[Sugar-devel] Just wondering

[Martin Abente]

Hello Jerry,

Thanks for the report, I have reproduce it, will revert this commit.

@sam: to reproduce it just run ."/osbuild shell",  and from inside the
shell run "sugar-install-bundle <your-bundle-flle>".

With your patch, and as Jerry suggested, this is what happens:

[osbuild sugar-build-no-broot]$ sugar-install-bundle
/home/tch/Downloads/Abacus-47.1.xo
Traceback (most recent call last):
  File
"/home/tch/Devel/sugar-build-no-broot/build/out/install/bin/sugar-install-bundle",
line 19, in <module>
    registry = get_registry()
  File
"/home/tch/Devel/sugar-build-no-broot/build/out/install/lib/python2.7/site-packages/jarabe/model/bundleregistry.py",
line 652, in get_registry
    _instance = BundleRegistry()
  File
"/home/tch/Devel/sugar-build-no-broot/build/out/install/lib/python2.7/site-packages/jarabe/model/bundleregistry.py",
line 78, in __init__
    self._mime_defaults = self._load_mime_defaults()
  File
"/home/tch/Devel/sugar-build-no-broot/build/out/install/lib/python2.7/site-packages/jarabe/model/bundleregistry.py",
line 147, in _load_mime_defaults
    f = open(os.environ["SUGAR_MIME_DEFAULTS"], 'r')
  File
"/home/tch/Devel/sugar-build-no-broot/build/out/sandbox/install/lib64/python2.7/UserDict.py",
line 23, in __getitem__
    raise KeyError(key)
KeyError: 'SUGAR_MIME_DEFAULTS'

Without the commit works fine.

Lets discuss what other solution we can find!

Regards,
tch.


On Sat, Dec 20, 2014 at 6:28 AM, Sam P. <sam.parkinson3 at gmail.com> wrote:
>
> Hi James,
>
> On Sat, Dec 20, 2014 at 12:37 PM, James Cameron <quozl at laptop.org> wrote:
>
>> Thanks for highlighting that patch, Jerry.
>>
>> Yes, it looks like it would break olpc-os-builder, and would be a
>> hurdle for adopting any release of Sugar that includes the patch.
>>
>
> OK, let's fix this.  Maybe we could run a check to check is sugar is
> running.  If sugar is running, use the registry (so the list is updated).
> If sugar is not running, use bundle.install (like pre-patch).
>
>
>>
>> I'd like the patch reverted and rewritten, and tested without X or
>> DBus present.  That can be simulated by running it in a process with
>> none of the critical environment variables.
>>
>
>
> What is your tip for replicating this environment?
>
>
>>
>> I didn't get any update from the bug system on ticket #4722, which is
>> a surprise.  I also didn't get any notification of a patch available
>> for review.
>>
>
> That's weird.  I did submit a pull request which was merged by Gonzalo a
> little while back (a few days after submitting the pull request).
>
> Thanks,
> Sam
>
>
>>
>> On Fri, Dec 19, 2014 at 06:22:41PM -0600, Jerry Vonau wrote:
>> > Hi Sam,
>> >
>> > > On December 19, 2014 at 5:40 PM "Sam P." <sam.parkinson3 at gmail.com>
>> > > wrote:
>> > >
>> > >
>> > > Hi Jerry,
>> > >
>> > > I would guess that it does require X.  The rational for using
>> > > BundleRegistry is that it emits all the right signals to the gui of
>> > > sugar,
>> > > which requires X.
>> > >
>> > > I havn't tested it against OOB, but maybe I could add a
>> "--no-registry"
>> > > option in sugar-install-bundle?  Would that fix the issue?
>> > >
>> >
>> > I'll defer that to quozl as he is the reporter and the current
>> maintainer
>> > of OOB. I haven't looked at other distros yet, but mktinycorexo would
>> > suffer the same fate. Given the low priority assigned to the bug I'd
>> revert
>> > and not break other scripts that might be out in the wild.
>> >
>> > Just my thoughts,
>> >
>> > Jerry
>> >
>> > > Thanks,
>> > > Sam
>> > >
>> > > On Sat, Dec 20, 2014 at 10:18 AM, Jerry Vonau <me at jvonau.ca> wrote:
>> > >
>> > > >
>> > > > Does the use of BundleRegistry()[1] require X to be started? I'm
>> > > > thinking
>> > > > yes as 'protected_activities' in the __init__ part is really a
>> > > > dconf-service setting that is returned as a string. I'm asking as a
>> > > > recent
>> > > > commit was added[2] that just might break olpc-os-builder(OOB). OOB
>> > > > relies
>> > > > upon
>> > > > sugar-install-bundle[3] running in a chroot'ed cli environment where
>> > > > dconf
>> > > > may not be running. Has this change been tested with OOB?
>> > > >
>> > > >
>> > > > Jerry
>> > > >
>> > > > 1.
>> > > >
>> > > >
>> https://github.com/sugarlabs/sugar/blob/master/src/jarabe/model/bundleregistry.py
>> > > > 2.
>> > > >
>> > > >
>> https://github.com/sugarlabs/sugar/commit/d0d045379f097e35c686e9f5dc7f1a0a231041bd
>> > > > 3.
>> > > >
>> > > >
>> http://dev.laptop.org/git/projects/olpc-os-builder/tree/modules/sugar/kspost.75.install_bundles.inc
>> > > > _______________________________________________
>> > > > Sugar-devel mailing list
>> > > > Sugar-devel at lists.sugarlabs.org
>> > > > http://lists.sugarlabs.org/listinfo/sugar-devel
>> > > >
>> > > _______________________________________________
>> > > Sugar-devel mailing list
>> > > Sugar-devel at lists.sugarlabs.org
>> > > http://lists.sugarlabs.org/listinfo/sugar-devel
>> > _______________________________________________
>> > Sugar-devel mailing list
>> > Sugar-devel at lists.sugarlabs.org
>> > http://lists.sugarlabs.org/listinfo/sugar-devel
>>
>> --
>> James Cameron
>> http://quozl.linux.org.au/
>>
>
>
> _______________________________________________
> Sugar-devel mailing list
> Sugar-devel at lists.sugarlabs.org
> http://lists.sugarlabs.org/listinfo/sugar-devel
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sugarlabs.org/archive/sugar-devel/attachments/20141220/c1fecbaa/attachment.html>