feat(translationtool): Allow vue-translator comments

[jotoeri]

Hi all,

i just gave it a go to extract translator-comments from vue-files. I fell over that in nextcloud/documentation#5422
I'm neither the great php pro, and not at all for Regex. The attempt is probably not the ideal one and it needs a bit more looping than before. But it seems to work. 🙈
So comments and review-suggestions are welcome! 😊

Three things from my side:

• The regex check for TRANSLATORS is executed twice - Once for HTML and once for JS. This is due to the fact, that i didn't find a Regex to match either the HTML-EndTag --> or just nothing for JS-Comments. Does one of you have an idea here?
  EDIT: Just had an idea and pushed a fixup: Now checking for <!--|//|/* TRANSLATORS and then separately removing possible end-tags -->|*/
• One issue i realised, but it is more or less independent and already existed before: Double-Quote matches most probably use double-quotes as they want to use a single quote within their strings. However, they are parsed to be single-quotes here, which does not work anymore. (I saw an issue on some app, but i don't find it anymore 😞) Any idea for this? Or probably do it as separate topic? 🤔
• Can we test this somehow properly before merging / going live? I was only able to test my new code out of scope, but therefore changing some small code parts to make it run.

Greets! :)
Jonas

cc @rakekniven 😉

Fake-File Before:

Fake -File After:

while the first comment is out of HTML and the second one out of JS:

[rakekniven]

@jotoeri Wow, thank you! Hope that reviews will be approved and it will hit ci soon ;-)

[jotoeri]

Any comments @juliushaertl @MorrisJobke @ChristophWurst ? 🤔

[ChristophWurst]

if you read line per line, do multi line t/n calls still match? I doubt it

try parsing code that says

const text = t(
  'test',
  'my text'
);

[jotoeri]

Ahh, you're right yes.

So - the only idea i have then is a quite complex regex, that catches the full block. Just quickly tested this one and it seems to work on my testfile, also with the multiline translations.

Would you be ok with using such a regex? Or do you think it's too complex & error-prone? 🤔

$translatorsRegex = "(\W(\/\/|\/\*|<!--)\s(TRANSLATORS)\s(.+))*"; // Indicating an optional TRANSLATORS-comment
$inbetweenRegex = "(\s*.*\s*)"; // Text between comment and translation-function

preg_match_all("/". $translatorsRegex . $inbetweenRegex . "\Wt\s*\(\s*'([\w]+)',\s*'(.+)'" . "/", $vueSource, $singleQuoteMatches);

[ChristophWurst]

is a regex ever not complex or error-prone? 😉

is there any measurable performance impact?

[jotoeri]

So - implemented and pushed now a fully working version.
However - yes, it seems like this currently has a huge impact on performance... So needs for sure some more love before it is ready to go. 😞

[rakekniven]

Any progress here? The translators will happily starting a party. At least me 😃

[jotoeri]

Ahh, I'm sorry, @rakekniven. I still have it on backlog & in my notifications, just recently had some other PRs to complete and due to exams at university not too much of a time for some hacking.
So it's not lost, i still want to complete that, i just need to take the time for. 😉

[rakekniven]

The last weeks I have seen many questions about context of strings based on vue files.

I am kindly asking @nickvergessen to have a look at this one.

Maybe @juliushaertl can spend some time on it as well?

[skjnldsv]

I would honestly be a little more comfortable with a bit of easy testing on this script 🙈

[rakekniven]

@skjnldsv For sure anyone is invited and appreciated by all translators :-)

[rakekniven]

Any update on this one?