[Info-vax] Looking into C-include files on VMS

[Joerg Schilling]

In article <xcydnWIE6b8Z13bXnZ2dnUVZ_u2dnZ2d at giganews.com>,
Richard B. Gilbert <rgilbert88 at comcast.net> wrote:
>Joerg Schilling wrote:

>> If I find a solution for the hangs in "configure" and
>> if I find a solution for the "cc always exits with 0"
>> problem that leads to wrong configure results, a 
>> make/configure controlled compilation seems to be close.
>> 
>> 
>
>Are you using the DEC/Compaq/HP C compiler?  That should NOT exit with 
>"0".  Gnu or other flavors of CC may exit with 0 or almost anything 

I am sorry, I have to correct me. on VMS 7.2 the problem was indeed
the zero exit code from the compiler wrapper, even if the compiler
was given a bad C source.

The "compiler" is a wrapper called "cc", written in C around the 
VMS C-compiler that implements compatibility to standard compiler options.

On EISNER with VMS 8.3, the compiler returns a correct exit code 
but the problem now is the "bash" binary that seems to have problems with
I/O redirection (in the current case, so-called "here documents").

calling "bash somescript"

with "somescript" containing:

cat > x.c << EOF
main()
{
printf("test\n");
}
EOF

creates an empty file "x.c" and a file $HOME/sh869670
that has the content expected for x.c, the small C program from above.

As a result of the empty C-source files, the configure tests cause the 
C-compiler to create warnings about empty source files and a (in this 
case correct) exitcode 0. This is interpreted as "success" :-(

As "bash" also from time to time hangs while exeuting similar code
as mentioned above, this I/O redirection problem currently seems to be 
the only problem for getting "configure" to work.

-- 
EMail:joerg at schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
      js at cs.tu-berlin.de		(uni)  
      joerg.schilling at fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
URL:  http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily