Fixes for non-Int based lengths #37741

dlfivefifty · 2020-09-25T09:08:33Z

This call requires a dodgy overload Int(::Infinity) = ∞ which is causing unnecessary latency. I can't see a reason why it was needed in the first place.

dlfivefifty · 2020-09-25T10:56:04Z

Alternatively it could be a call to Base.to_shape but as the tests are passing I assume that's unnecessary?

KristofferC · 2020-09-25T11:10:08Z

The Int comes from 7711aef#diff-c8c9f86af939c6f462b38975999f448cR50 cc @timholy

timholy · 2020-09-25T14:53:49Z

Pretty old commit though; we might have needed it in 2016 but not now? I would just test stuff. Make sure you measure invalidations with some package that defines a new Integer subtype and see whether it increases badly when you eliminate the Int.

dlfivefifty · 2020-09-25T15:01:37Z

Can you do what you propose with something like:

julia> x = 1:BigInt(10)
1:10

julia> size(x) |> typeof
Tuple{BigInt}

julia> SubArray(x,(1:2,)) |> size |> typeof # Tuple{BigInt} with this change
Tuple{Int64}

pabloferz · 2020-09-25T15:08:43Z

I'm afraid this more recent one might also be relevant for JuliaArrays/InfiniteArrays.jl#44

dlfivefifty · 2020-09-25T15:18:02Z

Ouch, that one looks a disaster....

I should probably add a lightweight ∞-array implementation in the tests (a la the OffsetArray tests) to avoid these types of changes killing the package in future versions of Julia...

timholy · 2020-09-25T16:10:03Z

But it's also pretty important to avoid invalidation. Do we really need to support arrays that don't have Int dimensions? Is that really an array or is it a DiscreteFunction?

dlfivefifty · 2020-09-25T16:24:32Z

We do if you want ApproxFun to continue to work. InfiniteArrays.jl is working quite well right now so suddenly insisting arrays have Int sizes, just for a faster concatenation implementation is a significant breaking change for very little reward.

In any case, adding ::Int everywhere is a clear case of "code smell". I don't really understand why it's not just mapreduce(length,+,...)

timholy · 2020-09-25T16:28:58Z

It has nothing to do with a faster concatenation implementation, though that would be a side effect. See https://julialang.org/blog/2020/08/invalidations/.

Not saying we can't change this, but you need to understand the reasons first and then we can talk about what to do about it.

dlfivefifty · 2020-09-25T19:33:34Z

Yes I understand that. I'll try to come up with an alternative, I'm thinking something like:

_vcat_similar(T, V) = similar(V[1], T, mapreduce(length, +, V))
_typed_vcat(::Type{T}, V::AbstractVecOrTuple{AbstractVector}) where T =
   _typed_vcat!(_vcat_similar(T, V), V)
function _typed_vcat!(a, V)
    pos = 1
    … # current code with all the Int(...)::Int calls
end

This would allow me to overload based on the type of a.

timholy · 2020-09-26T12:05:52Z

I will bet this PR is probably OK for invalidation, but you should still check (until we have #37193 it requires a manual check). Maybe try the subject of #36459? It might be useful to temporarily back out #37088, though I bet that fixed problems in the cat pipeline (#37741 (comment)) moreso than this subarray one.

timholy · 2020-09-26T12:15:02Z

Or actually, probably easier:

julia> using MethodAnalysis

julia> cd(joinpath(dirname(dirname(pathof(MethodAnalysis))), "demos"))

julia> include("abstract.jl")

julia> include("abstract_tests.jl")
Test Failed at /home/tim/.julia/dev/MethodAnalysis/demos/abstract_tests.jl:58
  Expression: meancounts < 32
   Evaluated: 36.07617360496014 < 32
ERROR: LoadError: There was an error during testing
in expression starting at /home/tim/.julia/dev/MethodAnalysis/demos/abstract_tests.jl:58

Looks like we've already regressed substantially, but you'll find out if this PR makes it that much worse.

dlfivefifty · 2020-09-26T13:39:25Z

It throws this:

julia> include("abstract_tests.jl")
Test Failed at /Users/solver/.julia/packages/MethodAnalysis/VLLrU/demos/abstract_tests.jl:53
  Expression: length(badcounts) < 1000
   Evaluated: 1071 < 1000
ERROR: LoadError: There was an error during testing
in expression starting at /Users/solver/.julia/packages/MethodAnalysis/VLLrU/demos/abstract_tests.jl:53

dlfivefifty · 2020-09-26T14:05:57Z

Sorry, that was on the last tagged version of MethodAnalysis.jl. All tests pass on MethodAnalysis#master, apart from

@test function_returns(ispow2, Bool)

I don't see why that would be connected to this though.

But I think it's probably best if I make all necessary changes for InfiniteArrays.jl in a single PR (there's some cases in printing where UnitRange{Int}(...) is called which causes problems for both infinite arrays and BigInt sized arrays)

timholy · 2020-09-26T16:36:36Z

If you stay under the bounds in those tests then I think any changes you make should be fine, at least for the purpose of Julia itself. (Some packages might experience more invalidation, but in a sense that's their own fault for having poorly-inferred code---although in the case of cat it's actually pretty hard to avoid some form of inference trouble.)

dlfivefifty · 2020-11-18T16:00:09Z

InfiniteArrays tests now pass, so I believe this PR is ready to be merged.

@timholy I had to revert some of #37080 because it triggered a breaking change independent of InfiniteArrays.jl: it caused [1; 1:BigInt(10)] to error out, see #37741. Hopefully this doesn't cause too many issues with invalidation.

timholy · 2020-11-30T11:19:02Z

@dlfivefifty, note test failures.

dlfivefifty · 2020-11-30T11:32:43Z

@timholy It doesn't look like the test failures are caused by this PR, unless I'm not understanding.

KristofferC · 2020-11-30T11:38:19Z

Try rebase on master where CI should be improved.

timholy · 2020-12-17T11:49:03Z

Something like

julia> mis = methodinstances(Base.print_matrix)
6-element Vector{Core.MethodInstance}:
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Int64})
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Matrix{Int64})
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Int64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Matrix{Int64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)

julia> mi = mis[end]  # pick one with abstract types in the signature
MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)

julia> mi.backedges
4-element Vector{Any}:
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Int64})
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Matrix{Int64})
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Core.MethodInstance})
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Any})

julia> using AbstractTrees

julia> print_tree(mi)
MethodInstance for Base.print_matrix(::IOContext{TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
├─ MethodInstance for Base.print_matrix(::IOContext{TTY}, ::Vector{Int64})
│  └─ MethodInstance for Base.print_array(::IOContext{TTY}, ::Vector{Int64})
│     └─ MethodInstance for show(::IOContext{TTY}, ::MIME{Symbol("text/plain")}, ::Vector{Int64})
├─ MethodInstance for Base.print_matrix(::IOContext{TTY}, ::Matrix{Int64})
│  └─ MethodInstance for Base.print_array(::IOContext{TTY}, ::Matrix{Int64})
│     └─ MethodInstance for show(::IOContext{TTY}, ::MIME{Symbol("text/plain")}, ::Matrix{Int64})
├─ MethodInstance for Base.print_matrix(::IOContext{TTY}, ::Vector{MethodInstance})
│  └─ MethodInstance for Base.print_array(::IOContext{TTY}, ::Vector{MethodInstance})
│     └─ MethodInstance for show(::IOContext{TTY}, ::MIME{Symbol("text/plain")}, ::Vector{MethodInstance})
└─ MethodInstance for Base.print_matrix(::IOContext{TTY}, ::Vector{Any})
   └─ MethodInstance for Base.print_array(::IOContext{TTY}, ::Vector{Any})
      └─ MethodInstance for show(::IOContext{TTY}, ::MIME{Symbol("text/plain")}, ::Vector{Any})

but I'm more interested in the backedges of _print_matrix (which doesn't exist on master).

dlfivefifty · 2020-12-18T10:11:26Z

Below is what I get. Should I try again with Base.invokelast?

julia> mis = methodinstances(Base.print_matrix)
6-element Vector{Core.MethodInstance}:
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Int64})
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Matrix{Int64})
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Int64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Matrix{Int64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)

julia> mi = mis[end]  # pick one with abstract types in the signature
MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)

julia> mi.backedges
ERROR: UndefRefError: access to undefined reference
Stacktrace:
 [1] getproperty(x::Core.MethodInstance, f::Symbol)
   @ Base ./Base.jl:33
 [2] top-level scope
   @ REPL[7]:1

julia> print_tree(mi)
MethodInstance for Base.print_matrix(::IOContext{TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)

julia> mis = methodinstances(Base._print_matrix)
4-element Vector{Core.MethodInstance}:
 MethodInstance for Base._print_matrix(::IOContext{Base.TTY}, ::Vector{Int64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64, ::UnitRange{Int64}, ::UnitRange{Int64})
 MethodInstance for Base._print_matrix(::IOContext{Base.TTY}, ::Matrix{Int64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64, ::UnitRange{Int64}, ::UnitRange{Int64})
 MethodInstance for Base._print_matrix(::IOContext{Base.TTY}, ::Vector{Core.MethodInstance}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64, ::UnitRange{Int64}, ::UnitRange{Int64})
 MethodInstance for Base._print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64, ::UnitRange{Int64}, ::UnitRange{Int64})

julia> mi = mis[end]  # pick one with abstract types in the signature
MethodInstance for Base._print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64, ::UnitRange{Int64}, ::UnitRange{Int64})

julia> mi.backedges
4-element Vector{Any}:
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Int64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Matrix{Int64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Core.MethodInstance}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Any}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)

julia> print_tree(mi)
MethodInstance for Base._print_matrix(::IOContext{TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64, ::UnitRange{Int64}, ::UnitRange{Int64})
├─ MethodInstance for Base.print_matrix(::IOContext{TTY}, ::Vector{Int64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
│  └─ MethodInstance for Base.print_matrix(::IOContext{TTY}, ::Vector{Int64})
│     └─ MethodInstance for Base.print_array(::IOContext{TTY}, ::Vector{Int64})
│        └─ MethodInstance for show(::IOContext{TTY}, ::MIME{Symbol("text/plain")}, ::Vector{Int64})
├─ MethodInstance for Base.print_matrix(::IOContext{TTY}, ::Matrix{Int64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
│  └─ MethodInstance for Base.print_matrix(::IOContext{TTY}, ::Matrix{Int64})
│     └─ MethodInstance for Base.print_array(::IOContext{TTY}, ::Matrix{Int64})
│        └─ MethodInstance for show(::IOContext{TTY}, ::MIME{Symbol("text/plain")}, ::Matrix{Int64})
├─ MethodInstance for Base.print_matrix(::IOContext{TTY}, ::Vector{MethodInstance}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
│  └─ MethodInstance for Base.print_matrix(::IOContext{TTY}, ::Vector{MethodInstance})
│     └─ MethodInstance for Base.print_array(::IOContext{TTY}, ::Vector{MethodInstance})
│        └─ MethodInstance for show(::IOContext{TTY}, ::MIME{Symbol("text/plain")}, ::Vector{MethodInstance})
└─ MethodInstance for Base.print_matrix(::IOContext{TTY}, ::Vector{Any}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
   └─ MethodInstance for Base.print_matrix(::IOContext{TTY}, ::Vector{Any})
      └─ MethodInstance for Base.print_array(::IOContext{TTY}, ::Vector{Any})
         └─ MethodInstance for show(::IOContext{TTY}, ::MIME{Symbol("text/plain")}, ::Vector{Any})

KristofferC · 2020-12-27T09:43:24Z

What's the status here?

timholy · 2020-12-27T14:48:26Z

Oh, sorry I didn't notice your response, @dlfivefifty.

So, as feared the backedges do go all the way back to show. I'd recommend invokelatest; thoughts, @KristofferC? There's a bit of a runtime penalty (~50ns), but to me that seems nowhere near the problem of invalidation given that

julia> io, a = IOBuffer(), [1,2]
(IOBuffer(data=UInt8[...], readable=true, writable=true, seekable=true, append=false, size=0, maxsize=Inf, ptr=1, mark=-1), [1, 2])

julia> @btime show($io, MIME("text/plain"), $a)
  559.135 μs (303 allocations: 35.20 KiB)

So the invokelatest is 1 part in 10^4.

dlfivefifty · 2020-12-27T17:17:08Z

Why should runtime matter for printing?

timholy · 2020-12-27T19:08:07Z

If someone is using it to dump data to a file? If you're printing 10^5 arrays you'd probably notice. But I doubt that is very common.

My main point was that we have to think about runtime performance, but having thought about it I don't think it's a problem in this case.

dlfivefifty · 2020-12-27T19:15:56Z

I’m sure they can what 50ns...

timholy · 2020-12-27T19:25:12Z

Yes, my point too.

dlfivefifty · 2020-12-28T11:35:52Z

I changed it to Base.invokelatest and to my untrained eye it looks better:

julia> mis = methodinstances(Base.print_matrix)
10-element Vector{Core.MethodInstance}:
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Int64})
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Matrix{Int64})
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Matrix{Float64})
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Core.MethodInstance})
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Int64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Matrix{Int64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Matrix{Float64}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::Vector{Core.MethodInstance}, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)
 MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)

julia> mi = mis[end]  # pick one with abstract types in the signature
MethodInstance for Base.print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)

julia> mi.backedges
ERROR: UndefRefError: access to undefined reference
Stacktrace:
 [1] getproperty(x::Core.MethodInstance, f::Symbol)
   @ Base ./Base.jl:33
 [2] top-level scope
   @ REPL[36]:1

julia> print_tree(mi)
MethodInstance for Base.print_matrix(::IOContext{TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64)

julia> mis = methodinstances(Base._print_matrix)
1-element Vector{Core.MethodInstance}:
 MethodInstance for Base._print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64, ::UnitRange{Int64}, ::UnitRange{Int64})

julia> mi = mis[end]  # pick one with abstract types in the signature
MethodInstance for Base._print_matrix(::IOContext{Base.TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64, ::UnitRange{Int64}, ::UnitRange{Int64})

julia> mi.backedges
ERROR: UndefRefError: access to undefined reference
Stacktrace:
 [1] getproperty(x::Core.MethodInstance, f::Symbol)
   @ Base ./Base.jl:33
 [2] top-level scope
   @ REPL[40]:1

julia> print_tree(mi)
MethodInstance for Base._print_matrix(::IOContext{TTY}, ::AbstractVecOrMat{T} where T, ::String, ::String, ::String, ::String, ::String, ::String, ::Int64, ::Int64, ::UnitRange{Int64}, ::UnitRange{Int64})

Is this also a good solution for concatenation? Probably run time penalty matters there.

dlfivefifty · 2020-12-30T12:57:44Z

The linux32 test failure seems unrelated:

ERROR: LoadError: UndefVarError: rr not defined
Stacktrace:
 [1] getproperty
   @ ./Base.jl:26 [inlined]
 [2] rr
   @ /buildworker/worker/tester_linux32/build/rr_capture.jl:24 [inlined]
 [3] (::var"#1#7")(dir::String)
   @ Main /buildworker/worker/tester_linux32/build/rr_capture.jl:25
 [4] mktempdir(fn::var"#1#7", parent::String; prefix::String)
   @ Base.Filesystem ./file.jl:729
 [5] mktempdir(fn::Function, parent::String)
   @ Base.Filesystem ./file.jl:727
 [6] top-level scope
   @ /buildworker/worker/tester_linux32/build/rr_capture.jl:17
in expression starting at /buildworker/worker/tester_linux32/build/rr_capture.jl:17

timholy · 2021-01-04T19:44:23Z

Thanks for getting this in, @dlfivefifty!

(cherry picked from commit e2f4073)

KristofferC · 2021-03-31T16:46:30Z

This almost looks like it is adding a "secret" extendible API with the oneto and unitrange methods. It doesn't feel sustainable that people "hack in" their own undocumented extensions into the language.

dlfivefifty · 2021-03-31T17:05:48Z

I can add documentation

(cherry picked from commit e2f4073)

Remove Int call

476a59b

dlfivefifty changed the title ~~Remove Int call~~ Remove Int call in size(::SubArray) Sep 25, 2020

Add tests

1d8f1fb

_typed_vcat!

433fb9e

Support infinite arrays in printing

a73d161

timholy mentioned this pull request Sep 29, 2020

Fix some string-related invalidation risks #37799

Merged

Merge branch 'master' into patch-14

8bf23e2

dlfivefifty mentioned this pull request Nov 18, 2020

Concat broken with BigInt ranges #38480

Closed

Revert some of JuliaLang#37080 to fix JuliaLang#37741

e1159be

dlfivefifty changed the title ~~Remove Int call in size(::SubArray)~~ Fixes for non-Int based lengths Nov 18, 2020

Merge remote-tracking branch 'JuliaLang/master' into patch-14

7179d5d

KristofferC mentioned this pull request Dec 19, 2020

WIP: Backports 1.6-beta #38949

Merged

26 tasks

Merge branch 'master' into patch-14

3051d7e

Update arrayshow.jl

69ba1d7

timholy approved these changes Dec 29, 2020

View reviewed changes

Merge branch 'master' into patch-14

a6ddb9e

timholy merged commit e2f4073 into JuliaLang:master Jan 4, 2021

KristofferC pushed a commit that referenced this pull request Jan 5, 2021

Fixes for non-Int based lengths (#37741)

b4f2c66

(cherry picked from commit e2f4073)

KristofferC removed the backport 1.6 Change should be backported to release-1.6 label Jan 6, 2021

timholy mentioned this pull request Jan 12, 2021

Improve (no)specialization in print_matrix #39194

Merged

mkitti mentioned this pull request Jan 14, 2021

Export oneto rather than implement range(stop) #39242

Closed

staticfloat pushed a commit that referenced this pull request Jan 15, 2021

Fixes for non-Int based lengths (#37741)

d29b695

(cherry picked from commit e2f4073)

dlfivefifty mentioned this pull request Feb 15, 2021

UnitRange{Int} -> unitrange #39668

Merged

dkarrasch mentioned this pull request Feb 16, 2021

Make length type match index type in sparse vectors #39645

Merged

KristofferC mentioned this pull request Mar 31, 2021

What to do about the new oneto and unitrange methods #40284

Open

ElOceanografo pushed a commit to ElOceanografo/julia that referenced this pull request May 4, 2021

Fixes for non-Int based lengths (JuliaLang#37741)

1333a3b

dlfivefifty mentioned this pull request Jul 9, 2021

Julia v1.7: Multiplication broken with BigInt ranges #41517

Closed

staticfloat pushed a commit that referenced this pull request Dec 23, 2022

Fixes for non-Int based lengths (#37741)

0abe143

(cherry picked from commit e2f4073)

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fixes for non-Int based lengths #37741

Fixes for non-Int based lengths #37741

dlfivefifty commented Sep 25, 2020

dlfivefifty commented Sep 25, 2020

KristofferC commented Sep 25, 2020

timholy commented Sep 25, 2020

dlfivefifty commented Sep 25, 2020

pabloferz commented Sep 25, 2020 •

edited

Loading

dlfivefifty commented Sep 25, 2020

timholy commented Sep 25, 2020

dlfivefifty commented Sep 25, 2020

timholy commented Sep 25, 2020

dlfivefifty commented Sep 25, 2020

timholy commented Sep 26, 2020

timholy commented Sep 26, 2020

dlfivefifty commented Sep 26, 2020

dlfivefifty commented Sep 26, 2020

timholy commented Sep 26, 2020

dlfivefifty commented Nov 18, 2020

timholy commented Nov 30, 2020

dlfivefifty commented Nov 30, 2020

KristofferC commented Nov 30, 2020 •

edited

Loading

timholy commented Dec 17, 2020 •

edited

Loading

dlfivefifty commented Dec 18, 2020 •

edited

Loading

KristofferC commented Dec 27, 2020

timholy commented Dec 27, 2020

dlfivefifty commented Dec 27, 2020

timholy commented Dec 27, 2020

dlfivefifty commented Dec 27, 2020

timholy commented Dec 27, 2020

dlfivefifty commented Dec 28, 2020

dlfivefifty commented Dec 30, 2020

timholy commented Jan 4, 2021

KristofferC commented Mar 31, 2021

dlfivefifty commented Mar 31, 2021

Fixes for non-Int based lengths #37741

Fixes for non-Int based lengths #37741

Conversation

dlfivefifty commented Sep 25, 2020

dlfivefifty commented Sep 25, 2020

KristofferC commented Sep 25, 2020

timholy commented Sep 25, 2020

dlfivefifty commented Sep 25, 2020

pabloferz commented Sep 25, 2020 • edited Loading

dlfivefifty commented Sep 25, 2020

timholy commented Sep 25, 2020

dlfivefifty commented Sep 25, 2020

timholy commented Sep 25, 2020

dlfivefifty commented Sep 25, 2020

timholy commented Sep 26, 2020

timholy commented Sep 26, 2020

dlfivefifty commented Sep 26, 2020

dlfivefifty commented Sep 26, 2020

timholy commented Sep 26, 2020

dlfivefifty commented Nov 18, 2020

timholy commented Nov 30, 2020

dlfivefifty commented Nov 30, 2020

KristofferC commented Nov 30, 2020 • edited Loading

timholy commented Dec 17, 2020 • edited Loading

dlfivefifty commented Dec 18, 2020 • edited Loading

KristofferC commented Dec 27, 2020

timholy commented Dec 27, 2020

dlfivefifty commented Dec 27, 2020

timholy commented Dec 27, 2020

dlfivefifty commented Dec 27, 2020

timholy commented Dec 27, 2020

dlfivefifty commented Dec 28, 2020

dlfivefifty commented Dec 30, 2020

timholy commented Jan 4, 2021

KristofferC commented Mar 31, 2021

dlfivefifty commented Mar 31, 2021

pabloferz commented Sep 25, 2020 •

edited

Loading

KristofferC commented Nov 30, 2020 •

edited

Loading

timholy commented Dec 17, 2020 •

edited

Loading

dlfivefifty commented Dec 18, 2020 •

edited

Loading