On 31 Jul 2002, Turbo Fredriksson wrote: > >>>>> "Adam" == Adam Conrad <[EMAIL PROTECTED]> writes: > > >> -----Original Message----- From: [EMAIL PROTECTED] > >> [mailto:[EMAIL PROTECTED] Sent: Wednesday, July 31, 2002 12:21 > >> AM > >> > >> Package: libc6 Version: 2.2.5-10 > >> > >> ----- s n i p ----- meduza:/tmp# dpkg -i > >> libc6_2.2.5-9.woody.4_i386.deb > > Adam> This version doesn't exist in the archive. > > Sorry about that. The version in the bug report (2.2.5-10) is correct. > The dpkg line was a cut-and-past from the discussion on debian-devel. > > >> In conversation on debian-devel, the problem is that it's using > >> the md5sum output from a newer dpkg, which is yet to be > >> unpacked/installed. > > Adam> The argument makes little sense, since the dpkg with the > Adam> changed behaviour only exists in sid, and not woody. It's > Adam> likely that -9.woody.4 did in fact have this problem (as it > Adam> was a backport from unstable), but -10 doesn't appear to. > > It was when installing the -10 version that the problem first showed > up, so I tried a version from the local mirror (ftp.acc.umu.se). > > Adam> You have 2.2.5-10 installed right now. Was it an issue with > Adam> that version? > > Yes. The only way to solve it was to unpack (dpkg -x) the new dpkg > package...
What version of dpkg? The error 'malformatted' does not appear in any context with the binary 'md5sum'. dpkg does not call /usr/bin/md5sum to check, it does it internally.
