lacks dependency on python-metakit

Bug #269936 reported by Luis Villa
8
Affects Status Importance Assigned to Milestone
Gourmet
Invalid
Undecided
Unassigned
gourmet (Ubuntu)
Invalid
Low
Unassigned

Bug Description

Tracebacks on first startup because python-metakit is not installed; should have python-metakit as one of the dependencies.

[update] This is not the case, metakit is not in modern Ubuntu, depending on it is broken behaviour, attaching revert patch

Tags: bitesize
Changed in gourmet:
assignee: nobody → nick.ellery
status: New → In Progress
importance: Undecided → Low
Revision history for this message
Nick Ellery (nick.ellery) wrote :

gourmet (0.14.0-1ubuntu1) intrepid; urgency=low

  * Added python-metakit to dependencies (LP: 269936).
  * Modify Maintainer value to match the DebianMaintainerField
    specification.

 -- Nick Ellery <email address hidden> Sat, 13 Sep 2008 16:56:17 -0700

Revision history for this message
Nick Ellery (nick.ellery) wrote :

New debdiff

Changed in gourmet:
assignee: nick.ellery → nobody
Changed in gourmet:
status: In Progress → Fix Released
Changed in grecipe-manager:
status: New → Invalid
Revision history for this message
Rolf Leggewie (r0lf) wrote :

I'm not sure this is right. Are you, Nick? python-metakit in ubuntu is ancient and all I can say about this bug is that the package runs fine for me on hardy without it.

Changed in gourmet:
status: Fix Released → Incomplete
Revision history for this message
Nick Ellery (nick.ellery) wrote :

Hm, that is a good point Rolf. python-metakit was last in Gutsy, so perhaps it shouldn't be added.

Revision history for this message
Luis Villa (luis-villa) wrote :

Hrm. I do indeed have gutsy sources still in my sources.list (though /etc/lsb-release says I'm running hardy.) That said, gourmet definitely was failing (tracebacking on startup) before I installed python-metakit, and now works. (well, 'works', see bug 269935.)

So maybe a dependency on metakit isn't right, but it either needs some replacement for metakit, or it needs to be removed, since it doesn't work.

Revision history for this message
Rolf Leggewie (r0lf) wrote :

Nick, I get the impression that Luis does a lot of and loud handwaving. Please be assured that the package works just fine for me without any of the stuff he proposes. I get the feeling his error might be "self-inflicted", possibly by incorrectly mixing package versions.

Nick, as Debian package maintainer for gourmet, I strongly suggest you reverse your change and consult with me first on any future changes. Let's not fuel the preconception in some debian circles that ubuntu takes some perfectly fine packages and patches them to a broken state which is just what adding the dependency on python-metakit did.

That being said, I am of course interested in fixing any problems with gourmet packaging if we can find out where they really come from.

Changed in gourmet:
assignee: nobody → r0lf
status: Incomplete → Invalid
Revision history for this message
Rolf Leggewie (r0lf) wrote :

I suggest to close this bug as invalid (the request for adding the dependency is invalid and results in a broken, uninstallable package) once the dependency has been removed from the package in the feed.

Let's move discussion of gourmet startup problems to bug 269935.

Changed in gourmet:
status: Invalid → In Progress
Rolf Leggewie (r0lf)
Changed in gourmet:
assignee: r0lf → nick.ellery
Revision history for this message
Luis Villa (luis-villa) wrote :

You're right that the suggested fix results in an uninstallable package, so we should continue this discussion and attempt to get to the root of it, but it is a distinct bug from 269935. I'm running all the most up-to-date packages from Hardy and ran a dist-upgrade again to check last night, removing all remnants of gutsy from my apt sources.list. Versions (at least of gourmet) aren't mixed, as gourmet had never been installed until about 30 minutes before this bug was reported. I'm happy to do whatever is necessary to clarify my dependencies and figure out what you've got installed that I don't (or vice-versa), and why.

If there is a problem here, you should be working with me to fix it, not pulling the old 'well, it works for me' trick. That's a very old, very boring, very unconstructive 'solution'.

(As with regards to being someone who 'does a lot of loud handwaving', well, google me. I've constructively dealt with and resolved more bugs on good days than your launchpad record indicates you having ever dealt with, period. Yes, this bug irritates me, and yes, I'm behaving slightly petulantly here, but yes, you're shipping software that appears completely broken in at least two separate ways on at least one user's computer, and as usual launchpad doesn't even offer me the option to do the right thing (set the proper priority/severity, much less get attention for it.))

(And no, I haven't run Debian in since about 1999. So that's not my pet peeve, but you can continue making up my motivations if it makes you feel better.)

Revision history for this message
Rolf Leggewie (r0lf) wrote :

Luis, get your facts straight is all I can say. And say again. I don't have time for your childish behaviour or to correct your mistakes, problems and misunderstandings. Welcome to my ignore list.

We work on this in our free time, so I don't understand what you try to accomplish by raising priority. It's not like your paying someone and you could force them to fix any problem that *you* have. You can send a patch, of course. But I hope that Nick will check back with me before applying any of your quick-shots next time around.

You're welcome to "continue this discussion and attempt to get to the root of it", just don't count me in for triaging your bugs. Maybe some other kind soul will help. From me, WFM is the best you'll get from now on. In any case, this bug is titled "lacks dependency on metakit" and that is invalid as you admitted. I also don't see any other issue being addressed here and "one issue, one report" should be familiar to a bug superhero like yourself. Expect this report to be closed soon.

Revision history for this message
Rolf Leggewie (r0lf) wrote :

$ dpkg -l *metakit*
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Installed/Config-f/Unpacked/Failed-cfg/Half-inst/t-aWait/T-pend
|/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad)
||/ Name Version Description
+++-=========================-=========================-==================================================================
un libmetakit-python <none> (no description available)

Jo Shields (directhex)
description: updated
Revision history for this message
Daniel Holbach (dholbach) wrote :

Nick: the python-metakit change needs to be reverted, at least for now. This is what happens in intrepid:

We're in a funny situation here: according to http://people.ubuntu.com/~ubuntu-archive/removals.txt libmetakit2.4.9.3 was removed from hardy because it was orphaned in Debian. At the time when it was removed this was possible because there was no package depending on it (gourmet didn't have the depends back then).

Now either we need to see if it's feasible to bring libmetakit2.4.9.3 back (which looks like a tough job judging by the number of bugs saying that it failed to build somewhere) or ask gourmet upstream if there is no alternative to libmetakit.

A brief glance through the code suggest that there are alternative code paths almost everywhere and that metakit should not be a hard requirement. Luis: can you please let us know what the problem/traceback exactly is, when python-metakit is not installed?

Revision history for this message
Jo Shields (directhex) wrote : Re: Broken dependency on metakit
Revision history for this message
Jo Shields (directhex) wrote :
Revision history for this message
Daniel Holbach (dholbach) wrote :

And please: all calm down. There's a lot of misunderstanding in the bug and once we have more data about what's happening without metakit, we can take a closer look at it.

Revision history for this message
Daniel Holbach (dholbach) wrote : Unsubscribing Sponsors.

I'm unsubscribing the Sponsors' Team for now. Please re-subscribe when ready.

Revision history for this message
Daniel Holbach (dholbach) wrote : Re: Broken dependency on metakit

directhex: can you please remove the " (LP: 269936)" from debian/changelog? I'm afraid the upload does not fix the underlying issue.

Revision history for this message
Jo Shields (directhex) wrote :

There's no obvious underlying issue that I can see. The app runs fine for me without metakit. Starts fine, anyway - not sure how to use the thing

Revision history for this message
James Westby (james-w) wrote :

Hi,

The CHANGES file has

  ==== Version 0.10.0 changes cince 0.9.0.3 ====
  * Switch database backend to SQLite. The backend is still abstract, so
    if in the future someone wants to switch back to metakit or to
    something else, it should be simple.

and the README has

  In earlier versions of Gourmet, the default backend was
  metakit. Metakit databases from previous versions should be
  automatically converted by newer versions.

Luis, was the database you wish to use with gourmet created with an earlier
version of gourmet?

Thanks,

James

Revision history for this message
James Westby (james-w) wrote :

If it was then you will have a recipes.mk database, as opposed to a
recipes.db one. I believe these live in ~/.gourmet.

Thanks,

James

Revision history for this message
Thomas M. Hinkle (thomas-hinkle) wrote :

Not sure what all this was about -- as James saw, metakit is only needed to convert old databases to SQLite.

It's good to see all this activity around packaging -- for what it's worth, my effort as the lead developer on Gourmet is entirely in the "unstable" tree, which will obsolete the 0.13.x branch soon enough (well, not soon enough -- I have a 10 month old at home and a day job that keeps me more than busy... but you know how that goes)...

Revision history for this message
Rolf Leggewie (r0lf) wrote :

+--[ changelog excerpt ]
|
| gourmet (0.14.0-1ubuntu2) intrepid; urgency=low
|
| * Revert broken dependency on Gutsy package.
|
| -- Jo Shields <email address hidden> Tue, 23 Sep 2008 14:06:17 +0100
|
+--

Thanks, directhex for reverting that erroneous change. Setting bug to invalid.

Changed in gourmet:
status: In Progress → Invalid
Revision history for this message
James Westby (james-w) wrote :

Rolf, shouldn't the bug not be set to Invalid, but changed to explain
the real problem, that gourmet doesn't fail gracefully when the user
has a metakit db, but no metakit installed?

Thanks,

James

Revision history for this message
Rolf Leggewie (r0lf) wrote :

no, as I believe I explained previously that is what bug 269935 shall be about. Too much noise in this one, so it is better to understand the scope of this item strictly as "please add dependency XY".

The task of having a better error message is not being lost.

Revision history for this message
James Westby (james-w) wrote : Re: [Bug 269936] Re: lacks dependency on python-metakit

On Thu, 2008-09-25 at 12:00 +0000, Rolf Leggewie wrote:
> no, as I believe I explained previously that is what bug 269935 shall be
> about. Too much noise in this one, so it is better to understand the
> scope of this item strictly as "please add dependency XY".
>
> The task of having a better error message is not being lost.
>

Great, thanks.

James

Changed in gourmet:
assignee: nick.ellery → nobody
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Bug attachments

Remote bug watches

Bug watches keep track of this bug in other bug trackers.