Replace sprintf with normal replace support about i18n-node HOT 18 CLOSED

on my "secret" ToDo List

from i18n-node.

+1

from i18n-node.

maybe like

__('Hello :firstname :lastname', {'firstname': 'Max', 'lastname': 'Mustermann'});

just to get a feeling of prefered syntax

from i18n-node.

:firstname looks better but it might introduce problems when there is no space afterwards?
Update: See my comment below #39 (comment)

from i18n-node.

what about ${firstname} so we have a defined start and end.

from i18n-node.

I'd prefer something like ::firstname ... two :s should very rarely happen in real live, and it's faster to type that than having to open and close parenthesis.

from i18n-node.

@enyo the problem is not the start the problem is the end in your examples That's why i would prefer %key% or ${key} that should normal never occur in a normal text.

from i18n-node.

In this case there is a very limited set of variable names that can be substituted, so it's not the same as with coffeescript variables (for example), since it just needs to go through all key names and replace :KEY_NAME in the string.
The end is only a problem if you have two variables like firstName and firstNameSuffix, in which case :firstNameSuffix could be confused with :firstName + Suffix. But that could easily be handled by the script as well (always take the longest key name possible).
(And.. yes, my first comment on that was wrong. Having a space after doesn't change anything)

from i18n-node.

But this could create very unexpected behaviour. A clear end delimiter is much better. Because with one I can easily see which keys I have for translation and what is normal translation.

from i18n-node.

Mh. I think that it's extremely rare to have substitutions immediately followed by letters. Normally a substitution will be followed by a point, or space. So this behaviour will lead to confusion very rarely.
I like the :key_name syntax because it's the same as in oracle's bind variables. Using ${keyName} is a bit strange because, if you use coffeescript or jade, you're used to #{keyName}. This could result in very confusing translation strings.

from i18n-node.

well and it should't interfere with possible templating engines, which leads to a configurable interpolation option anyway :)

from i18n-node.

That's the reason why I would vote for % % it's clear. Easy to read and shouldn't affect any template languages as far as I know. Oracles bind variables use : as the beginning and space as the end.

from i18n-node.

%key% looks fine to me too. I'm just really opposed to ${key}. (Also because it's common for programmers to write that, but the people writing translations mostly aren't programmers.)

from i18n-node.

+1

from i18n-node.

Any news on that?

from i18n-node.

I noticed the fork at matthias/i18n-node has support for mustache in the message strings

from i18n-node.

it's about to get merged - tracked in #76 I just wanted to get a quick review on it but didn't find time yet

from i18n-node.

released in 0.5.0 https://github.com/mashpie/i18n-node#mustache-support

from i18n-node.