Module:Template translation: Difference between revisions

Module:Template translation (edit)

Revision as of 00:21, 2 February 2014

1,415 bytes added , 10 years ago

m

there are other chars that generate exceptions in mw.language.isKnownLanguageTag, and some language codes still not validated correctly; check this correctly

Anonymous user

metawikimedia>Verdy p

Revision as of 16:38, 1 February 2014 (edit) metawikimedia>Verdy p mNo edit summary ← Older edit		Revision as of 00:21, 2 February 2014 (edit) (undo) metawikimedia>Verdy p m (there are other chars that generate exceptions in mw.language.isKnownLanguageTag, and some language codes still not validated correctly; check this correctly) Newer edit →
Line 1: local this = {} function this.checkLanguage(subpage~~, default~~) --[[Check first if there's an ~~apostrophe,~~any ~~because~~invalid ~~they~~character ~~break~~that ~~the~~would ~~isKnownLanguageTag~~cause the ~~if (~~mw.language.isKnownLanguageTag function(~~subpage)~~) to throw an exception:▼ ~~function. This test does not work with regexps, use plain search instead (no need~~ - all ASCII controls in [\000-\031\127], ~~to use Unicode parser, apostrophes can only appear isolated as one byte in UTF-8).~~ - double quote ("), sharp sign (#), ampersand (&), apostrophe ('), ]]▼ - slash (/), colon (:), semicolon (;), lower than (<), greater than (>), if (string.find(subpage, "'", 1, true) == nil)▼ - brackets and braces ([, ], {, }), pipe (\|), backslash (\\) All other characters are accepted, including space and all non-ASCII characters (including \192, which is invalid in UTF-8). ▲ --]] if mw.language.isValidCode(subpage) and mw.language.isKnownLanguageTag(subpage) --[[However "SupportedLanguages" are too restrictive, as they discard many valid BCP47 script variants (only because MediaWiki still does not define automatic transliterators for them, e.g. "en-dsrt" or "fr-brai" for French transliteration in Braille), and country variants, (useful in localized data, even if they are no longer used for translations, such as zh-cn, also useful for legacy codes). We want to avoid matching subpagenames containing any uppercase letter, (even if they are considered valid in BCP 47, in which they are case-insensitive; they are not "SupportedLanguages" for MediaWiki, so they are not "KnownLanguageTags" for MediaWiki). To be more restrictive, we exclude any character that is not ASCII and not a lowercase letter, minus-hyphen, or digit, and any code that does not start by a letter or does not finish by a letter or digit. of that has more than 8 characters between hyphens, or has two hyphens. --]] or string.find(subpage, "^[%l][%-%d%l]*[%d%l]$") ~= nil and string.find(subpage, "[%d%l][%d%l][%d%l][%d%l][%d%l][%d%l][%d%l][%d%l][%d%l]") == nil ▲ ifand (string.find(subpage, "'%-%-"~~, 1, true~~) == nil) then ~~-- Return the~~return subpage ~~only if it is a valid language code.~~ ▲ if (mw.language.isKnownLanguageTag(subpage)) ~~then~~ ~~return subpage~~ ~~end~~ end -- Otherwise there's currently no known language subpage