This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: real.c on unicosmk


Excerpt of message (sent 23 May 2002) by Stephen L Moshier:
> 
> > > pdp11 aborts on any floating constant.
> >
> > My patches should not have changed that.
> 
> On second reading, it seems what broke, by whomever's action, was actually
> the memory values.  They used to be printed in decimal as shown below
> (from gcc-2.95).  The PDP-11 floating point format is similar to a VAX but
> I think the byte order is slightly scrambled relative to VAX format.
> 
> 
> #define ASM_OUTPUT_DOUBLE(FILE,VALUE)  \
>   fprintf (FILE, "\tdouble %.20e\n", (VALUE))
> 
> /* This is how to output an assembler line defining a `float' constant.
> */
> 
> #define ASM_OUTPUT_FLOAT(FILE,VALUE)  \
>   fprintf (FILE, "\tfloat %.12e\n", (VALUE))

I'd have to check the memory order question.  But in any case, what
you've shown has gcc generating float numeric strings and the
assembler sorts it out.  The byte/word order question would matter to
gcc only if it was generating float and double literals with ".word"
and octal or decimal integer bitpatterns.

    paul


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]