add Stack.drop

[zapashcanon]

Hi,

This PR adds a val drop : 'a t -> unit function to the Stack module.

It behaves as Stack.pop except that it returns () instead of the topmost element.

This is useful when you're e.g. writing a WebAssembly interpreter. :)

More seriously, Sherlocode shows 118 results when searching for ignore (Stack.pop and there are probably more - IIRC Sherlocode doesn't index all opam packages so I bet the real number is higher.

With the feature freeze, I wasn't sure about when this can be merged. I used @since 5.00, let me know if this should be changed.

[gasche]

It's a useful proposal, I'm Approving, but indeed that looks like post-freeze material.

(I'm not sure whether the ice age will thaw off before or after the 5.00 release, so I just added a post-freeze milestone.)

[zapashcanon]

ping @gasche, I rebased on trunk and I believe it can be merged :)

[gasche]

We require stdlib changes to be approved by two maintainers instead of one.(cc @Octachron, @nojb)

[zapashcanon]

I changed @since 5.0 to @since 5.1.0.

[nojb]

LGTM

I am wondering though if drop : 'a t -> int -> unit may be more generally useful ?

[zapashcanon]

@nojb, fixed !

I am wondering though if drop : 'a t -> int -> unit may be more generally useful ?

From the result I get on sherlocode, people seem to drop a single element most of the time.

[xavierleroy]

From the result I get on sherlocode, people seem to drop a single element most of the time.

Agreed. The current API gives no facilities for pushing N elements at a time or popping N elements at a time, so dropping N elements at a time doesn't make much sense to me.

[nojb]

@nojb, fixed !

I am wondering though if drop : 'a t -> int -> unit may be more generally useful ?

From the result I get on sherlocode, people seem to drop a single element most of the time.

Makes sense, the PR diff looks good and it has two approvals. Merging.