Re: TF type (quad float) not implemented?

[Dylan_S_Cuthbert at hq dot scei dot sony dot co dot jp]

Hmm.. I couldn't find a similar definition to the one for intTI_type_node.
(although I can see why if below is true)

Or maybe within the source is it only referred to as a long double or something
like that?  Maybe that's where things got confusing.

If so, maybe some comments should be thrown around a bit in c-tree.h.  It is all
quite confusing searching out the relationship between machine mode and "type",
the documentation on www.gnu.org leaves a bit to be desired with regards to
this.  I think maybe the documentation needs to be split into an overview and a
reference guide.  Currently it kind of straddles the muddy patch inbetween and
doesn't quite work perfectly as either.  The overview could describe in detail
*how* the compiler works, ie. how all the bits connect together, from machine
modes, to constraints, to builtins, to functions.

So, with the current version of gcc, how do I specify in my machine description
files that I want to be able to use a type of mode TF?

Regards

Dylan Cuthbert





Richard Henderson <rth@cygnus.com> 2000/06/15 11:28:20 PM

宛先: Dylan S Cuthbert/SCEI@SCEI
cc:   gcc@gcc.gnu.org
件名: Re: TF type (quad float) not implemented?

On Thu, Jun 15, 2000 at 07:39:52PM +0900,
Dylan_S_Cuthbert@hq.scei.sony.co.jp wrote:
>
> Whilst trying to implement a TFtype using just the machine description I
> discovered that the TF type is documented but not actually implemented
> within the gcc main source.

Err.. what exactly lead you to this conclusion?  Several targets
make use of TFmode (alpha-osf5, sparc, sparc64).  However, not all
targets want to use TFmode, and so don't.


r~