[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: librpm / corrupt free list in an FD_t (help please)



On Mon, Jan 13, 2003 at 04:19:39PM -0500, Jim Knoble wrote:
> Circa 2003-01-13 14:57:44 -0500 dixit Jeff Johnson:
> 
> : The change from v3 to v4 packaging was reverted ~2 years ago,
> : rpm-4.0.3 IIRC.
> 
> That doesn't really matter at this point, does it?  The time that it

No it doesn't.

> mattered was back when rpm-3.0.4 was installed on millions of RHL-6.2
> systems worldwide, and rpm-4.0.whatever was creating incompatible
> packages that were released as official errata updates.

Wrong: rpm-3.0.3 and earlier not finding files was what necessitated v4
packaging.

> 
> (Sigh).  I don't know why i expect any sort of consistency between
> releases of RPM with the same major version....
> 

(Sigh). I don't know why i expect users to consistently upgrade
to the latest release, particularly when the problems they are
complaining about have long since been fixed.

> : And, if you whip out hexedit and change the "4" in byte #5 to "3",
> : you'll begin to appreciate the magnitude of the "incompatibility"
> : in all its regal glory.
> 
> You don't seriously expect end users of Red Hat Linux to whip out
> hexedit and edit *signed* errata updates for chrissakes, do you?
> If you do, you're even more of an iconoclast than i thought.
> 

No, of course not. the topic was building with rpm-4.0.x and the ever
present "incompatibilities". I suggest using rpm-4.0.4, and paying
attention to details to make the "incompatibilities" go away.

FYI: changing byte #5 has absolutely no effect on *signed* packages.

> : > This is on a relatively pristine RHL-6.2 system, complete with
> : > rpm-3.0.4-0.48.  The first package was produced using rpm-3.0.5-9.6x,
> : > while the second was produced with rpm-4.0.2-6x, both using 'rpm -ba
> : > blah.spec'.
> : 
> : So don't build with rpm-4.0.2. Or rpm-4.0.1, or rpm-4.0.
> 
> Latest official update for RHL-6.2 is rpm-4.0.2-6x.
> 
> : > rpm-4.0.x builds packages with a different magic major number.
> : 
> : Note: rpm-4.0 through rpm-4.0.2 only, lest you disseminate FUD. ;-)
> 
> It's stupid that i should even have to know that certain versions of
> rpm-4.0.x don't work in a backwards compatible fashion, and others do.
> It's also stupid that Red Hat didn't release an rpm-4.0.3 erratum for
> RHL-6.2.  And it's really stupid that rpm-4.0.3 wasn't called rpm-5.0
> so people could actually talk about rpm-4.0.x without being somehow
> confusing or confused.  You've heard my beef (and others'; i'm not the
> only one) about rpm version numbers before.
> 

Yes, Jim, We Know. Later versions have been released, and can
be found on ftp.rpm.org and at ftp://people.redhat.com/jbj.

It's equally "stupid" not to apprise yourself of known problems in
software and react accordingly. Or just complain, that'll work.

> : > The primary reason rpm-3.0.5 was released was [...].  Or do you
> : > remember something different?
> : 
> : I remember being given 24 hours to cobble together a solution to
> : 	Red Hat 7.0 packages in Raw Hide don't contain any files.
> : That's the raison d'etre for v4 packaging.
> 
> It's pretty stupid that errata updates for the stable Red Hat Linux at
> the time were being built with a beta release of RPM, isn't it?  That
> wasn't your dog food, Jeff, and you shouldn't have had to eat it.  The
> right thing to do would have been for whoever made that dog food to
> rebuild the errata updates with the extant rpm-3.0.x and rerelease
> them.  Sorry you had to eat the dog food.
> 

You don't know what you're talking about here

> : > Which is why i continue to advocate using rpm-3.0.5 or rpm-3.0.6 for
> : > software targeted at RHL-6.2 systems.
> : 
> : Yes, if you want an easy old age with Red Hat 6.2, rpm-3.0.[56] (and
> : there's a 3.0.7 floating around someplace, nothing worthwhile) is the
> : rocking chair of choice.
> 
> rpm-3.0.7 is floating around in your imagination, Jeff.  It's not on
> ftp.rpm.org, ftp.redhat.com, nor people.redhat.com/jbj/, and it's never
> been released, to my knowledge.
> 
> And if customers tell me they want to use a stable system that they
> know and that they know works, i let them.  It's not as if Red Hat's
> security errata for RHL-6.2 have been particularly aggressive for the
> last year and a half anyway....
> 

Do you feel better now Jim?

73 de Jeff

-- 
Jeff Johnson	ARS N3NPQ
jbj@redhat.com (jbj@jbj.org)
Chapel Hill, NC





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index] []