Added the possibility to configure the StyleLoader via the method Enc…

[tooltonix]

…ore.configureStyleLoader()

[Lyrkan]

Hi @tooltonix,

Thank you for this PR.

Besides the few remarks below the implementation looks good, but we'll also need some test cases to make sure that it works and will still work in the future :)

Also I'm wondering if this should be added as a separate method or directly as a disableCssExtraction() argument (since it is the only method that enables the loader)... what do you think? (not asking you to change it, I only want your opinion! Also, ping @weaverryan and @Kocal for the same question).

[Kocal]

Hi 👋

Also I'm wondering if this should be added as a separate method or directly as a disableCssExtraction() argument (since it is the only method that enables the loader)

Hum, I am bit divided on this point but I think a separate method is better.

If we had use a disableCssExtraction() argument:

• yes style-loader is only used if CSS extraction is disabled, but it feels weird to me to configure a loader in a method which is not named configure...Loader()
• what if one day we use style-loader somewhere else? (but I can't imagine cases where it would be possible 😛)

But if we have a configureStyleLoader:

• that's nice, a method for configuring a loader
• but the loader is not active by default, only when the CSS extraction is disabled. We should warn the user about this

[tooltonix]

Due to internal requirements, we use a different style loader, which we install per yarn, so that the options of the previous style loader are no longer applicable. In the latest version 1.1.3 of the StyleLoader, the loader automatically inject source maps when previous loader emit them.

It is correct that the method disableCssExtraction() activates the StyleLoader, but I would override the options with a separate method configureStyleLoader(), as it is more consistent for me in context and name.

[weaverryan]

Hey @tooltonix!

This looks good to me - and I agree with the 2 separate methods, but I'm happy we had the conversation - I think it will help us with some docs to show that these methods are connected.

Only minor comments. But we do need a functional test for this. See functional.js - there is already a case that includes disableCssExtraction(). We could duplicate that test case, maybe set the attributes option to TESTING_ATTRIBUTES and then assert that this string is present inside main.js.

Cheers!

[tooltonix]

Hey @tooltonix!

This looks good to me - and I agree with the 2 separate methods, but I'm happy we had the conversation - I think it will help us with some docs to show that these methods are connected.

Only minor comments. But we do need a functional test for this. See functional.js - there is already a case that includes disableCssExtraction(). We could duplicate that test case, maybe set the attributes option to TESTING_ATTRIBUTES and then assert that this string is present inside main.js.

Cheers!

Hi @weaverryan

I'm sorry, but I haven't written any tests in Mocha yet. But I tried anyway and the test was successful, as long as I used the correct options of the style loader in version 1.1.3. I hope that the test is sufficient. Otherwise I ask for appropriate suggestions.

[weaverryan]

Very excellent work! Thanks so much for this Patrick!