Re: some PRs

From: Kris Kennaway (kris_at_obsecurity.org)
Date: 07/18/04

  • Next message: Pawel Jakub Dawidek: "Re: some PRs" 
    Date: Sun, 18 Jul 2004 10:51:00 -0700
To: Robert Millan <zeratul2@wanadoo.es>

    On Sun, Jul 18, 2004 at 06:51:42PM +0200, Robert Millan wrote:
    > On Sun, Jul 18, 2004 at 06:16:49PM +0300, Giorgos Keramidas wrote:
    > > On 2004-07-18 15:33, Robert Millan <zeratul2@wanadoo.es> wrote:
    > > >
    > > > I think it's useful for compatibility.
    > >
    > > In general, I'm not against compatibility. However, what's the end of
    > > this route? To create one special device node in /dev for every
    > > possible errno value? :-(
    >
    > I don't claim that /dev/full is useful just for the sake of it. Your argument
    > (that having a device just for each errno value is silly) is something I
    > basicaly agree with.
    >
    > But if some applications depend on it, it's still helpful for portability.
    > I don't know what support for native compatibility is expected or planned for
    > FreeBSD, but I know you have a Ports Collection with thousands of packages,
    > and this might minimaly reduce the work of your port maintainers. IMHO, you
    > should ask the people working in the Ports Collection for their opinion before
    > taking a decision.
    >
    > (Note this patch comes from the context of the Debian GNU/kFreeBSD porting
    > effort, in which we port Debian GNU/Linux packages, which are a bit more
    > likely to introduce Linuxisms than the average candidate for FreeBSD Ports.
    > Thus, our requirements might differ somewhat.)

    There's no mention of /dev/full in the build logs (e.g. configure
    script output) of any of the 11000-odd ports in the ports collection.

    Kris

  • Next message: Pawel Jakub Dawidek: "Re: some PRs"

    Relevant Pages