Allow liveReload to be configurable in dev server

[t27duck]

By default, liveReload for the webpack_dev_server config is the inverse of the hmr setting. There may be a usecase where one may want to control both values manually.

This allows liveReload to be configurable through webpacker.yml via the live_reload option. If not set, then the default of the inverse of hmr is used.

[PikachuEXE]

Looks good except unsure about the contentHash change

[PikachuEXE]

Manually tested locally

[PikachuEXE]

For those who want to test this locally simply update package.json:

    "@rails/webpacker": "https://github.com/t27duck/webpacker.git#live_reload_configurable",

Update config/webpacker.yml if necessary

[t27duck]

Rebased with master. PR #3169 removed the hash part of the config so the "fix the deprication warning" is also removed from this PR.

[PikachuEXE]

Anything else needs to be done to get this merged?

[guillaumebriday]

Sounds good!

[justin808]

@t27duck

There may be a usecase where one may want to control both values manually.

What is the use case that you mention?

[t27duck]

@t27duck

There may be a usecase where one may want to control both values manually.

What is the use case that you mention?

Specifically, @PikachuEXE 🤷‍♂️

[t27duck]

@justin808 more specifically, this comment from the PR when I duct taped webpack-dev-server 4 into webpacker

#3122 (comment)

[PikachuEXE]

Yup, that comment right there
I have liveReload disabled for ages in current version of webpacker
It's annoying if any small change (sometimes just adding/updating code comment) triggers a reload and lose the state on my dev website page.

[justin808]

@t27duck @PikachuEXE, OK, I get it. If not HMR, then live reload may be on or off.

But what is the point of having liveReload true if HMR is true?

[PikachuEXE]

It allows "having liveReload true if HMR is true" doesn't mean it's meaningful
But if someone want's to try it I am fine :P (they probably will learn that's meaningless)