Bernhard Schmalhofer <[EMAIL PROTECTED]> wrote: > Leopold Toetsch wrote: > > > > This is what the "t" signature char is doing anyway - Oops or better, > > what it should to. While it's using string_to_cstring the created string > > isn't freed (unless the library would free it, which is unlikely).
> Actually it is freed right after the external function call. in nci.c yes in jit/i386 no. > Which is bad when the freed memory is accessed later. Yep. > > So what we need for the "t" signature is a string_pin with NUL append. > Sounds good to me. I'm just wondering how string pinning can be tested. > I guess I have to look into the garbage collection tests. Dead strings with the PObj_sysmem_FLAG set, do the right thing. You just have to take care that you hold a ref to that string as long as you need it (ot the lib of course). > CU, Bernhard leo