On 2022.09.10 13:56, David Haller wrote:
Hello,
On Sat, 10 Sep 2022, Jack wrote:
>I now get this error trying to emerge two different packages:
libofx-0.10.7
>and gnupg (both 2.2.39 and 2.3.6). It might also be the same
problem for a
>few bugs on b.g.o found by searching on "cannot create exectuables."
>
>The relevant lines from build.log are
>
>checking for x86_64-pc-linux-gnu-gcc... x86_64-pc-linux-gnu-gcc
>checking whether the C compiler works... no
>configure: error: in
>`/var/tmp/portage/dev-libs/libofx-0.10.7/work/libofx-0.10.7':
>configure: error: C compiler cannot create executables
>See `config.log' for more details
>
>and from config.log:
>
>configure:2941: x86_64-pc-linux-gnu-gcc -V >&5
>x86_64-pc-linux-gnu-gcc: error: unrecognized command-line option '-V'
>x86_64-pc-linux-gnu-gcc: fatal error: no input files
>compilation terminated.
>configure:2952: $? = 1
>configure:2941: x86_64-pc-linux-gnu-gcc -qversion >&5
>x86_64-pc-linux-gnu-gcc: error: unrecognized command-line option
'-qversion';
>did you mean '--version'?
>x86_64-pc-linux-gnu-gcc: fatal error: no input files
>compilation terminated.
>configure:2952: $? = 1
These tests are normal fails with gcc, they are version checks for
other compilers.
>configure:2972: checking whether the C compiler works
>configure:2994: x86_64-pc-linux-gnu-gcc -march=native -O2 -pipe -og
-ggdb
>-Wl,-O1 -Wl,--as-needed conftest.c >&5
>configure:2998: $? = 0
>configure:3036: result: no
>configure: failed program was:
[..boilerplate..]
>configure:3041: error: in
>`/var/tmp/portage/dev-libs/libofx-0.10.7/work/libofx-0.10.7':
>configure:3043: error: C compiler cannot create executables
>See `config.log' for more details
>
>The thing I find curious is that it appears to me that the output of
the test
>compile is a file called "g" which I don't recall ever seeing, and
so I
>wonder if the problem is that something has changed with gcc
defaults and
>configure does not yet recognize that change. I also don't know the
>significance of the two "fatal error: no input files".
>
>The fact that this happens with two unrelated packages suggests that
it's
>not specific to either of them, but something in my system or
configuration.
And it's a standard autoconf macro, namely AC_PROG_CC that results in
the error and the stuff before that is also standard. And as no
autoreconf is called, autotools versions should not matter.
>Any thoughts or suggestions?
I use gcc 11.3.0 here as well, and have no problem. Check for the
variables CC, CFLAGS, CPPFLAGS, LDFLAGS and LIBS in
/var/tmp/portage/dev-libs/libofx-0.10.7/temp/environment.
Somehow, that '-og' must have crept in there.
David - thank you, thank you, thank you.
No wonder my recent debugging wasn't doing what I expected. That stray
"-og" was supposed to be "-Og" (upper instead of lower case) and it is
in /etc/portage/env/debug.conf, which is referenced in
/etc/portage/package.env for those packages I wish to compile with
various debugging related compiler settings, as well as not deleting
the build dir, even on success.
It shows what a fresh pair of eyes can do. When looking at the various
error related files, I really just saw -Og, and when you pointed that
out, my first thought was why any debugging setting would possible
cause this type of failure. Of course in hindsight, it now all makes
sense.
Thanks again to the list for all sorts of ongoing assistance.
Jack
