Created attachment 162465 [details]
Screenshot of wrong free space calculation

I'm also affected by this issue too but I detected it selecting w32codec-all package (from Packman repo) to install.

*** Bug 308446 has been marked as a duplicate of this bug. ***

fix? workaround?

how is this better than beta 2? at least i could install software.

i noticed that the sax2 bug still exists if using a logitech g5 gaming mouse - on the bug list for months now.

Please, attach the YaST logs.

Created attachment 162857 [details]
y2log from sw_single session

Same here when selecting packages from the KDE:KDE4 repo.
Or when selecting the kernel-source package from the Updates repo.

Seems to be a problem with rpm-md repos...

I have the same - yast reports that kde4 pattern will require >30 GB of disk space. The issue is the same with rpm-md and yast2 factory repo.

Elevating severity to critical

Just for the record; is it the same problem as described in:
https://bugzilla.novell.com/show_bug.cgi?id=222556 ?

If so, it would mean that the above bug is not solved....

*** Bug 309052 has been marked as a duplicate of this bug. ***

As suggested in comment #9, this issue occurs in both rpm-md and yast repos, so it seems to be in the disk usage code. Michael?

*** Bug 308475 has been marked as a duplicate of this bug. ***

I have fixed the problem with wrong units in SVN, see bug #308475.

But I have found another problem, selecting a package for update reports the same final disk usage size when I select the package for removing. It seems that update is wrongly counted (update = remove A, install B, probably only remove A is counted).

(In reply to comment #16 from Ladislav Slezak)
> I have fixed the problem with wrong units in SVN, see bug #308475.
> 
> But I have found another problem, selecting a package for update reports the
> same final disk usage size when I select the package for removing. It seems
> that update is wrongly counted (update = remove A, install B, probably only
> remove A is counted).
> 

FWIW: the disk space calculation in beta-2 was correct for my test updating from 10.1 to 10.3 beta-2 (I did not really update for other reasons).
trying to update the same systems to beta-3 now gives wrong disk usage...

I have partly fixed the problem in libzypp SVN rev. 7165.

The problem which remains is that disk usage for installed package is empty. But when a repository provides the package then it's disk usage is used as a fallback. If there is no source providing the package DU will be wrong if the package is selected for removal.

*** Bug 309384 has been marked as a duplicate of this bug. ***

attention!!!

this problem also seems to be having a HORRIBLE impact on kmail as well

http://x.sodpit.com/screenshots/x49.png

kmail was stable in betas 1 and 2, now it crashes constantly!

HELP!!

should i file a separate report??

(In reply to comment #22 from darren winter)

> should i file a separate report??


Yes KMAIL has absolutely nothing to do with zypp ,hence, with this report. 

reported, https://bugzilla.novell.com/show_bug.cgi?id=309404

*** Bug 309745 has been marked as a duplicate of this bug. ***

Fixed (or at least improved) in libzypp-3.23.0.

Susetags, plaindir and rpmdb now provide more detailed disk usage information.
Yum metadata don't, so we book the package size to '/'.

So, do we have a work-around? 

I get this on one of my test systems, though not on the other - I can't get past the "Error" dialogue.

10.3 is looking pretty good otherwise...

To comment 28: Try RC1, it's fixed there.

rc1...  i am only about 12 hours in, but so far this bug hasn't shown up yet in rc1.

Darren, thank you for testing! It's nice to hear a positive feedback!

...same results here.  Just installed a TON of packages in RC1 and didn't get the error at all!  :-)

no problems at all here, awesome!

well, when i use truecrypt to mount a 20 gb image to a local folder then i start to see some warnings, but hey, that's my problem! :>