On 2019-08-27 22:16, Anders Magnusson wrote:
Den 2019-08-27 kl. 15:09, skrev Paul Koning:Last problem I had was with Python IIRC, mostly because things won't compile cleanly and require lot of manual intervention (missing headers, missing #defines etc...). Don't know whether it's fixed or not.A program that depends on a specific rounding rule is probably broken in other ways.On Aug 27, 2019, at 8:56 AM, Anders Magnusson <ragge%ludd.ltu.se@localhost> wrote: Morning Johnny, ... IEEE Floating point:+ Compatibility with rest of the world. Simple to compile programs. Possible to use otherwise unusable programs. - Require update of toolchain. Will not support IEEE rounding (would be too slow then).Can you give an example of a program that only works if the float is IEEE, but doesn't mind how rounding is done? I'm having a hard time conceiving of a useful and reasonably reliable program for which this would be true.The point is that we get unneccessary and time-consuming bugs where we can avoid it with quite simple fixes.
Python have all kind of test suites that really expect various INF and NaN to behave like IEEE expects. I wonder if this one will be solved by faking it. And if you just accept that it won't pass the test suites, I think it can already be made to work (if we just get a working C compiler back).
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt%softjar.se@localhost || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol