[mbstring][PHP 8.4] Add mb_ucfirst and mb_lcfirst to polyfills

[Ayesh]

Adds polyfills for mb_ucfirst and mb_lcfirst functions based on the polyfill shown in PHP.Watch. It basically takes the first mb character, calls mb_convert_case with MB_CASE_TITLE or MB_CASE_LOWER, and returns with the concat of the remaining string.

The tests are taken from the php-src PR.

• RFC
• php-src PR: php-src#13161
• PHP.Watch - PHP 8.4: New mb_ucfirst and mb_lcfirst functions

[zonuexe]

self::mb_internal_encoding() performs validation checks when setting internal values, so adding a try-catch to the else avoids a small amount of overhead. What do you think?

Suggested change

	}
	try {
	$validEncoding = @self::mb_check_encoding('', $encoding);
	} catch (\ValueError $e) {
	throw new \ValueError(sprintf('mb_ucfirst(): Argument #2 ($encoding) must be a valid encoding, "%s" given', $encoding));
	}
	// BC for PHP 7.3 and lower
	if (!$validEncoding) {
	throw new \ValueError(sprintf('mb_ucfirst(): Argument #2 ($encoding) must be a valid encoding, "%s" given', $encoding));
	}
	} else {
	try {
	$validEncoding = @self::mb_check_encoding('', $encoding);
	} catch (\ValueError $e) {
	throw new \ValueError(sprintf('mb_ucfirst(): Argument #2 ($encoding) must be a valid encoding, "%s" given', $encoding));
	}
	// BC for PHP 7.3 and lower
	if (!$validEncoding) {
	throw new \ValueError(sprintf('mb_ucfirst(): Argument #2 ($encoding) must be a valid encoding, "%s" given', $encoding));
	}
	}

[Ayesh]

You are right I understand the else block would avoid the non-null calls.i tried to follow the rest of the polyfill methods, so I'd prefer to keep this one like it is now.

[zonuexe]

Perhaps this code can be condensed by combining it into a method like self::assertEncoding(). Let's try another PR.

[Ayesh]

Thanks a lot for the review @zonuexe those were great. I addressed all the points except the else block remark, mainly to make sure we follow the other polyfill methods, although there is indeed a merit if that changes. Perhaps it's something we can do for all methods in a separate PR?

[nicolas-grekas]

Thank you @Ayesh.