Handle Currencies Stored as Strings in Formulas

[fdjohnston]

Added a function to the new FormattedNumber helper that will use the currency code pulled from localeconv by StringHelper::getCurrencyCode(). The currency code is preg_quoted and then dropped into a regexp that is modelled on the expression developed for convertToNumberIfPercent. This will allow locale independent operation. Unfortunately localeconv only provides information about standard currency formats and not accounting formats. This means that we can't say for sure whether or not a locale has an accounting format where the currency symbol shows up in a non-standard position (eg. left justified for some countries instead of appearing directly in front of the value). The regexp should handle these cases.

The primary issue with this approach is that the regexp will incorrectly match invalid currency formats for some locals as it checks for the symbol before and after the value.

A possible improvement would be to pull p_cs_precedes and n_cs_precedes from localeconv and using them to determine if the currency symbol should appear before or after the value for the current locale. Since white-space is ignored by the regexp, accounting formats should still work, as long as the accounting format doesn't involve moving the symbol to the opposite side of the value.

This is:

- [ ] a bugfix
- [x] a new feature
- [ ] refactoring
- [ ] additional unit tests

Checklist:

• Changes are covered by unit tests
  • Changes are covered by existing unit tests
  • New unit tests have been added
• Code style is respected
• Commit message explains why the change is made (see https://github.com/erlang/otp/wiki/Writing-good-commit-messages)
• CHANGELOG.md contains a short summary of the change
• Documentation is updated as necessary

Why this change is needed?

See issue #3165

[MarkBaker]

I've made a couple of small changes

• Use of the unicode flag for the regexp, for currency codes that aren't a part of the standard ASCII character set
• Some additional unit tests for positive sign
• Handling of currency/accounting strings with a thousands separator (not locale-aware)

Thanks for your work on this

[fdjohnston]

Great improvements; thank you.

Happy to contribute to such a great project, and may I say I've really enjoyed going back and forth with you on this. I have definitely learned some new tricks with regular expressions.

If you ever decide to make the solution Locale aware or explore the PHP Intl option you mentioned over in #3165, and if you'd like a hand, I would be happy to try to help out.