allow passing a callback to useFetch's run to modify init

[phryneas]

Heyo :)

This adresses #75.

Instead of using a parameter to be spread to init, I chose for a callback function.
This has two reasons

• if run is just passed as an onClick handler, a React SyntheticEvent would have been passed and possibly spread on to init. Detecting all kinds of events would be a pain, but typeof x === "function" is a simple test that it is NOT an event.
• this way, existsing properties can be modified (like adding a new header) without introducing an additional dependency to something like deep-merge

I also split up the method definition between deferFnand promiseFn because the handling of 2 parameters vs 3 parameters was already confusing enough (sometimes props is ctrl, sometimes ctrl is ctrl).
I think this improves readability without impacting code size too much.

Also, I reduced boolean logic at two places to better handle the undefined case of defer and json.

Of, and of course there's a new unit test and a bit more README :)

Looking forward to your feedback. 🙋

[phryneas]

Okay, now eslint should be happy, too.

[phryneas]

Okay, I've tried the failing ./node_modules/.bin/lerna run --scope '*-example' test -- --passWithNoTests --watchAll=false locally and it's working. I have no idea what is failing there, could you take a look?

[phryneas]

Seems like the tests are generally breaking with react 16.9 and something yarn ci does causes the newest react version to be installed.
So nothing associated with this PR :/

[phryneas]

I opened #77 to adress the test issue.

As I'll be gone for vacation tomorrow, I can only do any changes you request today during the day - otherwise feel free to just push any modifications to the top of my branch. (Due to #77, I fear this might take a little longer unti you get to this :/ )

[ghengeveld]

I'm very happy with where you're going with this. I'm in doubt about the callback though. I like the flexibility it provides but I still would prefer to have the option to just pass an object. We can detect a SyntheticEvent using one of the attributes they have, such as preventDefault.

My preference would be:

• if function, use it to determine init
• if object and not SyntheticEvent, spread it after init
• else do nothing with init

[phryneas]

Sounds like a plan! I'll do that as soon as I'm back. Am 10. August 2019 13:41:34 MESZ schrieb Gert Hengeveld <notifications@github.com>:

…

I'm very happy with where you're going with this. I'm in doubt about the callback though. I like the flexibility it provides but I still would prefer to have the option to just pass an object. We can detect a SyntheticEvent using one of the [attributes](https://reactjs.org/docs/events.html#overview) they have, such as `preventDefault`. My preference would be: - if function, use it to determine init - if object and not SyntheticEvent, spread it after init - else do nothing with init

-- Sent from my Android phone with K-9 Mail. Please excuse my brevity.

[phryneas]

Okay, that should do it for the functionality. I also pushed a commit that gives run in the context of useFetch better typings.

And I added an example. But unfortunately, I can't get the examples to run - are these examples currently runnable? It seems as if it can't parse the jsx tokens in packages/react-async/Async.jsx.

[phryneas]

I just rebased this on master and it seems to fail again - I guess this is due to some change on master?

[ghengeveld]

Yes the introduction of React v16.9 is causing some issues. On CircleCI things work fine but on Travis they fail. I suspect updating the examples to 16.9 might help.

[ghengeveld]

That seems to work. Should be fixed now.

[codecov]

Codecov Report

Merging #76 into master will increase coverage by 0.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master      #76      +/-   ##
==========================================
+ Coverage   98.91%   98.93%   +0.01%     
==========================================
  Files           8        8              
  Lines         370      375       +5     
  Branches      136      140       +4     
==========================================
+ Hits          366      371       +5     
  Misses          4        4

Impacted Files	Coverage Δ
packages/react-async/src/useAsync.js	99.05% <100%> (+0.04%)	⬆️

---

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 dda874a...becc9c3. Read the comment docs.

[phryneas]

Okay, tests are now passing.
Also, I found the problem with my example: my IDE had imported useFetch from react-async/src instead of react-async and I had missed that.
Now everything should be good.

[ghengeveld]

Released in v8.0.0

[ghengeveld]

@all-contributors please add @phryneas for ideas

[allcontributors]

@ghengeveld

I've put up a pull request to add @phryneas! 🎉