[Bug 221476] New: zypper up --skip-interactive (used in automatic online update) makes zypper exit
https://bugzilla.novell.com/show_bug.cgi?id=221476 Summary: zypper up --skip-interactive (used in automatic online update) makes zypper exit Product: openSUSE 10.2 Version: Beta 2 plus Platform: Other OS/Version: Other Status: NEW Severity: Major Priority: P5 - None Component: YaST2 AssignedTo: mvidner@novell.com ReportedBy: mvidner@novell.com QAContact: jsrain@novell.com CC: jsuchome@novell.com, jkupec@novell.com yast2 online_update_setup uses --skip-interactive if the corresponding checkbox is selected (it is selected by default). In zypper, this option is not implemented yet which makes zypper abort. 30 21 * * * root zypper up -y -t patch --skip-interactive And the same with -d (download only) but that is not selected by default. Actually for -d exiting if not implemented seems to be the preferred behavior. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=221476 mvidner@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jsrain@novell.com, ma@novell.com Status|NEW |ASSIGNED ------- Comment #1 from mvidner@novell.com 2006-11-16 04:38 MST ------- There is bool zypp::Patch::interactive() which is intended to do what we want: skip selecting these for "update" (maybe mark them specially in "list-updates"?). It checks whether a reboot is needed, whether there are messages. And there is also a check for package licenses, but it is commented out (with FIXME). Why? Svn ann says the code is ancient and written by jsrain. The code for interactive() is present in three(!!!) copies: detail/PatchImpl.cc source/yum/YUMPatchImpl.cc and target/store/xml/XMLPatchImpl.cc. Never mind the last one, already installed patches are not interesting in this respect, but I am really confused why the duplication is in detail and source/yum. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=221476 ------- Comment #2 from ma@novell.com 2006-11-16 05:35 MST ------- Because it's so easy to cut'n'paste ;( The interactive code seems to contain no implementation specific details. It just analyzes properties of public available items. Obviously this should be the default implementation provided by the interface class PatchImplIf. I'll relocate the code and remove the superfluous overloads. But I don't know why the license test is disabled. Jiri? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=221476 ------- Comment #3 from ma@novell.com 2006-11-16 06:13 MST ------- Had a closer look at the code. The way it is implemented, it does not work: Interactive wants to iterate all not_installed Atoms. In fact not_installed_atoms returns all Atoms. This can't be fixed, because a Resolvable has no State. State is maintained by the ResPool. A Resolvable is just a data provider. A not_installed method in a Resolvable is nonsense, same any_atom_selected and mark_atoms_to_freshen. We should remove these methods ASAP from the patch interface. A method Patch::intercactive could return whether a Patch contains at least one interactive Atom. But if you want to know whether an intractive Atom is part of a transaction (e.g. to be installed), you have to query the Pool, not the Resolvable. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=221476 ------- Comment #4 from mvidner@novell.com 2006-11-16 06:21 MST ------- huh? It does not need to care about ResStatus. I intend to filter for Needed patches beforehand, that's fine.
We should remove these methods ASAP from the patch interface. I hope that you mean not_installed_atoms() and not interactive().
-- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=221476 ------- Comment #5 from ma@novell.com 2006-11-16 06:29 MST ------- Yes. interactive (as described above) is something a Patch can offer. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=221476 ------- Comment #6 from ma@novell.com 2006-11-16 15:50 MST ------- I cleaned the Patch code in libzypp-2.7.5. intercactive() is now defined once in detail/PatchImplIf, and used by specialized Patch implementations. Interactive returns true, iff Patch needs reboot, contains a Message or an Atom with licenseToConfirm. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
https://bugzilla.novell.com/show_bug.cgi?id=221476 mvidner@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED ------- Comment #7 from mvidner@novell.com 2006-11-20 10:11 MST ------- --skip-interactive is fixed in zypper-0.6.13, tested with [patch]kernel-2285-0.noarch -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.
participants (1)
-
bugzilla_noreply＠novell.com