[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
On Sat, 4 Aug 2007 20:19:23 +0200
Love Hörnquist Åstrand <email@example.com> wrote:
> Hello Michael,
> > I want to implement krb5_config_set, krb5_config_vset ... etc. Is
> > it safe
> > to delete, replace and elements to the krb5_config_section *cf
> > member of
> > the krb5_context? Or could deleting an element create a dangling
> > pointer
> > somewhere? I'm just wondering what you think about the whole idea.
> Much of the code today assumes that krb5_config_get_string returns a
> string, so if you want to change the strings you have to modify all
> to make a local copy.
> krb5_config is heimdal specific, so its possible to change the
It seems to me you could just change the semantics of the API slightly
such that the man page might read:
"The string returned by krb5_config_get_string may not be valid after
a call to krb5_config_set_string"
Which is to say, if the caller needs to save the string for anything
beyond the immediate stack frame, they better make a copy.