On Sat, Apr 23, 2011 at 8:59 AM, <stef...@apache.org> wrote: > Author: stefan2 > Date: Sat Apr 23 13:59:23 2011 > New Revision: 1096150 > > URL: http://svn.apache.org/viewvc?rev=1096150&view=rev > Log: > * subversion/svnadmin/main.c > (subcommand_crashtest): silence 'not all control paths return a value' > warning in VisualStudio > > Modified: > subversion/trunk/subversion/svnadmin/main.c > > Modified: subversion/trunk/subversion/svnadmin/main.c > URL: > http://svn.apache.org/viewvc/subversion/trunk/subversion/svnadmin/main.c?rev=1096150&r1=1096149&r2=1096150&view=diff > ============================================================================== > --- subversion/trunk/subversion/svnadmin/main.c (original) > +++ subversion/trunk/subversion/svnadmin/main.c Sat Apr 23 13:59:23 2011 > @@ -2001,4 +2001,7 @@ subcommand_crashtest(apr_getopt_t *os, v > > SVN_ERR(open_repos(&repos, opt_state->repository_path, pool)); > abort(); > + > + /* merely silence a compiler warning (this will never be executed) */ > + return SVN_NO_ERROR; > } > >
Heh. Now gcc emits an unreachable code error: subversion/svnadmin/main.c: In function ‘subcommand_crashtest’: subversion/svnadmin/main.c:2006: warning: will never be executed I suppose we could if-def the above line depending on the compiler. If we *don't* want to play such games, my heavily biased bikeshed would be to omit the admittedly superfluous line. -Hyrum
