On Sun, Jul 10, 2016 at 12:52 AM, Isaiah Norton <[email protected]> wrote: > (also a work-around for a symbol hygiene issue: > https://github.com/JuliaLang/julia/issues/15085)
I don't think any of these requires a specific form. > > On Sun, Jul 10, 2016 at 12:46 AM, Isaiah Norton <[email protected]> > wrote: >> >> I believe the second form may be vestigial at this point, and agree that >> it is inconsistent. The usage is non-existent in base (except for one test), >> but there is at least one issue for which the current work-around requires >> the second form: https://github.com/JuliaLang/julia/issues/14208 >> >> On Fri, Jul 8, 2016 at 4:52 AM, Hans-Peter <[email protected]> wrote: >>> >>> Acc. to the manual [1] either `Mod.@mac` or `@Mod.mac` can be used. What >>> is the reason for the second form? Is it necessary to make me (other >>> people?) think which form should be used preferably? Maybe there is a reason >>> (didn't find something in issues/maillist) but with my current knowledge I'd >>> propose to 'push' the first form and deprecate the second. >>> >>> Both forms are not used often (in Julia source and my v0.4 packages). The >>> first (Mod.@mac) form is more frequent. >>> >>> ~~~ >>> find ~/.julia/v0.4/ -name "*.jl" | xargs ack " \w+\.@\w+" >>> find ~/.julia/v0.4/ -name "*.jl" | xargs ack " @\w+\." >>> ~~~ >>> >>> Sometimes both forms are being used, e.g.: >>> >>> ... v0.4/DataArrays/src/DataArrays.jl >>> 78: Base.@deprecate removeNA dropna >>> ... v0.4/DataArrays/test/reducedim.jl >>> 12: Base.Test.@test DataArrays._any(bits, i, j) == v >>> >>> vs. >>> >>> ... v0.4/Calculus/src/Calculus.jl >>> 62: @Base.deprecate integrate(f,a,b) quadgk(f,a,b)[1] >>> ... v0.4/Distributions/test/truncate.jl >>> 70: @Base.Test.test_approx_eq_eps(logpdf(d, x), lp, sqrt(eps())) >>> >>> >>> [1] >>> http://docs.julialang.org/en/latest/manual/modules/#namespace-miscellanea >> >> >
