Re: Regression: "wchar: fix bug when checking for ‘inline’" breaks distcc usage

[Bruno Haible]

Thomas Deutschmann wrote:
> > From 285334ca5ac8f537bc183abd121aa68984e5a515 Mon Sep 17 00:00:00 2001
> > From: Paul Eggert
> > Date: Sun, 1 Jul 2018 18:57:01 -0700
> > Subject: wchar: fix bug when checking for ‘inline’
> 
> (https://git.savannah.gnu.org/gitweb/?p=gnulib.git;a=commit;h=285334ca5ac8f537bc183abd121aa68984e5a515)

This commit has enabled a test that
  - creates a file conftest1.c that defines the main() function
    to call an external function zero(),
  - creates a file conftest2.c that defines the zero() function,
  - compiles both,
  - links both together.

> When using distcc, configure will now fail with
> 
> > configure:21565: checking whether <wchar.h> uses 'inline' correctly
> > configure:21590: x86_64-pc-linux-gnu-gcc -c -O2 -pipe -march=sandybridge 
> > -mtune=sandybridge  conftest1.c >&5
> > configure:21593: $? = 0
> > configure:21612: x86_64-pc-linux-gnu-gcc -c -O2 -pipe -march=sandybridge 
> > -mtune=sandybridge  conftest2.c >&5
> > configure:21615: $? = 0
> > /usr/lib/gcc/x86_64-pc-linux-gnu/8.2.0/../../../../x86_64-pc-linux-gnu/bin/ld:
> >  conftest2.o: in function `main':
> > conftest2.c:(.text.startup+0x0): multiple definition of `main'; 
> > conftest1.o:conftest1.c:(.text.startup+0x0): first defined here
> > /usr/lib/gcc/x86_64-pc-linux-gnu/8.2.0/../../../../x86_64-pc-linux-gnu/bin/ld:
> >  conftest1.o: in function `main':
> > conftest1.c:(.text.startup+0x1): undefined reference to `zero'
> > /usr/lib/gcc/x86_64-pc-linux-gnu/8.2.0/../../../../x86_64-pc-linux-gnu/bin/ld:
> >  conftest2.o: in function `main':
> > conftest2.c:(.text.startup+0x1): undefined reference to `zero'
> > collect2: error: ld returned 1 exit status

Looks like distcc, or the distcc installation and configuration on the
particular user's machine (*), is broken. I won't investigate that.

> > distcc[20786] ERROR: compile (null) on localhost failed

While at it, you can also investigate why distcc passes a NULL string
to fprintf. (It surely does not pass the string "(null)"?) This is not
valid in POSIX.

Bruno

(*) AFAIK, the same version of the same program frequently behaves differently
    on different Gentoo machines. Please correct me if I'm wrong.