Front page | perl.ponie.dev |
Postings from July 2005
On Thu, Jul 21, 2005 at 02:36:32PM -0400, Andy Dougherty wrote: > On Thu, 21 Jul 2005, Andy Dougherty wrote: > > > And the compiler seems much happier with the following patch. I'll test > > more carefully to make sure it doesn't raise a hackle somewhere else. > > > > Index: perl/sv.h > > =================================================================== > > --- perl/sv.h (revision 331) > > +++ perl/sv.h (working copy) > > @@ -1501,5 +1501,5 @@ > > #endif > > > > #ifndef MUMBLE > > -#define MUMBLE(p) p > > +#define MUMBLE(p) ((void *)p) > > #endif > > Well, it did give me 7 new warnings: > > "gv.c", line 1428: warning: a cast does not yield an lvalue > "op.c", line 6724: warning: a cast does not yield an lvalue > "sv.c", line 308: warning: a cast does not yield an lvalue > "regexec.c", line 3232: warning: a cast does not yield an lvalue > "universal.c", line 337: warning: a cast does not yield an lvalue > "universal.c", line 337: warning: a cast does not yield an lvalue > "opmini.c", line 6724: warning: a cast does not yield an lvalue > > but everything seemed to work (I ran out of disk space before > completing a successful make test). Also, as I understand the plan, > this is going to change some more anyway, so these 7 warnings are much > better than the 2000 they replace. Thanks. I found that if I tweak it to this: Index: sv.h =================================================================== --- sv.h (revision 331) +++ sv.h (working copy) @@ -1501,5 +1501,5 @@ #endif #ifndef MUMBLE -#define MUMBLE(p) p +#define MUMBLE(p) ((void *)(p)) #endif then I don't see those errors either. (from gcc in maximal bitchiness mode) Nicholas ClarkThread Previous | Thread Next