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.

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

Restrict arguments to Number in comparisons between scalars and EndPoint #143

Open
wants to merge 2 commits into
base: master
Choose a base branch
from

Conversation

pabloferz
Copy link
Contributor

Looking at some packages invalidations, I found some introduced by this one. As far as I can see, much of these methods aren't supposed to work for things other than Numbers so I think these all should be harmless.

@pabloferz
Copy link
Contributor Author

I think I spoke too soon, I see that Intervals are expected, at least from the tests, to work with things like Dates. Maybe these really are "irreducible" invalidations, I'll have to think about it a bit more.

Copy link
Collaborator

@omus omus left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Currently the only restriction the element type of Interval and AnchoredInterval is that the eltype must be comparable with isless. We may be able to get away with using promotion to convert a scalar (x) into a closed-interval (Interval(x, x)) which could work around the method invalidations by not having to define some of these methods. I suspect there will be a performance hit in doing this though.

src/endpoint.jl Outdated
Comment on lines 148 to 149
Base.:(==)(a::Number, b::Endpoint) = a == b.endpoint && isclosed(b)
Base.:(==)(a::Endpoint, b::Number) = b == a
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm pretty sure we can update this to be this safely (tests will show if that is the case):

Suggested change
Base.:(==)(a::Number, b::Endpoint) = a == b.endpoint && isclosed(b)
Base.:(==)(a::Endpoint, b::Number) = b == a
Base.:(==)(a::T, b::Endpoint{T}) where T = a == b.endpoint && isclosed(b)
Base.:(==)(a::Endpoint{T}, b::T) where T = b == a

I'm not sure these are actually used so we may be able to just drop them.

@codecov
Copy link

codecov bot commented Oct 1, 2020

Codecov Report

Merging #143 into master will not change coverage.
The diff coverage is 100.00%.

Impacted file tree graph

@@           Coverage Diff           @@
##           master     #143   +/-   ##
=======================================
  Coverage   75.76%   75.76%           
=======================================
  Files          11       11           
  Lines         491      491           
=======================================
  Hits          372      372           
  Misses        119      119           
Impacted Files Coverage Δ
src/endpoint.jl 94.59% <100.00%> (ø)

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c37cbb7...ca64ff6. Read the comment docs.

@pabloferz
Copy link
Contributor Author

pabloferz commented Oct 1, 2020

From code coverage, it looks like some of the new definitions are not used in the tests and we might get away with removing them. For others, it might be better to add code tests paths.

I'm not sure how does this look for someone more familiar with direct use cases of the package (I have only use it through other packages that have it as dependency).

@omus
Copy link
Collaborator

omus commented Oct 1, 2020

I'm looking into alternative approaches. I think we may be able to fix Base code such the + and - invalidations no longer occur (see #144). As for converting to a scalar I think we can define an alternative function for this behaviour.

@pabloferz
Copy link
Contributor Author

pabloferz commented Oct 1, 2020

That's great. I guess I will remove the arithmetic code changes. Yes, a custom scalar conversion function sounds sensible, but is probably better done in another PR.

@pabloferz pabloferz changed the title Restrict some arguments to Number to reduce invalidations Restrict arguments to Number in comparisons between scalars and EndPoint Oct 1, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants