fix: fake all supported timers by default

[SimenB]

Purpose (TL;DR) - mandatory

It's very weird to me that this method is excluded from default mocking. I've read #126 and to me the solution seems to be for users to specify a toFake themselves, not force other people to specify "yes, when I say fake timers I do mean all timers".

Note that this should be considered a breaking change.

Whenever we start supporting some new timer, it should by default not be mocked, but I think when we later make a new major release it should be mocked by default again.

[codecov[bot]]

Codecov Report

Merging #323 into master will decrease coverage by 0.03%. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #323      +/-   ##
==========================================
- Coverage   92.75%   92.71%   -0.04%     
==========================================
  Files           1        1              
  Lines         552      549       -3     
==========================================
- Hits          512      509       -3     
  Misses         40       40              

Flag	Coverage Δ
#unit	92.71% <100.00%> (-0.04%)	:arrow_down:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/fake-timers-src.js	92.71% <100.00%> (-0.04%)	:arrow_down:

---

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 7f8322e...21d726f. Read the comment docs.

[benjamingr]

So here's the thing: promise libraries and other libraries use nextTick and microticks for scheduling so this is a pretty big breaking change.

In principle I agree that we should mock this by default.

[SimenB]

So here's the thing: promise libraries and other libraries use nextTick and microticks for scheduling so this is a pretty big breaking change.

Good point, should probably call it out in the readme.

In principle I agree that we should mock this by default.

👍

[SimenB]

@benjamingr added a section to the readme, feedback appreciated 😀

[benjamingr]

@SimenB I am +1 on this change in principle but I know it will break pretty much every fake-timers install "in the wild" that uses the current API.

[SimenB]

Should work if people use the builtin Promise and async-await, no? If it breaks builtin Promise global I'd consider that a bug in node.

My main issue is that to enable "fake everything" I have to enumerate all timers from the outside, which is not obvious I think. Maybe we could rather add a fakeNextTick explicit option (defaulting to false I guess) instead of piggy-backing toFake?

[benjamingr]

Should work if people use the builtin Promise and async-await, no? If it breaks builtin Promise global I'd consider that a bug in node.

It doesn't break the built in promises but mostly because we don't mock its timings at the moment (we probably should with async_hooks). In practice this means that it's a lot harder than it has to to test promises with fake-timers (I recall a lot of discussions about this + async versions of timers somewhere?)

I don't mind seeing what breaks with mocking nextTick on some large codebases I maintain - though I assume I know the answer is "every place that relies on nextTick executing and doesn't .tick or runMicrotasks before that :]

[SimenB]

As a data point, in a work project where I migrated to use Jest's fake timers now that Jest ships with Lolex it broke node-fetch (and/or integration with nock, not entirely sure, I just added a runMicrotasks call)

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[mroderick]

Stale bot?!

[fatso83]

This makes my head hurt and you guys know best anyhow. Not sure why it has stalled, though?

[fatso83]

B,B,B,B,bump? It's not super problematic to release a major version, as long as you also document how to deal with the fallout. Like using node-fetch as an example case, show what needs to be done to fix the test cases. @SimenB

[mroderick]

This needs a rebase to resolve the conflicts.

@SimenB do you have time to take this across the finish line? If not, would you be ok with someone else doing it?

[fatso83]

I want to finish this (and do #490) and release a new breaking major. Soonish

I don't mind seeing what breaks with mocking nextTick on some large codebases I maintain

There's been 4 years. I guess the number of places that rely on non-native Promises are a lot fewer and further between. Maybe it's not that bad? I think we would just need to add something to the changelog.

[fatso83]

Since I closed #490 and I haven't heard anything in months, I think I'll just pick this up and release a new breaking major.

[fatso83]

Weird how much you can get done in a weekend without kids 😄