Re: New perl packages

[Brendan O'Dea <bod@debian.org>]

On Tue, Jan 30, 2001 at 09:16:49AM +0100, Jérôme Marant wrote:
>. Did you have a look through the bug list in order to apply some fixes=
> in you
>packages? For instance, it would be great to get the #76966 fixed as so=
>on as
>possible, since the bug is back in your packages.

Odd.  I carried over many of the patches Darren had applied, but not
that one as the fix of adding parenthesis seemed rather dubious to me
("&_PATH_LOG || croak" and "&_PATH_LOG() || croak" should behave
identically in this instance), and I was further unable to duplicate the
error:

  # perl -Mstrict -MSys::Syslog -we 'Sys::Syslog::setlogsock "inet";' \
                                 -e 'syslog "info", "syslog test (inet)"'
  # perl -Mstrict -MSys::Syslog -we 'Sys::Syslog::setlogsock "unix";' \
                                 -e 'syslog "info", "syslog test (unix)"'
  # tail -2 /var/log/messages
  Jan 30 23:37:22 duende bod: syslog test (inet)
  Jan 30 23:37:22 duende bod: syslog test (unix)

I'd be very interested to see some example code which triggers this bug.

>. Can a core package depend on a vendor package? (I know, I'm back agai=
>n with
>the cgi::fast issue :-)) If not, you'll have to install cgi::fast as a =
>vendor
>package and make it conflict with libcgi-pm-perl ... Otherwise, everyth=
>ing's OK.=20

I've made a note to add a conflict to libcgi-fast-perl in the next
revision.

Regards,
-- 
Brendan O'Dea                                        bod@compusol.com.au
Compusol Pty. Limited                  (NSW, Australia)  +61 2 9810 3633