Re: Patch: tfformat.c doesn't compile under hpux 10.20

[Jeffrey A Law <law at cygnus dot com>]

  In message <200009111848.OAA26783@hiauly1.hia.nrc.ca>you write:
  > >   In message <200009111654.MAA26451@hiauly1.hia.nrc.ca>you write:
  > >   > This patch fixes a problem compiling the test program tfformat.c unde
  > r
  > >   > hpux 10.20.  The include of stdlib.h has been removed because under h
  > pux
  > >   > it results in the inclusion of stdio.h.  The definitions in stdio.h
  > >   > conflict with those in iostdio.h.
  > >   > 
  > >   > The header stdlib.h was included for `exit'.  The single use of exit 
  > in
  > >   > the program has been changed to a return.  Tested under hpux and i686
  > >   > linux.
  > > I must have missed the original thread.  What definitions conflict and is
  > > it possible to fix the conflict?
  > 
  > The original thread was "libio testsuite: timeout compiling tFile.cc".  Bun
  > dled
  > with the patches to work around the timeout was a patch to fix the compile
  > problem for tfformat.
  > 
  > This problem was created by the following patch:
  > 
  > 2000-05-23  Zack Weinberg  <zack@wolery.cumb.org>
  > 
  >         * tfformat.c: Include stdlib.h.
  > 
  > Here is a sample of the conflicts:
  > 
  > Executing on host: /xxx/gnu/gcc-2.96/objdir/gcc/xgcc -B/xxx/gnu/gcc-2.96/ob
  > jdir/gcc/ ../../../../libio/testsuite/../tests/tfformat.c  -O3 -I.. -I../..
  > /../../libio/testsuite/..  -DTEST_LIBIO -DTEST_EXACTNESS -L/xxx/gnu/gcc-2.9
  > 6/objdir/hppa1.
  > 1-hp-hpux10.20/threads/libio -L/xxx/gnu/gcc-2.96/objdir/hppa1.1-hp-hpux10.2
  > 0/threads/libio -L/xxx/gnu/gcc-2.96/objdir/hppa1.1-hp-hpux10.20/threads/lib
  > iberty -g -lio -liberty -lm   -threads -o /xxx/gnu/gcc-2.96/objdir/hppa1.1-
  > hp-hpux10.20/libio/testsuite/tfformat    (timeout = 300)
  > In file included from /usr/include/pwd.h:74,^M
  > 		 from /xxx/gnu/gcc-2.96/objdir/gcc/include/stdlib.h:337,^M
  > 		 from ../../../../libio/testsuite/../tests/tfformat.c:11:^M
  > /xxx/gnu/gcc-2.96/objdir/gcc/include/stdio.h:69:33: warning: "_IOFBF" redef
  > ined^M
  > ../../../../libio/iostdio.h:47:1: warning: this is the location of the prev
  > ious definition^M
  > /xxx/gnu/gcc-2.96/objdir/gcc/include/stdio.h:72:33: warning: "_IONBF" redef
  > ined^M
  > ../../../../libio/iostdio.h:49:1: warning: this is the location of the prev
  > ious definition^M
  > /xxx/gnu/gcc-2.96/objdir/gcc/include/stdio.h:76:33: warning: "_IOLBF" redef
  > ined^M
  > ../../../../libio/iostdio.h:48:1: warning: this is the location of the prev
  > ious definition^M
  > 
  > I don't believe that the conflicts can be eliminated because the header
  > iostdio.h is supposed to define "a stdio-like environment".
Whoops.  Got it backwards.  The ones in libio are the originals, the ones
in stdio.h are the redefinitions.

Hmmm.  Send mail to Uli & Benjamin and ask them for advice, they own this
code -- I don't think they're following this list carefully.

jeff