I found the b'' idea on a recent message here between you and Nick I think, it seemed interesting. It's gone now, as well as the typo, thanks MRAB. The summary is, it is a superset of PEP 498 with i18n integrated, additional background to inform the arbitrary expression decision, and security policy specified for those who will be concerned about it. As for deferring interpolation, I suppose I'd recommend existing N_... syntax until I've had a bit more time to think about it. -Mike On 08/20/2015 05:39 PM, Guido van Rossum wrote:
Can you give a brief discussion of how your version differs from PEP 498? So far I've found:
- lots of language summarizing the discussion following PEPs 498 and 501 - %(name)s in byte strings (which I think is abominable) - t prefix for translated strings - some optional ideas (which I'm skipping for now)
Am I missing something?
Second, do you have a proposal for marking translatable strings that should be extracted by pygettext but not interpolated in the spot where they occur? (This is the N_(...) format from the pygettext docs.)