I've pushed the fix to the "mob" branch, commit b776bfaa. As the commit message says, I was able to confirm the value for AARM64, but only indirectly for RISCV.
On Thu, Dec 12, 2024 at 1:02 AM avih <avih...@yahoo.com> wrote: > > The re-definition of LDBL_MAX_EXP in aarch64/riscv with the same value is > indeed a bug, and your suggested patch does seem correct to me. > > Preferably you should first test it with aarch64 and riscv gcc compilers as > described at the comment in this section, to confirm this value in both. > > Then you can push to the mob branch yourself, as described here: > https://repo.or.cz/tinycc.git > > If you can't test it, then I still think it's a useful fix, but in such case > do write at the commit message that this was not tested with the respective > compilers. > > avih > > > On Wednesday, December 11, 2024 at 11:56:32 PM GMT+2, Keith Thompson > <keith.s.thomp...@gmail.com> wrote: > > > > > > In include/float.h, there are three sections defining the LDBL_* > macros for different architectures. > > The section for aach64 and riscv is missing a definition for LDBL_MAX_10_EXP. > It includes two matching definitions for LDBL_MAX_EXP. > > Here's a proposed patch (untested): > > diff --git include/float.h include/float.h > index 24b7410b..4caa4236 100644 > --- include/float.h > +++ include/float.h > @@ -53,7 +53,7 @@ > #define LDBL_MIN_10_EXP (-4931) > #define LDBL_MAX_EXP 16384 > #define LDBL_MAX 1.18973149535723176508575932662800702e+4932L > -#define LDBL_MAX_EXP 16384 > +#define LDBL_MAX_10_EXP 4932 > #define DECIMAL_DIG 36 > > #else > > _______________________________________________ > Tinycc-devel mailing list > Tinycc-devel@nongnu.org > https://lists.nongnu.org/mailman/listinfo/tinycc-devel _______________________________________________ Tinycc-devel mailing list Tinycc-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/tinycc-devel
