Closed lgoettgens closed 2 months ago
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 84.93%. Comparing base (
22e3908
) to head (a837dca
).
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Hm, alternatively one could put them all into the user facing matrix_space
functions, right?
Hm, alternatively one could put them all into the user facing
matrix_space
functions, right?
Yes, it was basically a coin flip what I did
Will this be merged soon? I have just ripped out my code (related to #1711), so that I could make PR #1732
Just not to lose this observation: it is strange that the ccalls to flint pass the numbers of rows/cols as Int
rather than UInt
. No doubt there is a reason for this...
Can you explain why you think this is strange?
Since the numbers of rows/cols must be non-negative and there is a type for non-negative values, UInt
, it seems strange to use a type Int
which also allows negative values (which should only lead to errors). I have not looked at the Flint sources...
The ccall
arguments must match the signature of the C function.
Ah, so my question was really about the flint UI; in C/C++ I understand one being wary of unsigned
values.
Thanks for merging!
Alternative to and thus closes https://github.com/Nemocas/Nemo.jl/pull/1711.
All matrix space methods now check
r
andc
in the constructor of the type.