Make recursive calls to flatMap stack safe #44

tomp4l · 2019-12-09T14:57:15Z

To make recursive algorithms easier to implement this pushes any
nested callbacks in flatMap onto a 'trampoline' which means only one
can run at any time. This is useful when synchronous futures are
created and used in recursive functions.

Without this patch the following test failure would occur:


    RangeError: Maximum call stack size exceeded

      17 |           } else {
      18 |             data[0] = Caml_option.some(result);
    > 19 |             Belt_List.forEach(Belt_List.reverse(callbacks[0]), (function (cb) {
         |                       ^
      20 |                     return Curry._1(cb, result);
      21 |                   }));
      22 |             callbacks[0] = /* [] */0;

      at forEachU (node_modules/bs-platform/lib/js/belt_List.js:785:18)
      at Object.forEach (node_modules/bs-platform/lib/js/belt_List.js:799:10)
      at src/Future.bs.js:19:23
      at Object._1 (node_modules/bs-platform/lib/js/curry.js:65:12)
      at src/Future.bs.js:42:30
      at Object._1 (node_modules/bs-platform/lib/js/curry.js:65:12)
      at make (src/Future.bs.js:13:9)
      at Object.value (src/Future.bs.js:41:10)
      at loop (__tests__/TestFuture.bs.js:364:54)
      at __tests__/TestFuture.bs.js:368:48

resolves #44

coveralls · 2019-12-09T14:58:36Z

Pull Request Test Coverage Report for Build 77

0 of 0 changed or added relevant lines in 0 files are covered.
1 unchanged line in 1 file lost coverage.
Overall coverage increased (+0.2%) to 98.901%

Files with Coverage Reduction	New Missed Lines	%
src/Future.bs.js	1	98.73%

Totals
Change from base Build 76:	0.2%
Covered Lines:	153
Relevant Lines:	154

💛 - Coveralls

tomp4l · 2019-12-09T15:09:00Z

Not sure why the second build failed, I only pushed a second test which passed.

tomp4l · 2019-12-09T15:26:45Z

Also this would ideally have the interfaces merged as the trampoline should not really be public. See #42

scull7 · 2019-12-10T03:32:33Z

I'm not sure we want to add the cost of a trampoline to every invocation of the Future library. Seems that we should think about it as an "opt-in" feature for cases where it would make sense. Elm has a library which does just that: https://github.com/elm-lang/trampoline/blob/master/src/Trampoline.elm

tomp4l · 2019-12-10T08:23:48Z

The overhead of the trampoline is very low, for non nested calls it just creates a thunk and immediately invokes it and tries to pop an array, otherwise there is just an additional unshift. For anyone doing anything asynchronous this overhead is likely negligible, a test that runs 100k loops takes well under 1s. For anyone doing anything with synchronous futures this fixes a breaking stack overflow for large nesting.

How would you use that trampoline library with futures? If you have a Trampoline(Future('a)) and flatMap it using jump then you would end up with a Trampoline(Future(Trampoline('a))). Vice versa If you have a Future(Trampoline('a)) and try to flatMap a Jump value you will end up with a Future(Trampoline(Future('a)), it doesn't compose. One option would be to add an optional executionContext similar to Scala to every function but seems like a lot more overhead.

If you don't think the benefits of this feature outweigh the costs then fair enough but this will bite users of your library at some point if they are writing recursive functions using futures.

scull7 · 2019-12-10T19:35:59Z

If you don't think the benefits of this feature outweigh the costs then fair enough but this will bite users of your library at some point if they are writing recursive functions using futures.

It's not that I think or believe that the costs will outweigh the benefits, it's the fact that I do not know. Nor do I believe that we can understand the usage of such a low level library to make an all encompassing decision. This is why I made the suggestion that we allow the user of our library to decide for themselves when and if they would like to incur the cost of the additional overhead.

For anyone doing anything asynchronous this overhead is likely negligible, a test that runs 100k loops takes well under 1s

Running a single recursive loop tells us nothing about the user who is utilizing the Future library throughout their code base. We would be making a decision that extra allocation is ok for their use case.

How would you use that trampoline library with futures?

I did not intend to suggest that we utilize the library or it's design pattern within the Future codebase. Merely showing that there are other ways to accomplish the task whereby we allow the user to decide when to incur the cost of the additional allocation.

All of that being said, this really should be implemented as an unrolled while loop to make this debate unnecessary.

http://glat.info/fext/

tomp4l · 2019-12-11T10:23:39Z

I'm definitely not an expert in this so I'm not sure I fully follow but the fastest solution in http://glat.info/fext/#section_setup_details_speed_test looks almost identical to this. The only difference being that it will allow the stack to grow to 5 before trampolining.

function runLoop(_callback) {
  while(true) {
    var callback = _callback;
    Curry._1(callback, /* () */0);
    var match = callbacks.pop();
    if (match !== undefined) {
      _callback = match;
      continue ;
    } else {
      return /* () */0;
    }
  };
}

If you have a solution to make this work currently please let me know but with the current implementation I couldn't find an easy way to make things stack safe. Or do you mean an option to allow opt in during make, i.e. provide an optional flag to use or not use trampoline?

tomp4l · 2019-12-11T21:52:07Z

Anyway this PR isn't really going anywhere right now so I'll open an issue and link this PR. Then you can make a decision on what changes if any can be made.

scull7 · 2019-12-11T21:56:59Z

If you want to make your implementation available under a constructor flag then we bump the major version, I think that’s a good stop-gap for the stack explosion.

…

On Dec 11, 2019, at 1:52 PM, Tom Mottram ***@***.***> wrote: Anyway this PR isn't really going anywhere right now so I'll open an issue and link this PR. Then you can make a decision on what changes if any can be made. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#44?email_source=notifications&email_token=AABZG4GOJAQVRXVDNMSTYKDQYFOIRA5CNFSM4JYKQZX2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEGUWJUA#issuecomment-564749520>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABZG4HEFYEBREZJYRVGFATQYFOIRANCNFSM4JYKQZXQ>.

scull7 · 2019-12-11T21:58:08Z

Just want to make sure we don’t explode someone’s performance expectations.

…

On Dec 11, 2019, at 1:52 PM, Tom Mottram ***@***.***> wrote: Anyway this PR isn't really going anywhere right now so I'll open an issue and link this PR. Then you can make a decision on what changes if any can be made. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#44?email_source=notifications&email_token=AABZG4GOJAQVRXVDNMSTYKDQYFOIRA5CNFSM4JYKQZX2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEGUWJUA#issuecomment-564749520>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABZG4HEFYEBREZJYRVGFATQYFOIRANCNFSM4JYKQZXQ>.

tomp4l · 2019-12-12T00:28:12Z

I pushed my latest changes now, I've kept backwards compatibility here. Let me know what you think of the API. I could also add another option to allow user specified executors other than trampoline

type executor = [|`executor(('a, 'a => unit) => unit)]

Again this would ideally be merged with the rei file to hide the internals.

tomp4l · 2019-12-12T00:35:46Z

Will need to update docs + not sure if trampoline is intuitive enough, maybe `stackSafe?

scull7 · 2019-12-12T01:20:43Z

src/Future.re

@@ -21,21 +51,23 @@ let make = resolver => {
  Future(
    resolve =>
      switch (data^) {
-      | Some(result) => resolve(result)
+      | Some(result) => trampoline(() => resolve(result))


Doesn't this mean the trampoline still runs on every invocation?

Yes my mistake, will fix in next commit.

scull7 · 2019-12-12T01:27:01Z

__tests__/TestFuture.re

+    ->Future.get(r => r |> expect |> toEqual(numberOfLoops) |> finish);
+  });
+
+  testAsync("async recursion is stack safe", finish => {


Make a test using the default executor to ensure it blows the stack and we're not just always using the trampoline

See "value recursion blows the stack with default executor"

scull7 · 2019-12-12T01:30:14Z

__tests__/TestFuture.re

@@ -311,4 +311,38 @@ describe("Future Belt.Result", () => {
        |> finish
      );
  });
-});
+
+  testAsync("value recursion is stack safe", finish => {


Make a deeply nested chain and test that we don't blow the stack there. Something like the following:

next(x) ->Future.flatMap(x => Future.value(x + 1)) ->Future.map(x => x + 1) ->Future.flatMap(x => Future.value(x + 1)->Future.flatMap(x => Future.value(x + 1)))

scull7 · 2019-12-12T01:31:06Z

src/Future.re

@@ -1,17 +1,47 @@
 type getFn('a) = ('a => unit) => unit;

+type executorOptions = [ | `none | `trampoline];


Maybe rename to executorType?

I think trampoline is good. If you run into the stack problem, trampoline is something you'll definitely find right away.

scull7

Looks good, great work

scull7 · 2019-12-12T21:47:54Z

@tomp4l @briangorman Once #42 is merged then this is good to go

tomp4l · 2019-12-14T12:47:20Z

Will need minor tweaks for the addition of the executor

To make recursive algorithms easier to implement this pushes any nested callbacks in flatMap onto a 'trampoline' which means only one can run at any time. This is useful when synchronous futures are created and used in recursive functions.

Use polymorphic variants for ease of use and to support the option of passing a parameterised version in future with user specified executor Keeps backwards compatibility.

tomp4l · 2019-12-17T11:42:45Z

I've updated the rei file, one change I did make was to make the future type abstract as I think it's more useful than exposing the underlying implementation and will make avoiding breaking changes much easier in future. If you disagree I can add the actual implementation to the interface but it will be a breaking change. As you can see the changes otherwise to the rei are minimal with just addition of optional arguments.

briangorman

Thanks!

tomp4l force-pushed the master branch from 7fe9068 to ddfb387 Compare December 9, 2019 15:04

tomp4l mentioned this pull request Dec 11, 2019

Flat map is not stack safe. #45

Closed

scull7 requested changes Dec 12, 2019

View reviewed changes

tomp4l requested a review from scull7 December 12, 2019 21:43

scull7 approved these changes Dec 12, 2019

View reviewed changes

tom added 4 commits December 17, 2019 11:33

Make recursive calls to flatMap stack safe

488a824

To make recursive algorithms easier to implement this pushes any nested callbacks in flatMap onto a 'trampoline' which means only one can run at any time. This is useful when synchronous futures are created and used in recursive functions.

Adds constructor option to opt-in for trampoline

59f03b3

Use polymorphic variants for ease of use and to support the option of passing a parameterised version in future with user specified executor Keeps backwards compatibility.

Address feedback and update README

d824fea

Add to interface file

baf85e9

tomp4l force-pushed the master branch from f3031fa to baf85e9 Compare December 17, 2019 11:39

briangorman requested changes Dec 30, 2019

View reviewed changes

briangorman approved these changes Dec 31, 2019

View reviewed changes

briangorman merged commit 543bbe6 into RationalJS:master Dec 31, 2019

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Make recursive calls to flatMap stack safe #44

Make recursive calls to flatMap stack safe #44

tomp4l commented Dec 9, 2019 •

edited by scull7

coveralls commented Dec 9, 2019 •

edited

tomp4l commented Dec 9, 2019

tomp4l commented Dec 9, 2019

scull7 commented Dec 10, 2019

tomp4l commented Dec 10, 2019

scull7 commented Dec 10, 2019

tomp4l commented Dec 11, 2019

tomp4l commented Dec 11, 2019

scull7 commented Dec 11, 2019 via email

scull7 commented Dec 11, 2019 via email

tomp4l commented Dec 12, 2019

tomp4l commented Dec 12, 2019

scull7 Dec 12, 2019

tomp4l Dec 12, 2019

scull7 Dec 12, 2019

tomp4l Dec 12, 2019

scull7 Dec 12, 2019

tomp4l Dec 12, 2019

scull7 Dec 12, 2019

scull7 Dec 12, 2019

scull7 left a comment

scull7 commented Dec 12, 2019

tomp4l commented Dec 14, 2019

tomp4l commented Dec 17, 2019

briangorman left a comment •

edited

		@@ -1,17 +1,47 @@
		type getFn('a) = ('a => unit) => unit;

		type executorOptions = [ \| `none \| `trampoline];

Make recursive calls to flatMap stack safe #44

Make recursive calls to flatMap stack safe #44

Conversation

tomp4l commented Dec 9, 2019 • edited by scull7

coveralls commented Dec 9, 2019 • edited

Pull Request Test Coverage Report for Build 77

💛 - Coveralls

tomp4l commented Dec 9, 2019

tomp4l commented Dec 9, 2019

scull7 commented Dec 10, 2019

tomp4l commented Dec 10, 2019

scull7 commented Dec 10, 2019

tomp4l commented Dec 11, 2019

tomp4l commented Dec 11, 2019

scull7 commented Dec 11, 2019 via email

scull7 commented Dec 11, 2019 via email

tomp4l commented Dec 12, 2019

tomp4l commented Dec 12, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

scull7 left a comment

Choose a reason for hiding this comment

scull7 commented Dec 12, 2019

tomp4l commented Dec 14, 2019

tomp4l commented Dec 17, 2019

briangorman left a comment • edited

Choose a reason for hiding this comment

tomp4l commented Dec 9, 2019 •

edited by scull7

coveralls commented Dec 9, 2019 •

edited

briangorman left a comment •

edited