Follow-up Comment #4, bug #17658 (project freeciv): > Without looking the code I wonder if this should be > translatable at all. Does the setting really work if > string is translated, or does it depend of it being > exactly "freeciv 2.3.0"? "freeciv 2.3.0" is the 'pretty' name of the value, and is not entered by users or saved anywhere. The corresponding 'machine'/'rule' name (which must not be translated or ever change) is "2.3.0".
This pretty/rule name pattern is common to all enumerated/bitwise options. While we could avoid this string appearing in .po files by removing the N_() markup, the pretty name will still be passed through _() anyway (and there's not a lot we can do about that), so it's more honest to keep the N_() markup. We can of course add a TRANS comment that tells translators not to worry about it. > Any 2.3.a is supposed to be compatible with any 2.3.b, > so last zero is at least misleading (you have to use > 2.3.0 format to save for 2.3.7) Well. From looking at the new save format stuff, I think part of the intended point of it is to make future new servers able to save "old"-format games (compatible with "old" servers back to 2.3.0, but not before). Given that, this allows us to be more relaxed about savefile format changes; we could, if we _really_ wanted to, make some change to the savefile format within the 2.3.x cycle, while still allowing server operators the choice to make backward-compatible savefiles. It looks like a range of versions have been reserved for this purpose. (I'm not sure what would drive us to this in a stable release series; perhaps an undiscovered design flaw requiring a format change.) _______________________________________________________ Reply to this item at: <http://gna.org/bugs/?17658> _______________________________________________ Message sent via/by Gna! http://gna.org/ _______________________________________________ Freeciv-dev mailing list Freecivemail@example.com https://mail.gna.org/listinfo/freeciv-dev