[predef][check] linux clang triggers BOOST_MAC_OS

Hi! I'm using clang + libc++ on Linux and the Predef "check" magic activates "BOOST_MAC_OS". That means I get: Performing configuration checks - symlinks supported : yes warn: Unable to construct ./info_as_objcpp warn: Unable to construct ./info_as_objc - BOOST_OS_MACOS : yes when running the predef test suite. Command chain: cd libs/predef/test rm -rf ../../../bin.v2/libs/predef/ ../../../bin.v2/check/ ../../../bin.v2/project-cache.jam b2 -a --verbose-test 2>&1 -d+2 --debug-configuration | tee verbose.log verbose.log is attached as zip file. And in fact the macos test are run: clang-linux.compile.c.without-pth BOOST_ROOT/src/devel/boost/bin.v2/libs/predef/test/macos_endian.test/clang- linux-3.7.0/debug/macos_endian.o which is only supposed to be run on macos. The test even passes because it internally checks for "__APPLE__" being defined, which should probably be removed, as the check should be done by the build system. Compiler: clang++ -v clang version 3.7.0 (trunk 227683) Target: x86_64-unknown-linux-gnu Thread model: posix That is current clang trunk. Though I'm already running with -d+2 --debug-configuration I get no output from the configuration test itself. Any ideas on how to debug this further? Yours, Jürgen -- * Dipl.-Math. Jürgen Hunold ! * voice: ++49 4257 300 ! Fährstraße 1 * fax : ++49 4257 300 ! 31609 Balge/Sebbenhausen * jhunold@gmx.eu ! Germany

On Sat, Jan 31, 2015 at 10:13 AM, Jürgen Hunold
Hi!
I'm using clang + libc++ on Linux and the Predef "check" magic activates "BOOST_MAC_OS". That means I get:
Performing configuration checks
- symlinks supported : yes warn: Unable to construct ./info_as_objcpp warn: Unable to construct ./info_as_objc - BOOST_OS_MACOS : yes
when running the predef test suite.
Strange. And in fact the macos test are run:
clang-linux.compile.c.without-pth BOOST_ROOT/src/devel/boost/bin.v2/libs/predef/test/macos_endian.test/clang- linux-3.7.0/debug/macos_endian.o
which is only supposed to be run on macos. The test even passes because it internally checks for "__APPLE__" being defined, which should probably be removed, as the check should be done by the build system.
I'd rather keep that.. As I don't like relying on a particular build system for any code to compile. Though I'm already running with -d+2 --debug-configuration I get no output
from the configuration test itself. Any ideas on how to debug this further?
The output from the configuration build parts is output to boost-root/bin.v2/config.log. Also you could cd to libs/predef/test. And run "b2 -a --verbose-test". And check the output to see what is detected for BOOST_OS_*. -- -- Rene Rivera -- Grafik - Don't Assume Anything -- Robot Dreams - http://robot-dreams.net -- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail

On Sat, Jan 31, 2015 at 10:19 AM, Rene Rivera
On Sat, Jan 31, 2015 at 10:13 AM, Jürgen Hunold
wrote: Hi!
I'm using clang + libc++ on Linux and the Predef "check" magic activates "BOOST_MAC_OS". That means I get:
Performing configuration checks
- symlinks supported : yes warn: Unable to construct ./info_as_objcpp warn: Unable to construct ./info_as_objc - BOOST_OS_MACOS : yes
when running the predef test suite.
Strange.
I've reproduced the problem myself now. I'll look into it. No need for you to dig any more. -- -- Rene Rivera -- Grafik - Don't Assume Anything -- Robot Dreams - http://robot-dreams.net -- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail

On Sat, Jan 31, 2015 at 10:23 AM, Rene Rivera
On Sat, Jan 31, 2015 at 10:19 AM, Rene Rivera
wrote: On Sat, Jan 31, 2015 at 10:13 AM, Jürgen Hunold
wrote: Hi!
I'm using clang + libc++ on Linux and the Predef "check" magic activates "BOOST_MAC_OS". That means I get:
Performing configuration checks
- symlinks supported : yes warn: Unable to construct ./info_as_objcpp warn: Unable to construct ./info_as_objc - BOOST_OS_MACOS : yes
when running the predef test suite.
Strange.
I've reproduced the problem myself now. I'll look into it. No need for you to dig any more.
It's fixed now. It seems that I lost some code while I was cleaning up before checking things in. Specifically the code that set the expressions on running the check program. -- -- Rene Rivera -- Grafik - Don't Assume Anything -- Robot Dreams - http://robot-dreams.net -- rrivera/acm.org (msn) - grafikrobot/aim,yahoo,skype,efnet,gmail
participants (2)
-
Jürgen Hunold
-
Rene Rivera