improve the Number.toLocaleString() implementation

[rbri]

• if 402 is not enabled for the context, ignore the given parameters
• if 402 is enabled, support string locales

this should at least avoid the exceptions described in #535

[rbri]

@aardvark179 please have a look at lets discuss....

[aardvark179]

Okay. I've done an exhaustive search of our internal code, and all the uses I can find pass 0 arguments, so should be good with this change. The only bug report I can find internally is that somebody wanted to pass a locale and couldn't, so this is probably fine - especially as it's neat gated behind the 402 feature.

It might be worth updating the 262 tests (I think we'd need to enable 402 mode in a context factory, remove the leading ~ from the intl402 line in the properties file and update then update the properties. I think we maybe pass about 20 tests from that entire group, but it's a start.

[rbri]

@aardvark179 ok, will go forward with this, but will be busy the next days with other stuff

[andreabergia]

Seems like you accidentally removed this directory

[andreabergia]

Well, I misclicked. I meant the one after:

language/arguments-object 183/263 (69.58%)

[rbri]

@andreabergia this is in draft because i had a hard time enabling the intl402 tests.
The setup for this is different from the other test folders and or regenertator fails on this (adds wrong entries). Spend already some time over the weekend to understand the generator and try to fix it. But i faild more or less and then tried to reimplement the generator using a different approach but i'm not 100% done.

The plan is to split this into two parts, fix only the problem with some unit tests and then create a second PR for the 262.properties handling. But you have to be a bit patient because of the workshop this week and next week i have to do a talk on a conference...