SIBR: (eqv? 0.0 -0.0) behaviour

[okuoku]

at https://github.com/okuoku/yunibase/commit/8bcb02f125b0ae80618493b1432622652905b576

Many Scheme implementations do not distinguish -0.0 and 0.0 with eqv?.

--     ERR: DIGAMMA
--     OK : GUILE3
--     OK : SAGITTARIUS
--     OK : CHIBI_SCHEME
--     OK : CHICKEN5_CSI
--     ERR: ICYC
--     OK : FOMENT
--     ERR: GOSH
--     OK : KAWA
--     OK : CHEZ_SCHEME
--     OK : GUILE
--     ERR: IRON_SCHEME
--     OK : RACKET
--     ERR: BIGLOO
--     OK : GSI
--     OK : MIT_SCHEME
--     ERR: S7
--     ERR: SCM
--     ERR: STKLOS

• DIGAMMA, ICYC (Cyclone), BIGLOO, S7 and GOSH (Gauche) -- do not distinguish
• SCM and IRON_SCHEME -- it cannot read -0.0 properly

[okuoku]

Reserved: SIBR0013

https://github.com/okuoku/yuni/commit/cd7638cc2c6c5d7174a9ae2ebff3d874ca80a4db

[okuoku]

ref: https://github.com/shirok/Gauche/commit/ca136f159390556a2b34dabf279ced88ab6edc89 Gauche adopted R7RS

[leppie]

From memory, at some point I did try to implement -0.0 for IronScheme, I think there were .NET compat errors (possibly not being able to get eqv? to work). From my point of view it was better to not support it than have broken behavior.

[lassik]

https://doc.scheme.org/surveys/zero/ has a section on negative zero.

More generaly, it would be great to integrate the data from SIBR into Scheme Surveys.

[okuoku]

(Now it's at docs : https://docs.scheme.org/surveys/zero/ )

[okuoku]

Done: https://github.com/okuoku/yuni/commit/c1e06a88dd4d5a700deb0db8f211ec03c61824ff?short_path=5d8e68c#diff-5d8e68ce8bdfde66d54edbbc7788a4e5d2ef5c6f1ccd44fa927c44bab4d251a6