pkgsrc-Bugs archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
pkg/51417: pkgin should put all interaction up-front
>Number: 51417
>Category: pkg
>Synopsis: pkgin should put all interaction up-front
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: pkg-manager
>State: open
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Wed Aug 17 07:40:00 +0000 2016
>Originator: he%NetBSD.org@localhost
>Release: NetBSD 7.0
>Organization:
>Environment:
System: NetBSD smistad.uninett.no 7.0 NetBSD 7.0 (GENERIC.201509250726Z) amd64
Architecture: x86_64
Machine: amd64
>Description:
pkgin can in at least two cases ask interactive questions at
... "inconvenient times":
1) During "pkgin fug" it may have done lots of downloads
before it discovers that the package repository doesn't
have either the "xv" program nor the "adobe-flash-plugin"
package (e.g. when one of the dependencies is going to be
upgraded).
pkgin will then abort the download, and at that point
forcing the operator to manually remove the package
(possibly after pkg_tarup) before re-doing "pkgin fug".
Instead, pkgin should discover this up-front before
starting the download phase, and possibly refuse to proceed
before the problem is fixed as suggested above.
This so that the operator can do away with all the
interaction up front and then proceed to let pkgin complete
the task on its own without the need for further
interaction.
2) During a "pkgin fug", when upgrading to a new pkgsrc
branch, it's likely that the pkg_install package will need
to be upgraded. However, pkgin insists on an
interactively-given permission for upgrading this package,
and insists on that interaction before upgrading this
particular package, which, it appears, is always the first
package to be upgraded (if it will be upgraded).
This happens after a long phase of non-interaction
(download of new packages and removal of all packages which
will be upgraded). Somehow pkgin appears to consider
pkg_install a "special" packge that it needs special
permission to upgrade (why?!?).
Meanwhile the system keeps sitting there with possibly
precious few packages installed, probably in a state of
strictly reduced functionality compared to what the
operator intended.
I would primarily suggest that the interactive promt for
upgrading pkg_install be abolished altogether (what happens
if you say "no"? Does it abort the whole upgrade?).
Alternatively if the prompt is retained, I would suggest
that the question be asked *before* the download of the new
packages is started, so that it happens at a time when the
operator has attention.
>How-To-Repeat:
Ref. description above.
>Fix:
Suggested fixes of behaviour given above. No code, though...
Home |
Main Index |
Thread Index |
Old Index