Limit number of possible suggestions

How the code is written, duplicates should not even be possible to be submitted. If they can be submitted, I'd consider it a bug. Also, kkaefer has an issue to support diffing between suggestions, so you can see the exact changes pretty darn fast. See #563128: Rework the translation UI. Not sure whether that covers the use cases you'd want to cover with this feature request.

Instead of setting a hard limit to the number of suggestions, we could instead opt to display a warning message when trying to submit a new suggestion when there are more than a certain number of suggestions already available for that string.

Yeah, that one sounds like a decently better idea to me at least! :)

I might have missed a point here, but as far as I can tell, duplicates (for translations) are very possible. I put probably 100 or so of them into l.d.o a week ago when doing mass imports.

If you look at Norwegian Bokmål translation of «Stylesheet URL» you'll see there's a duplicate (and btw, I believe they are both wrong, but I'll leave that for later). I had at least four identical suggestions for the same string at a point, there could be more of those. The server should definately leave out suggestions (or translations) that are already in, but some of those I put in could perhaps be explained by importing very many similar strings at the same time.

Yeah, then we have a bug here. The server should not allow the same suggestion for the same string multiple times.

I think this could be the result of import handling strings different than "normal" translate. Using the translate screen i can not add another duplicate, whereas import allows me to add duplicates in bulk.

I understand l10n server accepts translations from remote (l10n client (++?)). Shouldn't all translations/suggestions be put through a common filter before adding them to the database?

@Eirik_R: yes, a common filter would be best.

I've just submitted code (#494320: Unify translation checks when saving and skip whitespace only strings) and planning to deploy fixes to localize.drupal.org to not allow duplicate submissions from .po files or remote suggestions either, just like it was for UI based translations. Whitespace only submissions are also stopped before getting into the database. Maybe that would make this a duplicate then?

Is this still relevant given the duplicate submissions now averted? I'm still not sure this one is a very good idea.

Suppose we should divide translation & suggestion as grouping entity to allow extend suggestion functionality

Also this could make queries lighter

'l10n_server_string'
'l10n_server_translation' - only one which used as active
'l10n_server_suggestion' - community could poll on suggestion and some action will copy it's data to active-translation table

agree to #12