Javier Fernández-Sanguino Peña wrote: > (please CC: on replies, I'm not subscribed, sorry for breaking the thread > because of this) Please, don't set a M-F-T header with only the list address if you want to be CCed... >> Luk Claes wrote: >>> Javier Fernandez-Sanguino wrote: >>> - debian-goodies (0.27): uploaded yesterday, fixes important (non-RC) bug >>> #264985 (checkrestart is useless in previous releases) and also fixes >>> non-RC bugs in another script (network-test), it also enhances >>> documentation by providing one of the missing manpages (checkrestart.1) >> We have to draw the line somewhere, I'm not keen to unblock this... > > I really think this one should be unblocked, based on the guidelines for > changes accepted sent by Andreas: > >> Andreas Barth wrote: >> (...) here are the guidelines for changes that will be >> accepted into testing during the freeze: >> (...) >> - fixes for severity: important bugs in packages of priority: optional >> or extra, only when this can be done via unstable; > > #264985 is severity 'important' and debian-goodies is priority: optional > >> (...) >> - documentation fixes. > > The checkrestart.1 manpage is a documentation fix, and, really, an important > one. The manpage says it might be useful to determine if a service restart is > necessary after a system update to weed out security bugs. But, at the same > time, it explicitly warns that admins should not exclusively depend on this > tool (since there are sometimes false positives) to determine wether or not > the system needs to be rebooted after a system-upgrade. > > AFAIK there is currently no other tools to do this in Debian (libc6 has > something in a crude way in it's postint IIRC), so the need is there. I'd > rather we don't ship tools which are defective and users migh rely upon them > for security purposes [0] because they are pointed to [1] or recommended by our > users [2]. > > I know this should have been fixed a long time ago in the release cycle. > Sorry for bringing this up now, but I didn't had the Python skill to do it > and a patch for this bug was not provided until recently. This is no good excuse IMHO as you could have tagged the bug 'help' or asked on d-devel etc... Though unblocked anyway after another review and some days more testing. Cheers Luk -- Luk Claes - http://people.debian.org/~luk - GPG key 1024D/9B7C328D Fingerprint: D5AF 25FB 316B 53BB 08E7 F999 E544 DE07 9B7C 328D
Attachment:
signature.asc
Description: OpenPGP digital signature