Re: R on riscv64 (Re: building fewer packages or reduced functionality on i386)

Hi,

Am 23.08.26 um 11:46 schrieb Richard Lewis:
> Have you also tried to discuss this issue with whichever upstream
> project (glibc? kernel?) defines the NaN propogation behaviour?

Well, I am told (that's why the message says so) this is an _optional_ part of the IEEE standard.

If it's optional and riscv64 (or at least some CPUs, I am also told some have it...) didn't implement it you lost.


And no, in this case glibc/kernel is not involved itself (ttbomk), LO does everything here itself. Talked with LibreOffice

upstream, though, and they have no intention of changing it as current behaviour and expectiations are deep in the codebase.


All that together ended up in the debconf message. :) I could not build on riscv64 (especially since they added a test to test for this

exact issue after the above-mentioned discussion, which we disable in favour of that big fat debconf warning), but people might want to find

some use still (or use writer. etc..)


Regards,

Rene