cmpilato@collab.net writes:
> =?UTF-8?B?QnJhbmtvIMSMaWJlag==?= <brane@xbc.nu> writes:
>
> > > * We should replace svn_error_locate() with extra arguments to
> > > svn_create_error()'s internal name.
> > >
> > How do you propose to wrap svn_error_creatf without using C99 varargs
> > macros? I don't see how it's possible short of explicitly passing
> > __FILE__ and __LINE__ to every svn_error_createf call.
>
> svn_error_create and svn_error_createf can become macros that
> optionally pass this information to some real underlying function.
> This is the same way I did our pool debugging stuff, and the way
> Sander moved that to APR:
>
> APR_DECLARE(apr_status_t) apr_pool_create_ex(apr_pool_t **newpool,
> apr_pool_t *parent,
> apr_abortfunc_t abort_fn,
> apr_allocator_t *allocator);
> APR_DECLARE(apr_status_t) apr_pool_create_ex_debug(apr_pool_t **newpool,
> apr_pool_t *parent,
> apr_abortfunc_t abort_fn,
> apr_allocator_t *allocator,
> const char *file_line);
> #if APR_POOL_DEBUG
> #define apr_pool_create_ex(newpool, parent, abort_fn, allocator) \
> apr_pool_create_ex_debug(newpool, parent, abort_fn, allocator, \
> APR_POOL__FILE_LINE__)
> #endif
Ignore me. I'm not paying attention tonight. We already do the above
in our error generation. And it's not even relevant to the
discussion.
cmpilato goes off to hide somewhere...
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Oct 29 02:10:39 2002