But dlsym(RTLD_NEXT, "calloc") works, though - no need for dlopen. That's how test-runner looks up "malloc" and "free".
On Mon, Jul 23, 2012 at 1:02 PM, Daniel Stone <[email protected]> wrote: > Hi, > > On 23 July 2012 16:11, Kristian Høgsberg <[email protected]> wrote: > > On Mon, Jul 23, 2012 at 11:40:52AM +0300, Pekka Paalanen wrote: > >> We don't have a realloc() wrapper yet, do we? > >> Array uses realloc. We're missing calloc(), too, no? > > > > Yea, we don't have those. > > I came up against this a couple of weeks ago, and the patchset I just > sent out has a realloc() wrapper. I'd really liked to have seen > calloc in there, but unfortunately you can't (at least not AFAICT), > because to get the symbol for the real calloc, you need to call > dlopen, which calls calloc ... > > Cheers, > Daniel > _______________________________________________ > wayland-devel mailing list > [email protected] > http://lists.freedesktop.org/mailman/listinfo/wayland-devel >
_______________________________________________ wayland-devel mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/wayland-devel
