This is the mail archive of the gcc-patches@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: RFC: Address some fallout from "fractional modes"


kenner@vlsi1.ultra.nyu.edu (Richard Kenner) writes:

>     This patch is relative to the tree as of Sunday.  I see that a
>     "temporary workaround" was committed to gcc/ada this morning.
>     I am not about to dig through that very large bulk update to 
>     find out what it was
>
> Arno's ChangeLog entry was a little confusing there: the only change
> relevant to this was to use and define the two new fp_* functions.
> (The DIGIT stuff was fixing a bug with GC and VAX floating point
> types.)  They are in misc.c.  They need the obvious change after
> this patch and that's the better way to deal with it for the moment.

So, let me make sure I understand this: I should drop the patches I
already did for ada/.  Instead I should replace GET_MODE_BITSIZE(X)
with GET_MODE_PRECISION(X) and GET_MODE_SIZE(X)*BITS_PER_UNIT with
GET_MODE_BITSIZE(X).  That should be enough.  Correct?

>     (please, consider contributing each individual change as a separate
>     patch, as is done for all other work).  
>
> It has been considered, but isn't practical because of the amount of
> work that's involved.  First, gathering "each individual change"
> together between different files is non-trival.  But if each patch
> is submitted separately, it has to be *tested* separately, which
> will increase the number of test runs each day by over an order of
> magniture.

Oddly enough, everyone else can cope just fine with both of these
requirements.  I do not see why the Ada front end should be different.

zw


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