Re: Putting .so symlinks in libs package for dlopen()ing?
- To: Andrew Clausen <clausen@gnu.org>
- Cc: Scott James Remnant <scott@netsplit.com>, Timshel Knoll <timshel@debian.org>, bug-parted@gnu.org, debian-policy@lists.debian.org, debian-devel@lists.debian.org
- Subject: Re: Putting .so symlinks in libs package for dlopen()ing?
- From: Daniel Burrows <dnb114@psu.edu>
- Date: Mon, 9 Dec 2002 00:16:44 -0500
- Message-id: <[🔎] 20021209051644.GA24292@torrent.burrows.local>
- Mail-followup-to: Daniel Burrows <dnb114@psu.edu>, Andrew Clausen <clausen@gnu.org>, Scott James Remnant <scott@netsplit.com>, Timshel Knoll <timshel@debian.org>, bug-parted@gnu.org, debian-policy@lists.debian.org, debian-devel@lists.debian.org
- In-reply-to: <[🔎] 20021209050129.GG1135@gnu.org>
- References: <[🔎] 20021209002614.GA6420@localhost.localnet> <[🔎] 1039399332.9503.30.camel@nethack.netsplit.com> <[🔎] 20021209050129.GG1135@gnu.org>
On Mon, Dec 09, 2002 at 04:01:29PM +1100, Andrew Clausen <clausen@gnu.org> was heard to say:
> On Mon, Dec 09, 2002 at 02:02:12AM +0000, Scott James Remnant wrote:
> > I don't see why this is a problem, you'd only need to change the
> > dlopen() code if there's a SONAME change - and that should only change
> > if there's a binary-incompatible difference. A difference that might
> > not be picked up by the internal checking of the code.
>
> It would be picked up.
So reiserfs has an internal mechanism for reporting whether stuff
is really compatible? Is it using versioned symbols or something?
Daniel
--
/-------------------- Daniel Burrows <dburrows@debian.org> -------------------\
| "Hah, I can just see a real playsmith puttin' a..a DONKEY in a play!" |
| -- Terry Pratchett, _Lords and Ladies_ |
\------- (if (not (understand-this)) (go-to http://www.schemers.org)) --------/
Reply to: