Improved arithmetic primops

There are a few primops that report arithmetic overflow, which work on Ints and a couple or so for Words. But that's not enough! Let's add more and clean up the existing code (e.g., make overflow-reporting helpers proper primops) on the go. See the wiki page for the proposal.

changed weight to 5

added Ttask Trac import labels

changed milestone to %8.0.1

changed the description

changed the description

Could you clarify (on the wiki) what assembly you get if you just implement this in Haskell (on top of the "unsafe" primops) and what assembly you'd like to see. Elsewhere (e.g. with the shift primops) we have been able to implement "safe" primops on top of the unsafe ones.

changed the description

Replying to [ticket:10962#comment:108368 tibbe]:

Could you clarify (on the wiki) what assembly you get if you just implement this in Haskell (on top of the "unsafe" primops) and what assembly you'd like to see. Elsewhere (e.g. with the shift primops) we have been able to implement "safe" primops on top of the unsafe ones.

https://ghc.haskell.org/trac/ghc/wiki/ImprovedArithmeticPrimops#AsmfromsubWordCvs.asmfromauser-definedoverflow-checkingfunction

mentioned in commit 8160f42b

Sorry, to bring this up so late in the release cycle, but I have soem concerns about both the name and the type signature of this new function.

Before this was added we had:

plusWord2# :: Word# -> Word# -> (# Word#, Word# #)
timesWord2# :: Word# -> Word# -> (# Word#, Word# #)
quotRemWord#" :: Word# -> Word# -> (# Word#, Word# #)

and now we have added:

subWordC# :: Word# -> Word# -> (# Word#, Int# #)

which has a name and a type signature that is wildly out of whack with the others.

Edit : In addition for the other cases, the first element of the result tuple is the carry whereas for this one its the second.

changed milestone to %8.2.1

This bug won't be fixed in 8.0.1; bumping to 8.2.

Trac metadata

Trac field	Value
Differential revisions	- → D1966

changed milestone to %8.4.1

It doesn't seem likely that this will happen for 8.2. Bumping to 8.4.

changed milestone to %8.6.1

This ticket won't be resolved in 8.4; remilestoning for 8.6. Do holler if you are affected by this or would otherwise like to work on it.

changed milestone to %8.8.1

These won't be fixed for 8.6, bumping to 8.8.

changed milestone to %8.10.1

Bumping milestones of low-priority tickets.

added Pnormal label

removed milestone

marked this issue as related to #22665