/tmp/work/cortexa7hf-neon-vfpv4-poky-linux-gnueabi/python3/3.5.3-r1.0/recipe-sysroot' Ross, do you think is sane to submit something like this to python folks? --- a/configure.ac 2018-04-21 11:10:43.351156044 +0200 +++ b/configure.ac 2018-04-21 11:12:17.329798702 +0200 @@ -1318,11 +1318,11 @@ fi], [AC_MSG_RESULT(no)]) if test "$Py_LTO" = 'true' ; then case $CC in - *clang*) + *clang) # Any changes made here should be reflected in the GCC+Darwin case below LTOFLAGS="-flto" ;; - *gcc*) + *gcc) case $ac_sys_system in Darwin*) LTOFLAGS="-flto" @@ -1382,7 +1382,7 @@ then fi LLVM_PROF_ERR=no case $CC in - *clang*) + *clang) # Any changes made here should be reflected in the GCC+Darwin case below PGO_PROF_GEN_FLAG="-fprofile-instr-generate" PGO_PROF_USE_FLAG="-fprofile-instr-use=code.profclangd" @@ -1397,7 +1397,7 @@ case $CC in fi fi ;; - *gcc*) + *gcc) case $ac_sys_system in Darwin*) PGO_PROF_GEN_FLAG="-fprofile-instr-generate" @@ -1421,7 +1421,7 @@ case $CC in ;; esac ;; - *icc*) + *icc) PGO_PROF_GEN_FLAG="-prof-gen" PGO_PROF_USE_FLAG="-prof-use" LLVM_PROF_MERGER="true" @@ -1459,7 +1459,7 @@ then # Clang also needs -fwrapv case $CC in - *clang*) WRAP="-fwrapv" + *clang) WRAP="-fwrapv" ;; esac @@ -1541,7 +1541,7 @@ yes) # ICC doesn't recognize the option, but only emits a warning ## XXX does it emit an unused result warning and can it be disabled? case "$CC" in - *icc*) + *icc) ac_cv_disable_unused_result_warning=no ;; *) @@ -1796,7 +1796,7 @@ esac # ICC needs -fp-model strict or floats behave badly case "$CC" in -*icc*) +*icc) CFLAGS_NODIST="$CFLAGS_NODIST -fp-model strict" ;; esac @@ -2562,7 +2562,7 @@ then else CCSHARED="-Kpic -belf" fi;; IRIX*/6*) case $CC in - *gcc*) CCSHARED="-shared";; + *gcc) CCSHARED="-shared";; *) CCSHARED="";; esac;; esac @@ -2603,7 +2603,7 @@ then LINKFORSHARED="-Wl,--export-dynamic" fi;; SunOS/5*) case $CC in - *gcc*) + *gcc) if $CC -Xlinker --help 2>&1 | grep export-dynamic >/dev/null then LINKFORSHARED="-Xlinker --export-dynamic" @@ -5233,7 +5233,7 @@ if test "$have_gcc_asm_for_x87" = yes; t # http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46491 # http://gcc.gnu.org/ml/gcc/2010-11/msg00366.html case $CC in - *gcc*) + *gcc) AC_MSG_CHECKING(for gcc ipa-pure-const bug) saved_cflags="$CFLAGS" CFLAGS="-O2"
I opened an upstream bug report https://bugs.python.org/issue33341