refactor: use modern JS constructs to validate NRIC

[timotheeg]

Context

While following code from the NIRC component, I ended up on the nirc validation where I thought several minor improvements could be made. Below are some (very minor) issues I thought about when seeing the code

• The regex is using unnecessary constructs like {1}
• Regexp specifies both lowercase and uppercase matches, while it would be simpler to uppercase the input prior to running the test
• The regexp captures groups, but these are not used because the format validation and checksum validation are disjointed
• The digits string is exploded to iterate over the characters, but that is unnecessary since the coefficients is an array of the same format that could be used instead
• We can us the string array operator to read digits from the input, and to read checksum markers (i.e. there's no need to create local arrays.)
• the reduce() function does an unnecessary assignment to a local variable, which looks confusing because of the shadow variable name sum.

The code was working perfectly fine, so the refactor is unnecessary. I initially only wanted to remove the{1} and add a i flag to the regex, but I ended up refactoring the whole file 😑

I hesitated to open the PR since it is unplanned and unnecessary, and will consume precious reviewer time, but in the end, since I did the changes, I thought I'd open it up anyway for some discussions, like:

• does using for named groups in regexes affect our browser support matrix?
• does using array operator for string character access affect our browser support matrix? (do we have precedent in code?)

Approach

I've made changes that addresses all the points above. The code reduces significantly. I am obviously really biased, but I think it is more readable now.

Testing

• Create a form with a NIRC component
• Verify validation works as expected in major browsers (Chrome, FF, safari, Edge)

[mantariksh]

a few typo/casing suggestions first, will review the logic soon!

[tshuli]

lgtm!

[timotheeg]

On my 2 questions from the PR description:

does using for named groups in regexes affect our browser support matrix?

Support matrix for named character groups

does using array operator for string character access affect our browser support matrix? (do we have precedent in code?)

String array-like character accessed was introduced in EcmaScript 5 (see doc here).

which has this support matrix:

I guess the biggest question I have now is whether we support IE11?

I just checked the transpiled (tsc+babel) result to check if by chance the code would be transformed to use older constructs like .charAt() to be more backward compatible, but it doesn't look like it does:

function(e,t,n){"use strict";Object.defineProperty(t,"__esModule",{value:!0}),t.isNricValid=void 0;var r=/^(?<prefix>[STFG])(?<digits>\d{7})(?<checksum>[A-Z])$/;t.isNricValid=function(e){var t=null==e?void 0:e.toUpperCase().match(r);if(!t)return!1;var n=t.groups,i=n.prefix,o=n.digits;return n.checksum===("S"===i||"T"===i?"JZIHGFEDCBA":"XWUTRQPNMLK")[[2,7,6,5,4,3,2].reduce((function(e,t,n){return e+t*parseInt(o[n])}),"S"===i||"F"===i?0:4)%11]}}

I will defer merging this for now, and maybe schedule a short discussion for our next sync.

[timotheeg]

Confirmed with Sonjia we do not support IE11, so I'm just going to merge this now.