code in cvs is broken

[Linas Vepstas]

On Thu, Nov 15, 2001 at 02:37:38PM -0500, Derek Atkins was heard to remark:
> 
> Hrm, I hadn't thought of that; the problem is that I really wanted
> it to fail when it was being freed.  Unfortunately there is no
> way (that I know) to say:
> 	break in g_free when the argument is 0xXXXXX

another tried & true trick is to create your own library with
the routine g_free() in it.  You can then make it do whatever 
you want, including calling assert on address xxxx.  The
The only trick is to make sure it is linked in *after* any other 
libs that uses g_free, but *before* the real glib.  i.e.

cc -o myapp x.o -lgtk -lmy_g_freelib -lglib

the first symbol that the linker finds, left to right, is the one 
that gets linked in.  (That's how efence does its trick: it just 
overloads malloc & free).

(fwiw, that's also how you hunt mem leaks: you track all of the 
requests that come in through malloc(), but haven't gone out with
free().)

you sure 'mem prof' the gnome thingy, doesn't do this kind of
trap/assert?


--linas


-- 
pub  1024D/01045933 2001-02-01 Linas Vepstas (Labas!) <linas@linas.org>
PGP Key fingerprint = 8305 2521 6000 0B5E 8984  3F54 64A9 9A82 0104 5933