Re: Floating-point constant folding in Emacs byte compiler

[Paul Eggert]

On 03/26/2018 11:23 AM, Pip Cet wrote:
> if I understand things correctly, large integers in decimal notation 
> are automatically converted to floats, while large hex constants 
> signal an overflow error. I believe the former case applies to the 
> read syntax for byte-compiled functions.

Yes, that's right. This thread got started, though, partly because I'm 
thinking of changing the behavior for plain integers to be more like 
that of other radixes, because the automatic-float conversion causes too 
many problems in practice. This is still work in progress. For a recent 
discussion please see:

https://debbugs.gnu.org/30408

> is this really a problem for any actual code? Good code shouldn't 
> depend on more than the 30 bits Emacs guarantees,

It is a real issue, I'm afraid.(I suppose this means the code isn't 
good....) See these recent fixes in the master branch:

https://git.savannah.gnu.org/cgit/emacs.git/commit/?id=f92c6dd6cd09991c2ab4c0612574064ca061d927
https://git.savannah.gnu.org/cgit/emacs.git/commit/?id=011186279c1041e790b81af72981547d623d9b29
https://git.savannah.gnu.org/cgit/emacs.git/commit/?id=cae9e0fa0358d6323b57a288b5661dc72d36356d

Better treatment of large integers at compile-time will help catch these 
gotchas.