--On Wednesday, March 31, 2004 1:01 AM -0500 Greg Hudson <ghudson@MIT.EDU>
wrote:
> Hm. It seems like Nico's suggested approach is a much simpler change
> than what you're suggesting, and they both have weaknesses of roughly
> equal magnitude. (Nico's fails when the server doesn't have a language
> file for your language; yours fails when the client doesn't know about
> the particular error generated by the server.)
Yes, but whom do we think is going to be more likely to upgrade? If it
matters to you as a user to have the 'latest' translations, then I think it's
reasonable to place that burden solely on the client. If we place the burden
on the server, then we'd force them to upgrade to deliver new translations to
their users. To me, that seems harsh - especially when we're going to all
this effort to maintain client/server compatibility between versions.
Of course, if you don't recognize the error code, you just display the 'raw'
message. But, if you get back a bogus translation, that doesn't help you at
all as you can't fall back onto something better. So, I think the failure
mode with client-translated errors is easier to recover than being fed
poor/out-of-date translations over the wire. -- justin
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Mar 31 08:15:49 2004