Hello, Weird problem within GCC and I don't know if it is "mine", "lfs" or "upstream".
- the problem: I am successful to compile cmake. AND (everything almost equal) I am not successful to compile cmake:BUILD/cmake-3.11.2/Utilities/cmlibuv/src/unix/timer.c -o uv-src-unix-timer.c.o In file included from /home/jmp/rpmbuild/BUILD/cmake-3.11.2/Source/cmQtAutoGeneratorMocUic.h:9, from /home/jmp/rpmbuild/BUILD/cmake-3.11.2/Source/cmcmd.cxx:10: /home/jmp/rpmbuild/BUILD/cmake-3.11.2/Source/cmQtAutoGenerator.h:58:10: error: 'mutex' in namespace 'std' does not name a type
std::mutex Mutex_;
- The context:
I am trying to qualify LFS as a production system.
One criteria of a production system is to be able to regenerate
itself over 3 generations, packages being build in a
a random order and in a highly parallel sequences.
Previously, this was proved no problem with LFS-8.1 and LFS-8.2,
trying to do the same with SVN-20180612.
Chapter III. Building the LFS System
6. Installing Basic System Software
make a gcc 8.1.0 which CAN compile a working cmake.
This gcc is the first generation, and was created with
the bootstrap GCC (generation 0, chapter 5. Constructing a Temporary
System), the second generation GCC (compiled with first generation GCC)
is able to compile over 200 packages (vim, coreutils, httpd, etc..)
excepted cmake (I guess it is a gcc++ anomaly)
- Debugging
The default is proved to be coming from
/usr/include/c++/8.1.0
there is files present in gcc second generation (example stdc++.h)
missing from the first.
Obviously the second generation ./configure
(building gcc again with all chapter 6 components) is
injecting the "malfunction"
both GCC (first and second generation) make -k check report
OK.
Note: TCL, expect and dejagnu should be part of chapter6, as
currently runtest is the one from /tools when testing
(it should use a runtest within "basic system")
Only g++ seems in trouble.
I noticed BLFS-2018-06-15 is still using/compiling GCC-7.3.0
is this by purpose???
Am I missing something obvious?
Is someone in this list able to duplicate my finding?
could upstream team give us indication what packages detection
could trigger presence of include as stdc++.h by gcc configure?
Hopefully my bug report is meaningful, I can provide
more data if necessary.
Thanks for all suggestions which could come to your mind.
--
A bientôt
===========================================================
Jean-Marc Pigeon E-Mail: j...@safe.ca
SAFE Inc. Phone: (514) 493-4280
Clement, 'a kiss solution' to get rid of SPAM (at last)
Clement' Home base <"http://www.clement.safe.ca";>
===========================================================
smime.p7s
Description: S/MIME Cryptographic Signature
-- http://lists.linuxfromscratch.org/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
