Bug#719696: marked as done (apt: please improve reported meta-errors)
Your message dated Thu, 22 Aug 2013 17:24:18 +0200
with message-id <CAAZ6_fABPR_Yf69oVH0ng_TGBm6=ztQ6G_w14vsUOXuny1HMTQ@mail.gmail.com>
and subject line Re: Bug#58859: please improve reported meta-errors
has caused the Debian Bug report #719696,
regarding apt: please improve reported meta-errors
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@bugs.debian.org
immediately.)
--
719696: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=719696
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems
--- Begin Message ---
Package: dpkg,apt
Version: n/a
Severity: wishlist
I know this is one of those really annoying bugs to recieve, so I wouldn't
be submitting it, except I see it causing user headaches OVER and OVER and
OVER, and [insert witty comment about Jason Gunthorpe here].
Both dpkg and apt print out meta-errors, i.e. reports of other errors.
Users paste these into channels regularly, and although they do a good job
of letting the user know that it's not working, they don't tell them how to
fix it. (make is the worst offender - make: *** [install] Error 1 gets
pasted more places...)
Neither dpkg nor apt actually tell you what to do to fix a problem like this
(besides wait until some developer fixes the package). They have to consult
a Debian Expert(tm) who has to pry info from them like which script was
failing, then said Debian Expert(tm) has to refer them to
/var/lib/dpkg/info/packagename.scriptname (scriptname doesn't directly
correlate with dpkg's error messages either!), and only *then* can work
start on determining what the problem really was (this can be three emails
into a bug report, if the user chose that route), and by this time a newbie
can really start hating Debian.
Addressing this issue will improve the quality of a significant percentage
of bug reports.
<SuperPhly> anyone ahve any clue...
<SuperPhly> dpkg: error processing xdm (--configure):
<SuperPhly> subprocess post-installation script returned error exit status 1
> superphly: did it print out an error *above* that?
--
Robert Woodcock - rcw@debian.org
"Anybody else wanna negotiate?" -- The Fifth Element
--- End Message ---
--- Begin Message ---
On Wed, Aug 14, 2013 at 1:44 PM, Guillem Jover <guillem@debian.org> wrote:
> Splitting the bug report, because bugs assigned to multiple packages
> are suboptimal to deal with, and only removing the wontfix for dpkg
> because I don't know how the apt team might want to deal with this.
To be fair, I have no real idea what to do about it.
If we are talking about any specific message from APT we could probably
change it, but the bugreport doesn't mention any.
Then it comes to messages dpkg shows: I believe APT shouldn't suggest any
kind of "solution" for dpkg errors. dpkg has just way better information
to do that. And frankly, I have no idea what these "solutions" should be
in the general case.
The only suggestion I am reading very often is to add an "exit 0" at the
top of the maintainer script, but while this might be an acceptable
suggestion from a human who has looked at what the script does, letting
a machine suggest this by default is probably not a good idea as usually
the scripts who are buggy are also the scripts which need to be run to
ensure a successful transition…
So as I don't see any obvious point for improvement here I am therefore
closing this bug as being "too meta". If there are any specific examples
feel free to reopen/report new bugs (but please check for duplicates first),
but currently keeping this one here open would just ensure that we have yet
another bug in APT which is never closed as no goal can be achieved.
Best regards
David Kalnischkies
--- End Message ---
Reply to: