>>>>Is there a reason svn_client_status and svn_client_status2 take a NON-const
>>>>svn_opt_revision_t?
>>>>
>>>>
>>>I believe that adding const to svn_client_status would be an ABI
>>>change and so, strictly speaking, would break our interface rules.
>>>
>>>
>>No, it wouldn't be an ABI change (it would if the language were C++).
>>But some old code which uses the function would compile with warnings.
>>
>>
>It shouldn't cause any additional warnings
>
>casting non-const to const is not a warning (adding qualifiers is okay)
>casting const to non-const is a warning (removing qualifiers is not)
>IE you can do
>
>char * a;
>const char *b = a;
>
>you can't do
>
>const char *a;
>char *b = a;
>
>
Oh, of course... What was I thinking? I guess the only case it would
cause a warning is with pointers to pointers.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Apr 22 06:18:49 2005