Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

isa_tfunc not working around broken subtyping in the way I hoped it would #27078

Closed
jrevels opened this issue May 11, 2018 · 0 comments · Fixed by #27736
Closed

isa_tfunc not working around broken subtyping in the way I hoped it would #27078

jrevels opened this issue May 11, 2018 · 0 comments · Fixed by #27736

Comments

@jrevels
Copy link
Member

jrevels commented May 11, 2018 

julia> bar(T::Type{S}) where {S} = isa(T, UnionAll) ? bar(T.body) : T
bar (generic function with 1 method)

julia> V = Vector{Vector{T}} where T
Array{Array{T,1},1} where T

julia> bar(V)

ERROR: type DataType has no field body
Stacktrace:
 [1] getproperty at ./sysimg.jl:15 [inlined]
 [2] bar(::Type{Array{Array{T,1},1} where T}) at ./REPL[1]:1 (repeats 2 times)
 [3] top-level scope

If I remove the Type{S} where S:

julia> bar(T) = isa(T, UnionAll) ? bar(T.body) : T
bar (generic function with 1 method)

julia> V = Vector{Vector{T}} where T
Array{Array{T,1},1} where T

julia> bar(V)
Array{Array{T,1},1}

Title might not be quite correct, but the sadness this bug causes me is very real
This was referenced May 17, 2018
Closed
Closed
jrevels added a commit that referenced this issue Jun 22, 2018
@jrevels
add test for #27078
05cc1e2
@jrevels jrevels mentioned this issue Jun 22, 2018
Merged
jrevels added a commit that referenced this issue Jun 25, 2018
@jrevels
add test for #27078
0399394
jrevels added a commit that referenced this issue Jun 25, 2018
@jrevels
add test for #27078
396ab28
jrevels added a commit that referenced this issue Jul 2, 2018
@jrevels
add test for #27078
de1ce0d
jrevels added a commit that referenced this issue Jul 2, 2018
@jrevels
add test for #27078
a5dd37a
jrevels added a commit that referenced this issue Jul 6, 2018
@jrevels
add test for #27078
f7c470e
jrevels added a commit that referenced this issue Jul 6, 2018
@jrevels
add test for #27078
cdf6987
jrevels added a commit that referenced this issue Jul 7, 2018
@jrevels
Avoid broken kind type handling when compiling isa (#27736)
3ac28ee 
* small workaround checks against incorrect subtyping for kind types for isa_tfunc

* add test for #27078

* fix pretty crazy bug where Type{T}s were inferred as Ts

* work around broken subtyping in emit_isa codegen optimization

* a test that was checking for exact/optimal inference result is now broken

Justification for allowing this test to remain broken for now:
- benchmarking the expression (including downstream toy calculations on the output, e.g. broadcast sin) using BenchmarkTools reveals no actual performance difference
- inference returning an optimal result before was probably reliant on the broken subtyping behavior; correctness >>> performance
- inference is still returning a fairly tightly bounded, correct Union type
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Avoid broken kind type handling when compiling isa JuliaLang/julia
1 participant
@jrevels