RE: 0.99^H^H^H^H1.04 vs redhat

Thx! I'm still keeping the clean tree, but maybe the "BUG" note in 571 can
be removed now... :-)

Incidentally, there's a thread on rpm-list about labelcompare with a note
about using rangesoverlap in the last email:
http://www.geocrawler.com/mail/thread.php3?subject=labelcompare&list=87
								-Rob

> -----Original Message-----
> From: John Berninger [mailto:[email protected]]
> Sent: Wednesday, June 05, 2002 3:52 PM
> To: [email protected]
> Subject: Re: [Current-server] 0.99^H^H^H^H1.04 vs redhat
> 
> 
> Rob et al:
> 
>         Our code right now only checks for return values of 
> -1 or 0 when
> comparing package versions; obviously, this is a problem.  I've just
> sent a changeset to Hunter which addresses this issue; to get it
> working, you should be able to change line 588 in channel.py from:
> 
>         if ( compare == -1 ):
> 
> to
> 
>         if ( compare > 0 ):
> 
>         Sorry for the mistake there...
> 
> On Wed, 05 Jun 2002, Rob Hagopian wrote:
> 
> > Well, this problem hasn't gone away (I've been a freak 
> about keeping a clean
> > source tree since the problem hadn't been "resolved")... 
> :-( I just got the
> > latest errata for 6.2->7.3 and the xchat packages weren't correctly
> > versioned. This is a current server running v1.04 on redhat 
> 7.1 with all the
> > errata (for those that forget, there's a few tidbits about 
> rh7.1 vs. rh7.2
> > in the quoted emails; the consensus being that it 
> *shouldn't* matter).
> > 
> > dh-i386-6.2 (works great):
> > Jun 05 15:11:50 -------------------------------------
> > Jun 05 15:11:50 Examing xchat-1.8.9-1.62.0.i386.rpm
> > Jun 05 15:11:50 dep_names[xchat] = ['xchat:1.8.7:1.62.0:1']
> > Jun 05 15:11:50 hits = ['xchat:1.8.7:1.62.0:1']
> > Jun 05 15:11:50 dups = [['xchat', '1.8.7', '1.62.0', '1', 
> 'i386', 1066942]]
> > Jun 05 15:11:50 dups = [['xchat', '1.8.7', '1.62.0', '1', 
> 'i386', 1066942]]
> > Jun 05 15:11:50 Comparing packages: xchat-1.8.9-1.62.0 and
> > xchat-1.8.7-1.62.0 yields 2
> > Jun 05 15:11:50 Comparing packages: xchat-1.8.7-1.62.0 and
> > xchat-1.8.9-1.62.0 yields -2
> > pkg0 = xchat, pkg4 = i386, filename = xchat-1.8.7-1.62.0.i386.rpm
> > Jun 05 15:11:50 Removing xchat-1.8.7-1.62.0.i386.rpm
> > Jun 05 15:11:50 Older package deleted from db.
> > Jun 05 15:11:50 Adding xchat-1.8.9-1.62.0.i386.rpm
> > 
> > dh-i386-7.0 (uh oh):
> > Jun 05 15:14:55 Examing xchat-1.8.7-1.70.0.i386.rpm
> > Jun 05 15:14:55 dep_names[xchat] = ['xchat:1.8.9:1.70.0:1']
> > Jun 05 15:14:55 hits = ['xchat:1.8.9:1.70.0:1']
> > Jun 05 15:14:55 dups = [['xchat', '1.8.9', '1.70.0', '1', 
> 'i386', 1136276]]
> > Jun 05 15:14:55 dups = [['xchat', '1.8.9', '1.70.0', '1', 
> 'i386', 1136276]]
> > Jun 05 15:14:55 Comparing packages: xchat-1.8.7-1.70.0 and
> > xchat-1.8.9-1.70.0 yields -2
> > Jun 05 15:14:55 Comparing packages: xchat-1.8.9-1.70.0 and
> > xchat-1.8.7-1.70.0 yields 2
> > pkg0 = xchat, pkg4 = i386, filename = xchat-1.8.9-1.70.0.i386.rpm
> > Jun 05 15:14:55 Removing xchat-1.8.9-1.70.0.i386.rpm
> > Jun 05 15:14:55 Older package deleted from db.
> > Jun 05 15:14:55 Adding xchat-1.8.7-1.70.0.i386.rpm
> > 
> > dh-i386-7.1 (works):
> > Jun 05 15:19:23 -------------------------------------
> > Jun 05 15:19:23 Examing xchat-1.8.9-1.71.0.i386.rpm
> > Jun 05 15:19:23 dep_names[xchat] = ['xchat:1.8.7:1.71.0:1']
> > Jun 05 15:19:23 hits = ['xchat:1.8.7:1.71.0:1']
> > Jun 05 15:19:23 dups = [['xchat', '1.8.7', '1.71.0', '1', 
> 'i386', 1180276]]
> > Jun 05 15:19:23 dups = [['xchat', '1.8.7', '1.71.0', '1', 
> 'i386', 1180276]]
> > Jun 05 15:19:23 Comparing packages: xchat-1.8.9-1.71.0 and
> > xchat-1.8.7-1.71.0 yields 2
> > Jun 05 15:19:23 Comparing packages: xchat-1.8.7-1.71.0 and
> > xchat-1.8.9-1.71.0 yields -2
> > pkg0 = xchat, pkg4 = i386, filename = xchat-1.8.7-1.71.0.i386.rpm
> > Jun 05 15:19:23 Removing xchat-1.8.7-1.71.0.i386.rpm
> > Jun 05 15:19:24 Older package deleted from db.
> > Jun 05 15:19:24 Adding xchat-1.8.9-1.71.0.i386.rpm
> > Jun 05 15:19:24 
> > 
> > dh-i386-7.2 (doh!):
> > Jun 05 15:21:40 -------------------------------------
> > Jun 05 15:21:40 Examing xchat-1.8.7-1.72.0.i386.rpm
> > Jun 05 15:21:40 dep_names[xchat] = ['xchat:1.8.9:1.72.0:1']
> > Jun 05 15:21:40 hits = ['xchat:1.8.9:1.72.0:1']
> > Jun 05 15:21:40 dups = [['xchat', '1.8.9', '1.72.0', '1', 
> 'i386', 1141240]]
> > Jun 05 15:21:40 dups = [['xchat', '1.8.9', '1.72.0', '1', 
> 'i386', 1141240]]
> > Jun 05 15:21:40 Comparing packages: xchat-1.8.7-1.72.0 and
> > xchat-1.8.9-1.72.0 yields -2
> > Jun 05 15:21:40 Comparing packages: xchat-1.8.9-1.72.0 and
> > xchat-1.8.7-1.72.0 yields 2
> > pkg0 = xchat, pkg4 = i386, filename = xchat-1.8.9-1.72.0.i386.rpm
> > Jun 05 15:21:40 Removing xchat-1.8.9-1.72.0.i386.rpm
> > Jun 05 15:21:40 Older package deleted from db.
> > Jun 05 15:21:40 Adding xchat-1.8.7-1.72.0.i386.rpm
> > 
> > 
> > 
> > But this appears to be logic in current, as the compare 
> functions return the
> > same codes for all 4 platforms and the ones that failed 
> were doing the
> > comparison in reverse... I suspect the 2/-2 return codes 
> (that I haven't
> > seen before, usually 1/-1) aren't being processed correctly?
> > 								-Rob
> > 
> > > -----Original Message-----
> > > From: Hunter Matthews [mailto:[email protected]]
> > > Sent: Friday, March 08, 2002 2:58 PM
> > > To: Current Server Mailing List
> > > Subject: Re: [Current-server] 0.99 vs redhat
> > > 
> > > 
> > > On Fri, 2002-03-08 at 11:51, Vincent Cojot wrote:
> > > > 
> > > > Hello John,
> > > > 
> > > > 	The thing is that we had the same behaviour 
> than RH7.1 but on
> > > > RH6.2.. (rpm-4.0.2 as well)  Isn't that a little too 
> > > restrictive that it
> > > > works -only- in RH7.2? It's too bad that the headers can't 
> > > make their way
> > > > into the database or be used for a decision making process 
> > > before the
> > > > correct rpm is entered in the database..
> > > 
> > > Thats out of Current's control - because doing version 
> comparison is
> > > _so_ complex and fraught with danger, and becuase the 
> server must come
> > > to the same/similar conclusion as the client, Current uses rpmlib.
> > > 
> > > At this point, I'm confused as hell - I dont' see where rpm's code
> > > changed from the rpm that shipped in 7.1 to the rpm that 
> > > shipped in 7.2
> > > (in the version compare routines).
> > > 
> > > John's getting different results from Rob when he does 
> label compares.
> > > 
> > > Can anyone else post log entries (it'll have to be at the 
> maximum log
> > > level, but please trim them down to just the relevant parts) 
> > > of where a
> > > comparison is made and it looks like Current got it _wrong_?
> > > 
> > > We will NOT be adding the headers to the backends - we'd code
> > > versioncompare into Current first, and its not clear 
> thats where the
> > > problem lay yet.
> > > 
> > > 
> > > -- 
> > > Hunter Matthews                          Unix / Network 
> Administrator
> > > Office: BioScience 145/244               Duke Univ. 
> Biology Department
> > > Key: F0F88438 / FFB5 34C0 B350 99A4 BB02  9779 A5DB 8B09 F0F8 8438
> > > Never take candy from strangers. Especially on the internet.
> > > 
> > > _______________________________________________
> > > Current-server mailing list
> > > [email protected]
> > > http://lists.dulug.duke.edu/mailman/listinfo/current-server
> > > 
> > _______________________________________________
> > Current-server mailing list
> > [email protected]
> > http://lists.dulug.duke.edu/mailman/listinfo/current-server
> 
> -- 
> Thank you,
> John Berninger
> 
> Systems Administrator		[email protected]
> Department of Mathematics	Box 8205, Harrelson Hall
> NC State University		Raleigh, NC 27695
> Phone:  (919)515-6315		Fax:	(919)515-3798
> 
> GPG Key ID: A8C1D45C
>         Fingerprint: B1BB 90CB 5314 3113 CF22  66AE 822D 42A8 
> A8C1 D45C
> --
> _______________________________________________
> Current-server mailing list
> [email protected]
> http://lists.dulug.duke.edu/mailman/listinfo/current-server
>