Kouhei Sutou wrote:
> Hi,
>
> I want to initialize only NLS in Ruby bindings. The bindings
> doesn't need to initialize others (apr_pool, ...).
> svn_cmdline_init() initializes NLS but this initializes
> apr_pool too.
>
> This patch extract NLS initialization code from
> svn_cmdline_init() for reusing the code by bindings.
>
> In Ruby bindings, I don't want to call textdomain(). So I
> divided the code into svn_nls_environment_init() and
> svn_nls_init(). I'll use only svn_nls_environment_init() in
> Ruby bindings.
The only thing that seems weird to me about this is that it's still
rather command line specific. Taking a FILE * to print error messages
to and so forth. Is there any reason (and keep in mind, I know like
nothing about how the NLS stuff works, so there very well could be a
reason) this can't act more like normal svn functions, returning a
svn_error_t and so forth, then wrap around that for the command line
specifics? I mean it doesn't seem like the ruby bindings would really
want to write random stuff to stderr in the case of a failure in this
stuff, but on the other hand they wouldn't want to be limited to
reporting either "it worked" or "it didn't work", which is what the
current API would limit them to.
-garrett
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Jul 10 04:07:45 2005