chop! no longer produces empty coefficients #479
Conversation
Test case:
using ApproxFun
x,y = Fun(identity, Chebyshev() * Chebyshev())
ApproxFun.norm(0*x)
ApproxFun.norm(zero(x))
|
Without this patch, norm(zero(x)) gives the following error in ApproxFunBase/src/Multivariate/ProductFun.jl:177: ERROR: LoadError: BoundsError: attempt to access 0×0 Matrix{Float64} at index [1:1, 1:1] Fixing ProductFun to handle empty coefficients just leads to a similar error later on. So it seems this package generally assumes that the coefficients vector is not empty. Thus I modified chop! to never return an empty coefficients vector. |
Test case (requires also JuliaApproximation/ApproxFunBase.jl#479): using ApproxFun xdom = Chebyshev(-1..1) ydom = Chebyshev(-1..1) domain = xdom * ydom x,y = Fun(identity, domain) Dx = Derivative(Chebyshev()^2, [1,0]) Dy = Derivative(Chebyshev()^2, [0,1]) N(u, v) = [ 2*u - x; 3*v + y ] u0 = one(x) * one(y) v0 = one(x) * one(y) u, v = newton(N, [u0, v0]) Bug report: JuliaApproximation#887
|
Thanks! ideally the empty |
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #479 +/- ##
===========================================
- Coverage 71.22% 26.38% -44.84%
===========================================
Files 80 81 +1
Lines 8331 8443 +112
===========================================
- Hits 5934 2228 -3706
- Misses 2397 6215 +3818 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Test case:
using ApproxFun
x,y = Fun(identity, Chebyshev() * Chebyshev())
ApproxFun.norm(0*x)
ApproxFun.norm(zero(x))
I didn't create a unit test because I couldn't find a way to reproduce with ApproxFunBase only.