From @nwc10

Created by @nwc10

I've been playing with clang, and found a bug in it which prevents building
Errno​: http​://llvm.org/bugs/show_bug.cgi?id=3915

Specifically, clang -E foo.c works, but clang -E foo.h doesn't.
clang -E - <foo.h works too, which would mean that a cpp wrapper would work...

Now, whilst this is a bug in clang, we, in theory, probe for and work around
around buggy C pre-processors, we don't spot this one, because all the tests
in Configure are for testcpp.c

Ideally, to have trapped this, we'd try for testcpp.c and testcpp.h, and if
either fails, deem that approach bad, and try the next. Problem is, that
"next" looks like this​:

if $ok; then
  : nothing
elif echo 'Maybe "'"$cc"' -E" will work...'; \
  $cc -E <testcpp.c >testcpp.out 2>&1; \
  $contains 'abc.*xyz' testcpp.out >/dev/null 2>&1 ; then
  echo "Yup, it does."
  x_cpp="$cc -E"
  x_minus='';
elif echo 'Nope...maybe "'"$cc"' -E -" will work...'; \
  $cc -E - <testcpp.c >testcpp.out 2>&1; \
  $contains 'abc.*xyz' testcpp.out >/dev/null 2>&1 ; then
  echo "Yup, it does."
  x_cpp="$cc -E"
  x_minus='-';
elif echo 'Nope...maybe "'"$cc"' -P" will work...'; \
  $cc -P <testcpp.c >testcpp.out 2>&1; \
  $contains 'abc.*xyz' testcpp.out >/dev/null 2>&1 ; then
  echo "Yipee, that works!"
  x_cpp="$cc -P"
  x_minus='';
elif echo 'Nope...maybe "'"$cc"' -P -" will work...'; \
  $cc -P - <testcpp.c >testcpp.out 2>&1; \
  $contains 'abc.*xyz' testcpp.out >/dev/null 2>&1 ; then
  echo "At long last!"
  x_cpp="$cc -P"
  x_minus='-';
elif echo 'No such luck, maybe "'$cpp'" will work...'; \
  $cpp <testcpp.c >testcpp.out 2>&1; \
  $contains 'abc.*xyz' testcpp.out >/dev/null 2>&1 ; then
  echo "It works!"
  x_cpp="$cpp"
  x_minus='';
elif echo 'Nixed again...maybe "'$cpp' -" will work...'; \
  $cpp - <testcpp.c >testcpp.out 2>&1; \
  $contains 'abc.*xyz' testcpp.out >/dev/null 2>&1 ; then
  echo "Hooray, it works! I was beginning to wonder."
  x_cpp="$cpp"
  x_minus='-';
elif echo 'Uh-uh. Time to get fancy. Trying a wrapper...'; \
  $wrapper <testcpp.c >testcpp.out 2>&1; \
  $contains 'abc.*xyz' testcpp.out >/dev/null 2>&1 ; then
  x_cpp="$wrapper"
  x_minus=''
  echo "Eureka!"
else
  dflt=''
  rp="No dice. I can't find a C preprocessor. Name one​:"
  . ./myread
  x_cpp="$ans"
  x_minus=''
  $x_cpp <testcpp.c >testcpp.out 2>&1
  if $contains 'abc.*xyz' testcpp.out >/dev/null 2>&1 ; then
  echo "OK, that will do." >&4
  else
echo "Sorry, I can't get that to work. Go find one and rerun Configure." >&4
  exit 1
  fi
fi

Is there a clean way to adapt all of that to loop for both file names?
Is there a better approach?

Nicholas Clark

Flags:
    category=core
    severity=low

Site configuration information for perl 5.11.0:

Configured by nclark at Tue Mar 31 15:39:45 BST 2009.

Summary of my perl5 (revision 5 version 11 subversion 0) configuration:
  Commit id: cedab142c3511603d566403c134158cf036c418d
  Platform:
    osname=linux, osvers=2.6.24-23-generic, archname=i686-linux
    uname='linux vun---ha-ha-ha.cambridge.moo 2.6.24-23-generic #1 smp mon jan 26 00:13:11 utc 2009 i686 gnulinux '
    config_args='-Dusedevel=y -Dcc=/home/nclark/build/llvm/Debug/bin/clang -Dld=ccache /home/nclark/build/llvm/Debug/bin/clang -Ubincompat5005 -Uinstallusrbinperl -Dcf_email=nick@ccl4.org -Dperladmin=nick@ccl4.org -Dinc_version_list=  -Dinc_version_list_init=0 -Doptimize=-Os -Uusethreads -Uuse64bitint -Accflags=-DNO_MATHOMS -Uusemymalloc -Duseperlio -Dprefix=~/Sandpit/snap5.9.x-GitLive-blead-882-gcedab14 -Dinstallman1dir=none -Dinstallman3dir=none -Uuserelocatableinc -de'
    hint=recommended, useposix=true, d_sigaction=define
    useithreads=undef, usemultiplicity=undef
    useperlio=define, d_sfio=undef, uselargefiles=define, usesocks=undef
    use64bitint=undef, use64bitall=undef, uselongdouble=undef
    usemymalloc=n, bincompat5005=undef
  Compiler:
    cc='/home/nclark/build/llvm/Debug/bin/clang', ccflags ='-DNO_MATHOMS -fno-strict-aliasing -pipe -fstack-protector -I/usr/local/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64',
    optimize='-Os',
    cppflags='-DNO_MATHOMS -fno-strict-aliasing -pipe -fstack-protector -I/usr/local/include'
    ccversion='', gccversion='4.2.1 Compatible Clang Compiler', gccosandvers=''
    intsize=4, longsize=4, ptrsize=4, doublesize=8, byteorder=1234
    d_longlong=define, longlongsize=8, d_longdbl=define, longdblsize=12
    ivtype='long', ivsize=4, nvtype='double', nvsize=8, Off_t='off_t', lseeksize=8
    alignbytes=4, prototype=define
  Linker and Libraries:
    ld='ccache /home/nclark/build/llvm/Debug/bin/clang', ldflags =' -fstack-protector -L/usr/local/lib'
    libpth=/usr/local/lib /lib /usr/lib /lib64 /usr/lib64
    libs=-lnsl -ldb -ldl -lm -lcrypt -lutil -lc
    perllibs=-lnsl -ldl -lm -lcrypt -lutil -lc
    libc=/lib/libc-2.7.so, so=so, useshrplib=false, libperl=libperl.a
    gnulibc_version='2.7'
  Dynamic Linking:
    dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=undef, ccdlflags='-Wl,-E'
    cccdlflags='-fPIC', lddlflags='-shared -Os -L/usr/local/lib -fstack-protector'

Locally applied patches:
    PERL_GIT_UNPUSHED_COMMITS    	/* do not remove this line */
    PERL_GIT_UNCOMMITTED_CHANGES	/* do not remove this line */


@INC for perl 5.11.0:
    lib
    /home/nclark/Sandpit/snap5.9.x-GitLive-blead-882-gcedab14/lib/perl5/site_perl/5.11.0/i686-linux
    /home/nclark/Sandpit/snap5.9.x-GitLive-blead-882-gcedab14/lib/perl5/site_perl/5.11.0
    /home/nclark/Sandpit/snap5.9.x-GitLive-blead-882-gcedab14/lib/perl5/5.11.0/i686-linux
    /home/nclark/Sandpit/snap5.9.x-GitLive-blead-882-gcedab14/lib/perl5/5.11.0
    .


Environment for perl 5.11.0:
    HOME=/home/nclark
    LANG (unset)
    LANGUAGE (unset)
    LD_LIBRARY_PATH (unset)
    LOGDIR (unset)
    PATH=/home/nclark/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/sbin:/sbin:/usr/sbin
    PERL_BADLANG (unset)
    SHELL=/bin/bash

From @doughera88

On Tue, 31 Mar 2009, Nicholas Clark wrote​:

I've been playing with clang, and found a bug in it which prevents building
Errno​: http​://llvm.org/bugs/show_bug.cgi?id=3915

Specifically, clang -E foo.c works, but clang -E foo.h doesn't.
clang -E - <foo.h works too, which would mean that a cpp wrapper would work...

Now, whilst this is a bug in clang, we, in theory, probe for and work around
around buggy C pre-processors, we don't spot this one, because all the tests
in Configure are for testcpp.c

It's not the .c vs. .h issue per se. It's that all the tests
are piped from stdin. (This is the probe for $cppstdin, after all.)
The tests all look something like​:

  $cc -E <testcpp.c >testcpp.out 2>&1

The .c vs. .h distinction is irrelevant here.

I gather Errno is trying to do something subtly different, and not just
reading from stdin. Do you know where, specifically, in Errno the problem
is arising, or shall I go spelunking?

--
  Andy Dougherty doughera@​lafayette.edu

The RT System itself - Status changed from 'new' to 'open'

From @nwc10

On Fri, Apr 03, 2009 at 08​:53​:16AM -0400, Andy Dougherty wrote​:

It's not the .c vs. .h issue per se. It's that all the tests
are piped from stdin. (This is the probe for $cppstdin, after all.)
The tests all look something like​:

 $cc \-E \<testcpp\.c >testcpp\.out 2>&1

The .c vs. .h distinction is irrelevant here.

I gather Errno is trying to do something subtly different, and not just
reading from stdin. Do you know where, specifically, in Errno the problem
is arising, or shall I go spelunking?

ext/Errro/Errno_pm.PL, here​:

  # With the -dM option, gcc outputs every #define it finds
  unless(open(FH,"$Config{cc} -E -dM $Config{cppflags} $file |")) {
  warn "Cannot open '$file'";
  return;
  }

And yes, I realised soon after I sent the bug report that the wrapper would
have made no difference.

Nicholas Clark

The clang bug to which Nick linked was fixed in 2009. perl builds and tests fine with clang, including tests for Errno.

I am taking it for the purpose of closing it in 7 days unless someone seriously objects.

@nwc10

Thank you very much.
Jim Keenan

The clang bug to which Nick linked was fixed in 2009. perl builds and tests fine with clang, including tests for Errno.

I am taking it for the purpose of closing it in 7 days unless someone seriously objects.

@nwc10

Thank you very much.
Jim Keenan

Closing as per schedule.