deprecate spones by fredrikekre · Pull Request #25037 · JuliaLang/julia

fredrikekre · 2017-12-12T08:16:28Z

This is a rather useless function.

This is on top of #25030, relevant commit for this PR is a0ada0f
fix #22381

Sacha0 · 2017-12-12T17:32:54Z

NEWS.md

    respectively. The new `spdiagm` implementation now always returns a square matrix ([#23757]).

+  * `spones(A::AbstractSparseArray{T})` has been deprecated in favor of
+    `LinAlg.fillstored!(copy(A), one(T))` ([#TBD]).


The one(T) could simplify to 1 in the vast majority of cases, as fillstored! necessarily converts that 1 to eltype(A)? Suggesting the simplest replacements first is nice :).

Sacha0 · 2017-12-12T17:33:25Z

base/deprecated.jl


+# PR #TBD
+@eval SparseArrays @deprecate spones(A::SparseMatrixCSC{T}) where {T} fillstored!(copy(A), one(T))
+@eval SparseArrays @deprecate spones(A::SparseVector{T}) where {T} fillstored!(copy(A), one(T))


Likewise here, perhaps simplify the suggested replacement as much as possible?

Sacha0 · 2017-12-12T17:35:32Z

base/sparse/sparsematrix.jl

-"""
-spones(S::SparseMatrixCSC{T}) where {T} =
-     SparseMatrixCSC(S.m, S.n, copy(S.colptr), copy(S.rowval), ones(T, S.colptr[end]-1))
+LinAlg.fillstored!(S::SparseMatrixCSC, x) = (fill!(S.nzval, x); S)


fill!(view(S.nzval, 1:(S.colptr[S.n + 1] - 1)), x)?

Sacha0 · 2017-12-12T17:36:06Z

base/sparse/sparse.jl

 export AbstractSparseArray, AbstractSparseMatrix, AbstractSparseVector,
    SparseMatrixCSC, SparseVector, blkdiag, droptol!, dropzeros!, dropzeros,
-    issparse, nonzeros, nzrange, rowvals, sparse, sparsevec, spdiagm, spones,
+    issparse, nonzeros, nzrange, rowvals, sparse, sparsevec, spdiagm, fillstored!,


Why export fillstored! from SparseArrays while LinAlg owns fillstored!?

Sacha0 · 2017-12-12T17:36:41Z

base/sparse/sparsevector.jl


-# Construction of same structure, but with all ones
-spones(x::SparseVector{T}) where {T} = SparseVector(x.n, copy(x.nzind), ones(T, length(x.nzval)))
+LinAlg.fillstored!(x::SparseVector{T}, y) where {T} = (fill!(x.nzval, y); x)


Perhaps nix the unnecessary type parameter?

Sacha0 · 2017-12-12T17:38:07Z

test/sparse/sparse.jl

-@testset "spones" begin
-    @test spones(sparse(2.0I, 5, 5)) == Matrix(I, 5, 5)
+@testset "fillstored!" begin
+    @test SparseArrays.fillstored!(sparse(2.0I, 5, 5), 1) == Matrix(I, 5, 5)


Why the SparseArrays qualification when the definitions in SparseArrays extend the name from LinAlg? (Likewise below :).)

Sacha0 · 2017-12-12T17:38:41Z

Much thanks for clearing this one @fredrikekre! :)

fredrikekre · 2017-12-15T07:41:48Z

Rebased, planning to merge this when ci is done.

rfourquet · 2017-12-15T07:59:04Z

base/deprecated.jl

 # PR #25030
 @eval LinAlg @deprecate fillslots! fillstored! false

+# PR #TBD


Todo! the CI jobs to seem to not have started yet, so it's still time to update this without having CI resources wasted.

Thanks, fixed!

In the future maybe CI will check for #TBD #TODO, [#CATS] etc!

StefanKarpinski · 2017-12-15T08:12:03Z

I may have broken the build right when this was pushed, so probably worth fixing and rebasing.

StefanKarpinski · 2017-12-15T08:14:42Z

Nevermind, I fixed it and pushed it for you.

StefanKarpinski · 2017-12-15T08:15:32Z

I think we did the same thing at the same time...

fredrikekre · 2017-12-15T08:16:16Z

I'm so confused... I only fixed the missing PR number that Rafael pointed out :)

StefanKarpinski · 2017-12-15T08:18:12Z

Ok, I did that plus I rebased on top of a newer master where I fixed the merge mistake.

Sacha0 · 2017-12-15T17:07:26Z

The i686 AV linalg/bunchkaufman failure is identical to that in #25083. Seems something broke AV i686 linalg/bunchkaufman recently but was nonetheless merged? Having a look at the AV queue... (Edit: Seems AV was out of commission for a bit, during which period the AV i686 linalg/bunchkaufman breaking change likely went through innocently.)

Sacha0 · 2017-12-15T17:22:55Z

The AV i686 failure is #25096, fixed by #25103. FreeBSD hit a timeout, seemingly unrelated. Best!

Sacha0

lgtm for the most recent version! Thanks @fredrikekre! :)

fredrikekre added deprecation This change introduces or involves a deprecation sparse Sparse arrays labels Dec 12, 2017

fredrikekre requested a review from Sacha0 December 12, 2017 08:31

fredrikekre force-pushed the fe/spoons branch from e13e4ce to a0ada0f Compare December 12, 2017 10:10

Sacha0 reviewed Dec 12, 2017

View reviewed changes

fredrikekre force-pushed the fe/spoons branch 3 times, most recently from bb8f554 to 9ec0111 Compare December 15, 2017 07:41

rfourquet reviewed Dec 15, 2017

View reviewed changes

StefanKarpinski force-pushed the fe/spoons branch from 9ec0111 to ddd292d Compare December 15, 2017 08:14

fredrikekre force-pushed the fe/spoons branch from ddd292d to c4b83e0 Compare December 15, 2017 08:14

at-deprecate spones(A) fillstored!(copy(A), 1)

462ee64

StefanKarpinski force-pushed the fe/spoons branch from c4b83e0 to 462ee64 Compare December 15, 2017 08:17

Sacha0 approved these changes Dec 15, 2017

View reviewed changes

fredrikekre merged commit 7cfcca2 into JuliaLang:master Dec 15, 2017

fredrikekre deleted the fe/spoons branch December 15, 2017 18:56

fredrikekre added a commit to fredrikekre/julia that referenced this pull request Dec 29, 2017

fix deprecation of spones from JuliaLang#25037

5e2fe57

fredrikekre added a commit that referenced this pull request Dec 31, 2017

fix deprecation of spones from #25037 (#25330)

9018379

cormullion mentioned this pull request Mar 18, 2018

NEWS.md is getting a bit untidy #26508

Closed

Uh oh!

Conversation

fredrikekre commented Dec 12, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Sacha0 Dec 12, 2017

Choose a reason for hiding this comment

Uh oh!

Sacha0 Dec 12, 2017

Choose a reason for hiding this comment

Uh oh!

Sacha0 Dec 12, 2017

Choose a reason for hiding this comment

Uh oh!

Sacha0 Dec 12, 2017

Choose a reason for hiding this comment

Uh oh!

Sacha0 Dec 12, 2017

Choose a reason for hiding this comment

Uh oh!

Sacha0 Dec 12, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Sacha0 commented Dec 12, 2017

Uh oh!

fredrikekre commented Dec 15, 2017

Uh oh!

rfourquet Dec 15, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

fredrikekre Dec 15, 2017

Choose a reason for hiding this comment

Uh oh!

rfourquet Dec 15, 2017

Choose a reason for hiding this comment

Uh oh!

StefanKarpinski commented Dec 15, 2017

Uh oh!

StefanKarpinski commented Dec 15, 2017

Uh oh!

StefanKarpinski commented Dec 15, 2017

Uh oh!

fredrikekre commented Dec 15, 2017

Uh oh!

StefanKarpinski commented Dec 15, 2017

Uh oh!

Sacha0 commented Dec 15, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Sacha0 commented Dec 15, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Sacha0 left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

fredrikekre commented Dec 12, 2017 •

edited

Loading

Sacha0 Dec 12, 2017 •

edited

Loading

rfourquet Dec 15, 2017 •

edited

Loading

Sacha0 commented Dec 15, 2017 •

edited

Loading

Sacha0 commented Dec 15, 2017 •

edited

Loading