feat: add option for custom compare function

[lellimecnar]

It's not always reliable to use mtime, so allowing for a compare function to be passed makes it possible for custom logic to be used. (ie. hashing)

This PR contains a:

• bugfix
• new feature
• code refactor
• test update
• typo fix
• metadata update

Motivation / Use-Case

In our CI environment, the mtime doesn't always match when it probably should, so we'd like to use something else (like file hashing). This is a simple update to expose a compare option that can be used to provide custom logic.

It's been mentioned before (#34 #41 #29)

Breaking Changes

No breaking changes

Additional Info

I started writing some tests, but I'm not sure how best to test this update...

[jsf-clabot]

All committers have signed the CLA.

[codecov]

Codecov Report

Merging #71 into master will decrease coverage by 0.88%.
The diff coverage is 0%.

@@            Coverage Diff             @@
##           master      #71      +/-   ##
==========================================
- Coverage   58.13%   57.25%   -0.89%     
==========================================
  Files           2        2              
  Lines         129      131       +2     
  Branches       20       20              
==========================================
  Hits           75       75              
- Misses         46       48       +2     
  Partials        8        8

Impacted Files	Coverage Δ
src/index.js	57.69% <0%> (-0.91%)	⬇️

---

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 1369c05...0c10211. Read the comment docs.

[codecov]

Codecov Report

Merging #71 into master will decrease coverage by 0.88%.
The diff coverage is 0%.

@@            Coverage Diff             @@
##           master      #71      +/-   ##
==========================================
- Coverage   57.81%   56.92%   -0.89%     
==========================================
  Files           2        2              
  Lines         128      130       +2     
  Branches       20       20              
==========================================
  Hits           74       74              
- Misses         46       48       +2     
  Partials        8        8

Impacted Files	Coverage Δ
src/index.js	57.36% <0%> (-0.91%)	⬇️

---

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 1bdf3e6...66f90c7. Read the comment docs.

[alexander-akait]

/cc @mistic

[m2bright]

👍

[lellimecnar]

We have discovered why the mtime is unreliable in our environment:

By default, some tools (like tar) preserve mtimes that are precise to the second not the millisecond. Since the cache-loader is using === to compare mtimes, this causes some environments (like CI) to always invalidate the cache.

So, I've updated the default compare function to ignore milliseconds, which should solve the issue for these types of environments.

Once this PR is merged, I would be happy to create another PR that exposes a precision option that would allow this to be configured.

[mistic]

@lellimecnar I think removing the miliseconds part should not be the default, what happens if while developing there are 2 different changes to the same file within the same second? I think we should open 2 different PRs, one for simply provide a configurable compare function that receives 2 times and a second one to change the precision of the times passed to the compare function. The default should be the entire time with miliseconds too

What do u think @evilebottnawi ?

[lellimecnar]

I can see that @mistic. I'll make that change, and start the other PR.

[lellimecnar]

@mistic this PR has been updated

[lellimecnar]

@mistic Here is the PR for the precision option: #73

[mistic]

LGTM but I think we should add tests

[mistic]

@lellimecnar also can you solve the conflicts and merge with master please?

[mistic]

Closed as it will be merged through #71 with tests.