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

Add size comparison methods to IterableOps #6950

Merged
merged 1 commit into from
Jul 27, 2018

Conversation

NthPortal
Copy link
Contributor

@NthPortal NthPortal commented Jul 18, 2018

Add IterableOps.sizeCompare(Int), .sizeCompare(Iterable[_]), and
.sizeIs

Addresses this comment on scala/collection-strawman#338

@NthPortal NthPortal requested review from julienrf and Ichoran July 18, 2018 06:14
@scala-jenkins scala-jenkins added this to the 2.13.0-M5 milestone Jul 18, 2018
* is `O(size min _size)` instead of `O(size)`. The method should be overwritten
* if computing `size` is cheap.
*/
def sizeCompare(_size: Int): Int = {
Copy link
Contributor Author

Choose a reason for hiding this comment

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

not super happy with this parameter name, but I wasn't sure what else to call it, as shadowing the size method would be confusing

Copy link
Contributor

Choose a reason for hiding this comment

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

Maybe otherSize or thatSize?

* this.sizeIs > size // this.sizeCompare(size) > 0
* }}}
*/
@inline final def sizeIs: IterableOps.SizeCompareOps = new IterableOps.SizeCompareOps(this)
Copy link
Contributor Author

Choose a reason for hiding this comment

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

not sure if lengthIs should just alias to this, or what. The issue is, then the documentation for all of its ops would say "size" instead of "length"; I'm not sure if that's an issue.

As it is right now, IterableOps.SizeCompareOps and SeqOps.LengthCompareOps are pretty much the same

@NthPortal
Copy link
Contributor Author

Yes, this needs at least a test for compareSizesWith, as well as perhaps some test reorganization. However, I wanted to push the PR up to allow for feedback on the changes as soon as possible

@dwijnand
Copy link
Member

If the results of sizeCompare and compareSizesWith are to be tested with being less/greater than and/or equal to 0, why not have their type be a little LT/EQ/GT sum type instead of an Int?

@NthPortal
Copy link
Contributor Author

@dwijnand because Java (probably) :(

sizeCompare is equivalent to lengthCompare on Seqs, and lengthCompare already returns an Int prior to 2.13, so it's a little late to change that. And it would be pretty inconsistent for just compareSizesWith to introduce a new return type.

Early on, someone decided that Ordering, Ordered, and friends should return Ints, and... sadly, I feel like it's too late to change that, as much as I'd love to.

I am 100% with you, especially since I'm pretty sure returning an Int is what leads to the mistake of trying to subtract numbers to compare them, which isn't generally valid (see #6930, for example)

@dwijnand
Copy link
Member

Let's break some eggs 😄. There is the lengthCompare precedent but I'd say it shouldn't be followed. Perhaps the difference can be lived with, or perhaps it should be dealt by deprecating int-fraternizer lengthCompare?

because Java (probably) :(

Can't the Java user .equals the results? Alternatively:

  • the type could be a java enum rather than a sealed family (is there a precedent of shipping java enums in the standard library?)
  • the type could have the suite of lt/lte/.. methods on it for both Java and Scala users

@NthPortal
Copy link
Contributor Author

oh, I just meant because Java decided to return an int for comparison methods. I didn't mean for Java interop

@Ichoran
Copy link
Contributor

Ichoran commented Jul 18, 2018

We don't have zero-cost abstraction of enums that work in a Scala-friendly way, so I don't think there are good options for returning a trinary that won't slow things down. Since the point of these methods is for speed, typically, slowing things down seems like a bad idea.

(Point is--ints are really fast. If you do anything else, it should be accompanied by a benchmark indicating that the difference isn't an issue. Even if it were a good idea to break with precedent, which it probably isn't.)

@SethTisue
Copy link
Member

sadly, I feel like it's too late to change that, as much as I'd love to

agree

@dwijnand
Copy link
Member

ints are really fast

Is int < faster than reference equality?

@dwijnand
Copy link
Member

Note: my feedback shouldn't in any way be considered a blocker to this change landing. AFAIC this can land as is.

@Ichoran
Copy link
Contributor

Ichoran commented Jul 20, 2018

@dwijnand - It's a smaller factor than I thought. If we have

sealed trait Cmp {}
final object Lt extends Cmp {}
final object Eq extends Cmp {}
final object Gt extends Cmp {}

def compare(a: Int, b: Int) = if (a < b) Lt else if (a > b) Gt else Eq

then in a tight loop I measure it to be less than 5% slower than using java.lang.Integer.compare(a, b).

@dwijnand
Copy link
Member

That's interesting! Thanks for sharing.

* is `O(size min _size)` instead of `O(size)`. The method should be overwritten
* if computing `size` is cheap.
*/
def sizeCompare(_size: Int): Int = {
Copy link
Contributor

Choose a reason for hiding this comment

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

Maybe otherSize or thatSize?

* is `O(this.size min that.size)` instead of `O(this.size + that.size)`.
* The method should be overwritten if computing `size` is cheap.
*/
def compareSizesWith(that: Iterable[_]): Int = {
Copy link
Contributor

Choose a reason for hiding this comment

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

What about using the singular for “size” here: compareSizeWith?

Copy link
Contributor Author

@NthPortal NthPortal Jul 22, 2018

Choose a reason for hiding this comment

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

IANAG(rammarian), but I thought that "sizes" makes more sense, since I am comparing the size of this to the size of that (i.e. two sizes), rather than the size of this to that (which would be one size, and also nonsense). I am happy to be be corrected by someone that plural is not more grammatically correct in this case.

Edit: e.g. "I compared phones with my friend"

Copy link
Contributor Author

Choose a reason for hiding this comment

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

If most others thinks compareSizeWith is more natural or intuitive, I'm happy to change it

Copy link
Member

Choose a reason for hiding this comment

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

Do you feel more comfortable with compareBySizeWith?

Copy link
Contributor

Choose a reason for hiding this comment

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

Neither of the With ones are great because we normally use With to mean we're using a different comparator (e.g. sortWith). compareSizeTo is fine (shorthand for "compare my size to that's size"). Or, as we did with compare, we could simply drop the To--not compareTo, just compare, so not compareSizeTo but just compareSize.

The plural works too (shorthand for "compare our respective sizes"), but since the non-plural version has a sensible meaning I'd go for it instead. Having to remember random s's is annoying.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

honestly, that makes me more inclined to call it sizeCompare. I think it makes more sense grammatically than compareSize (which I still opine ought to use the plural), and is also consistent with the other sizeCompare method. I think that compareSizesWith is sufficiently different from sizeCompare that one would be unlikely to mix them up, but having both sizeCompare and compareSize would quite confusing.

Copy link
Contributor

Choose a reason for hiding this comment

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

Overloading sizeCompare looks good to me.

val seq = Seq(1, 2, 3)
assert(seq.compareSizesWith(Seq(1, 2)) > 0)
assert(seq.compareSizesWith(Seq(1, 2, 3)) == 0)
assert(seq.compareSizesWith(Seq(1, 2, 3, 4)) < 0)
Copy link
Contributor

Choose a reason for hiding this comment

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

I think we should explicitly use List here, and then write another test with a collection implementation that have a knownSize different from -1.

Add IterableOps.sizeCompare(Int), .sizeCompare(Iterable[_]), and
.sizeIs
@NthPortal NthPortal force-pushed the topic/sizeCompare/PR branch from 4323252 to 68161af Compare July 24, 2018 07:38
@NthPortal
Copy link
Contributor Author

ping

Copy link
Member

@dwijnand dwijnand left a comment

Choose a reason for hiding this comment

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

Thanks Nth!

@dwijnand dwijnand merged commit ea9795e into scala:2.13.x Jul 27, 2018
@NthPortal NthPortal deleted the topic/sizeCompare/PR branch July 27, 2018 13:26
@SethTisue SethTisue added the release-notes worth highlighting in next release notes label Jul 28, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
release-notes worth highlighting in next release notes
Projects
None yet
Development

Successfully merging this pull request may close these issues.

6 participants