Process available chunk names properly in dev mode

[fatfisz]

Fixes #4603.

Tests explain it the best:

    describe('development mode (no chunkhash)', () => {
      it('should strip the extension from the filename', () => {
        const filename = 'foo_bar_0123456789abcdef.js'
        expect(getChunkNameFromFilename(filename, true)).toBe('foo_bar_0123456789abcdef')
      })

      it('should only strip the extension even if there\'s a hyphen in the name', () => {
        const filename = 'foo-bar-0123456789abcdef.js'
        expect(getChunkNameFromFilename(filename, true)).toBe('foo-bar-0123456789abcdef')
      })
    })

    describe('production mode (with chunkhash)', () => {
      it('should strip the hash from the filename', () => {
        const filename = 'foo_bar_0123456789abcdef-0123456789abcdef.js'
        expect(getChunkNameFromFilename(filename, false)).toBe('foo_bar_0123456789abcdef')
      })

      it('should only strip the part after the last hyphen in the filename', () => {
        const filename = 'foo-bar-0123456789abcdef-0123456789abcdef.js'
        expect(getChunkNameFromFilename(filename, false)).toBe('foo-bar-0123456789abcdef')
      })
    })

[timneutkens]

Thanks!

[fatfisz]

@timneutkens Just wondering when the next release will be? After this regexp "fun" my code will finally work, so is it too much to ask for a canary release soon?

My code is experimental, so I don't care about using a stable version of Next.js as long as it does not break the base functionality, which in this case is dynamic context imports.