overhaul to allow for different labels #8
Conversation
simeonschaub
force-pushed
the
overhaul
branch
from
0a696be to
0860aa2
Compare
simeonschaub
force-pushed
the
overhaul
branch
from
0860aa2 to
6b88418
Compare
Codecov Report
@@ Coverage Diff @@
## master #8 +/- ##
=====================================
Coverage 100% 100%
=====================================
Files 1 1
Lines 26 29 +3
=====================================
+ Hits 26 29 +3
Continue to review full report at Codecov.
|
| end | ||
|
|
||
| @dynamo function skipargcheck(x...) | ||
| struct ElideCheck{label} | ||
| ElideCheck(label::Symbol) = new{label}() |
There was a problem hiding this comment.
Any particular reason, why label is not a field, but a type parameter?
There was a problem hiding this comment.
I need to do some benchmarks, whether there is actually any difference
There was a problem hiding this comment.
Just for my understanding. You are talking about compile time performance, not run time right? Because as I understand it at run time no ElideCheck objects exist anymore.
There was a problem hiding this comment.
Ok, now that I've looked into this, a type parameter is definitely the way to do this. A @dynamo cannot access any fields of ElideCheck, so the only way to access the label would be to insert a statement into the IR, which wouldn't even work in this case. This makes sense, if you think about it, because otherwise IRTools would have to recompile f every time, since label is not a compile time constant.
julia> using ArgCheck, OptionalArgChecks
[ Info: Precompiling OptionalArgChecks [dfbeeb84-381a-44da-9ec9-a723abf299c7]
julia> f(x) = @argcheck x > 0
f (generic function with 1 method)
julia> @skipargcheck f(-1)
julia> 👍 |
|
I registered an ArgCheck version that has markers: |
Co-Authored-By: Jan Weidner <[email protected]>
simeonschaub
force-pushed
the
overhaul
branch
from
992b99a to
2b87444
Compare
|
Great! @simeonschaub what do you think about tagging a release? |
|
Already done: JuliaRegistries/General#9359. Still waiting for 0.1 to be registered though:grin: |
Ref: #7, jw3126/ArgCheck.jl#32