Refactor CamelCaseToSeparator Filter

[ramchale]

Q	A
Documentation	no
Bugfix	no
BC Break	yes
New Feature	no
RFC	no
QA	no

Description

Refactoring CamelCaseToSeparator based on #177

[gsteel]

@ramchale

Had another look at the regex - I think this is more like it: https://3v4l.org/ISIYe

[ramchale]

Thanks
I think the solution to #10 needs some discussion, so I've split that into it's own Draft PR

[ramchale]

@ramchale

Had another look at the regex - I think this is more like it: https://3v4l.org/ISIYe

Awesome. Hadn't seen this before I created the additional draft. I'd put in my own unicode example, but your change seems to work for that as well 😄

[gsteel]

@froschdesign Can I have your opinion on the use of str_replace here?

#186 (comment)

I can't decide if input like 'Foo-Bar' should be filtered to 'Foo-Bar' or 'Foo---Bar' - I'm currently preferring the latter because it is consistent with 'Foo/Bar' => 'Foo-/-Bar'.

WDYT?

[ramchale]

@froschdesign Can I have your opinion on the use of str_replace here?

#186 (comment)

I can't decide if input like 'Foo-Bar' should be filtered to 'Foo-Bar' or 'Foo---Bar' - I'm currently preferring the latter because it is consistent with 'Foo/Bar' => 'Foo-/-Bar'.

WDYT?

A third option is to add an entry to the constructor Options (something like $this->deduplicateSeperator = $options['deduplicateSeperator'] ?? false;) to toggle the use of the str_replace. Would also need to add this to CamelCaseToUnderscore and CamelCaseToDash as it currently stands.

[gsteel]

A third option is to add an entry to the constructor Options (something like $this->deduplicateSeperator = $options['deduplicateSeperator'] ?? false;) to toggle the use of the str_replace. Would also need to add this to CamelCaseToUnderscore and CamelCaseToDash as it currently stands.

If we're doing extra options - it would be better to perhaps to add stripNonWord and then preg_replace anything else with a single separator, so:

I'mAMuppet becomes I-m-A-Muppet, but at this point, we're trying to save devs from themselves - this filter is supposed to convert camelcase input, if you've not provided camelCase input then any weird output is on you, so I'm in favour of filtering With-Dashes to With---Dashes and removing any attempt to normalise input.

Also, remember that filter chain can be used, i.e. Trim | Alnum | CamelCaseToDash

[ramchale]

Updated this in line with the conversation so far. Let me know if it needs further changes

[froschdesign]

Some changes to the documentation are required.

[froschdesign]

This new behaviour is not covered in the documentation. Example:

TIP: **New Behaviour since Version 3**
The filter will now treat numbers as a word boundary.
For example `ThisHas4Words` will filter to `This-Has-4-Words`.

[froschdesign]

This new behaviour is not covered in the documentation.

[froschdesign]

The note on constructor options is missing:

The constructor now only accepts an associative array of [documented options](../word.md#camelcasetoseparator).

[froschdesign]

This new behaviour is not covered in the documentation.

[gsteel]

LGTM! Thanks @ramchale

I'll wait for @froschdesign to approve the docs changes - I think behaviour wise, what we have is fine and any weird results from users supplying input that is not camel cased is the users problem.

[ramchale]

@gsteel awesome, thanks. I think you solved the trickiest bit 😄

[froschdesign]

@gsteel
Before the release I will check and update the documentation again.

[gsteel]

Thanks again @ramchale 👍