Re: [Gaim-devel] Resolving IP addresses in Gaim

[Ethan B. <ebl...@cs...>]

Mark Doliner spake unto us the following wisdom:
> On Mon, 21 Aug 2006 00:35:11 -0500, Gary Kramlich wrote
> > Mark Doliner wrote:
> > > We use either getaddrinfo() or gethostbyname(), and they are both blo=
cking.
> >
> > I haven't looked at the changes, but I hope, now that it's threaded that
> > we're using gethostbyname_r or gethostbyname2_r, since the man page say=
s:
> >=20
> >       The functions gethostbyname() and gethostbyaddr() may return
> >       pointers to static data, which may be overwritten by later
> >       calls. Copying the struct hostent does not suffice, since it
> >       con=1B$B!>=1B(B tains pointers; a deep copy is required.
> >=20
> > This of course, is implementation specific.
>=20
> I'm working on that right now, actually.  Unfortunately
> gethostbyname_r() is a GNU extension, which means we shouldn't use it.
> Which means that we'll have to use gethostbyname() in a blocking
> manner on systems without addrinfo.  Fortunately this should be a
> relatively small number of systems.

This is what I mentioned on IRC the other day, with respect to using
reentrant lookups where available in case some other library chooses
to do a lookup.  (Viz. the recent gnome-vfs deciding to go threaded,
for example.)  gethostbyname_r is most certainly *not* GNU-specific,
it exists on several systems that I am aware of.  Perhaps the GNU
implementation diverges in some way with, e.g., the Solaris version,
but I expect we can assume they are the same.  I'll peruse the man
pages later and find out.

Regression to blocking lookups on platforms without gethostbyname_r is
a very good argument for returning to cooperating processes, if you
ask me.  :-P

Ethan

--=20
The laws that forbid the carrying of arms are laws [that have no remedy
for evils].  They disarm only those who are neither inclined nor
determined to commit crimes.
		-- Cesare Beccaria, "On Crimes and Punishments", 1764