On Wed, Aug 26, 2009 at 07:22:47PM +0100, Adam D. Barratt wrote: > In future, it would be good if you could post debdiffs to -release and > wait for confirmation /before/ uploading. :-) OK > > * Remove open-whois.org as it is cybersquatted (Closes: #537477) > > * Backport IPv6 support for the Received header parser (Closes: #510711) > > * Correct the Maintainer field. > > * Fix numerous pod2man errors that caused warning messages to be > > embedded within several man pages. > > There are also a few other changes which weren't mentioned in the > changelog or your mail, at least the first of which looks unintentional. > The others are relatively minor, but I wasn't sure if they were intended > to form part of the update, as they weren't mentioned. I curse "svn merge" > - The documentation fix for spamhaus and SURBL not being free for all > has been reverted, as has the corresponding changelog entry from 3.2.5-2 > (and the changelog trailer for that version changed). > > - The perl-modules dependency has grown a "| libio-zlib-perl (>= 1.04)" > alternative. > > - GPG.KEY is no longer not compressed > > - sa-update is no longer run with an explicit umask by the cron job These all came from unstable. The libio-zlib-perl makes life easier for backporting this package to etch, and is the only one that I actually *meant* to include. The rest I blame on svn. > > The IPv6 header support is the only change to actual functionality. It > > fixes the problem where the config parser was unable to handle IPv6 > > addresses in config directives like "trusted_networks". Given that one > > of lenny's release goals was pervasive IPv6 support, I felt that this > > was worth fixing. Aside from myself, at least one other person is > > currently using packages including this patch on a busy mail site, and > > they work as expected. It is worth noting, though, that this change > > introduced a new dependency on libnetaddr-ip-perl. > > Personally, the patch is a little larger than I'd currently be happy > accepting for a stable update (at least for a non-security / RC fix). I > realise the changes are all in unstable, although only for a couple of > weeks. Yeah, I wasn't sure whether it'd be accepted or not. The change doesn't have any functional effect at all if you haven't configured any IPv6 addresses in trusted_networks, etc, so I believe them to be safe. Feel free to reject these packages, though; they're definitely not what I *intended* to upload. > I'd definitely suggest a lenny-backports upload including the changes > though. That's pretty much a given. SpamAssassin in bpo stays pretty well in sync with what's in testing. noah
Attachment:
signature.asc
Description: Digital signature