[svlug] Debian testing / unstable

Marc MERLIN marc_news at valinux.com
Fri Nov 30 00:31:01 PST 2001

On Thu, Nov 29, 2001 at 07:19:54PM -0500, R P Herrold wrote:
> hunh? "impossible"?  Your memory is flogging the wrong horse.
> RH 7.0 issued 9/25/00
>   http://www.redhat.com/about/presscenter/2000/press_rhl7.html
> Package type 4 (the RH 7.0 format, [before addition of  a 
> backwards compatability mode] was introduced in rpm-3.0.4
> I was reporting issues against RPM 3.0.5 on  2000-06-23 at:
>     https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=12876
> (rpm-3.0._5_ STILL 70 days before RH 7.0 release)
> rpm-3.0.4, with a -6x variant issued  Mar 16 2000
>     ftp://ftp.rpm.org/pub/rpm/dist/rpm-3.0.x/
> (rpm-3.0.5 on Jul 21 2000, rpm-3.0.6 on Sep 14 2000 [as did 
> rpm-4.0.1] -- still 11 days before RH 7.0 released)

I must suck because I really looked for it, and sure didn't find it.
I can assure you that I sure wasn't  the only one having a hard time finding
an glibc 2.1 rpm v3 package back then.
(but it might have to do with the  fact that most of us where indeed looking
for an rpm-4.x.x  package, and that there  didn't seem to be  an obvious FAQ
entry on this on the RH site)

> You are confusing packaging format version 4 and rpm-4.x.x -- 
> like confusing lightning and lightning bugs. -- Similar names; 

I stand corrected. I did indeed think that version 4 of the format appeared
in rpm 4.x. Would have been too logical :-)

Obviously, since you  follow RPM development, you know all  this much better
than I do.
I'm sure there's also a really good reason why they've had to break backward
compatiblity 3 (or was it 4?) times when debian never had to.

If you  design your format with  a little forsight, you  leave extra fields,
and allow  for extension while  keeping backward compatiblity (see  IPV4 and
TCP, especially TCP extensions, or look at ext2)

I'm sorry, but it's  plain lame that a v4 package that  may not even contain
binaries (config or perl) is uninstallable with a v3 rpm binary.

> As I recall  it, the "real pain in  the ass" was people unable  to RTFM in
> the release notes, the RPM mailing list, of the RPM errata notice, EACH of
> which outlined  the steps -- and  yet insisting on forcing  their systems,
> with --nodeps and --force's into  unstable states; and then loudly blaming
> Red Hat.
Agreed. I don't do that myself, or if I do, I don't complain.
(I've half upgraded to RH 7.0 by hand from the command line)

PS: I have to  admit that I was  actually impressed that the  RH 7.2 install
was able  to figure out the  state my system  was in and installed  a mostly
working 7.2 system.
Microsoft is to operating systems & security ....
                                      .... what McDonalds is to gourmet cooking
Home page: http://marc.merlins.org/   |   Finger marc_f at merlins.org for PGP key

More information about the svlug mailing list