On Sat, Dec 17, 2016 at 9:06 PM, <[email protected]> wrote: > So, I believe that the fact that we allow for function argument values to be > complex strings is an unnecessary vector for complicated parser/resolver code > and is not something FTL users would expect. > > Does it answer your question?
To some extent :) It tells me that we haven't seen enough usage of this feature yet to really make good decisions. From my point of view, I'd say it would be surprising if strings marked with "" could or could not have placeables in them depending on where they are used. But I'm aware that this might just be my inclination towards consistency and not something users of FTL would have an issue with. Regarding your other point about complicated parser/resolver code, I would think that the current design which simply uses the existing TextElemenet syntax with all its semantics is already very lean implementation-wise. I agree that it's rather useless to allow for complex strings in function arguments. I don't know right now if it's better to simplify by removing this feature (and possibly introducing new syntax) or to keep it consistent with the current semantics of "". Perhaps there is a middle-ground to be found in that we don't change the syntax but instead impair our implementation (via a ParseError), so that we can decide later and not have to deal with existing translations already using the feature. Also, we already have a symbol which has a value and doesn't allow interpolations: keywords. We could consider the following as an alternative: NUMBER($num, style: [percent]). _______________________________________________ tools-l10n mailing list [email protected] https://lists.mozilla.org/listinfo/tools-l10n
