[Debian-olpc-devel] [Sugar-devel] Assorted Debian+Sugar bugs

[Jonas Smedegaard]

Hi again,

On Tue, Oct 13, 2009 at 07:37:07PM -0400, Michael Stone wrote:
>>On Mon, Oct 12, 2009 at 07:52:11PM -0400, Michael Stone wrote:
>>>Jonas & co:
>>>
>>>I just wanted to report a couple of regressions that I found today 
>>>while trying out sugar-0.84 on Sid. In no particular order:
>>
>>Thanks for reporting this.
>
>You're very welcome; thanks again for your hard work packaging sugar!
>
>>Even better, however, is if you file issues like this as proper 
>>bugreports.
>
>I haven't diagnosed the causes of these issues so I don't actually know 
>what packages are at fault yet; I'm just reporting behavioral 
>regressions. Consequently, I don't know what packages to file the bugs 
>against. :)

I understand, but don't let that discourage you: file the issues, as 
separate as you feel they make sense to split, against some - any - 
package you feel is related to the issue.

It is easy later on to reassign to another package if that turns out to 
make more sense.  And it is easy to merge issues together (splitting is 
possible too, but little more work).

Problems experienced using a specific activity are probably best filed 
against the package containing that activity (e.g. sugar-pippy-activity 
or sugar-browse-activity-0.84).

Only package that would be "bad" to file against is perhaps that 
education-desktop-sugar package, as I would not notice bugs filed there 
and I suspect noone else will either.


>>I will respond here, cross-posted to both list, to respect your choice 
>>of communication platform.  But there is a higher risk that your 
>>information will not get tracked and the issues not solved by doing it 
>>this way.
>
>By all means, feel free to enter the information into the tracker of 
>your choice. I will be happy to follow your lead. I simply do not wish 
>to file reports with faulty information.

I can do that, but a) it is more work, b) it is more polite to play by 
the communication style of the initiator, causing me to *both* have this 
thread going *and* start a new one through the BTS which risks this 
thread stealing focus from the other one and thus c) draining attention 
away from where bug solving happens.

(exact same is most probably true about Sugarlabs and OLPC bug trackers, 
and I am guilty of not engaging very well with those either)

The "more work" a) it not (only) that work is passed from you to me, but 
that the "reportbug" tool bootstraps new bugreports elegantly, attaching 
you as bugreporter, gathering valuable details about your system, and 
guiding you through a few questions to improve effective processing.

Thing is, filling in data is only part of the process.  The more 
valuable part (in many cases) is a continued dialogue with the 
bugreporter - and with others chiming in later with additional info, 
something that is encouraged by discussion-style communication which 
feels natural in emails.

A thing that I love about the Debian BTS is that it is email-based, 
dynamically creating a "mailinglist" for each bugreport.  This makes 
ongoing discussion with the bugreporter quite fluent, unlike (IMHO) most 
web-based systems.  Others less fluent with email may obviously disagree 
with me on that, but I suspect most _developers_ are fluent in email, 
and in my opinion issue tracking should serve developers more than 
users. A user-friendly pendant to an issue tracker would IMO be a 
different thing, one that e.g. allowed users to "let off steam" even if 
irrelevant to the developers (similar to an issue tracker dealing with 
details of absolutely no interest to the user).


>You may, however, be interested to know that, as of yesterday, 
>sugar-0.86 depends on substantially more material than sugar-0.84. (To 
>the tune of 300 MB more, I think.)

Oh my :-(

A few days ago I installed a smallest-possible install of (the 
Debian-packaged parts of) Sugar 0.86.  I am not yet satisfied with the 
configurations, and have not gotten rid of some old cruft, but I believe 
it required less than 200MB, including X11, GNU and BSD utilities and 
the kernel.  If stripping unused locales I can probably save another 
50MB.

Maybe I am oldfashioned, but I still find it a worthy challenge to 
squeeze a full system into 256MB or 512MB USB sticks.


>>>P.S. - Please also find attached the output of "dpkg -l" run from 
>>>inside my testing chroot. This chroot was constructed by the code in 
>>>the "sugar" branch of
>>>
>>> http://dev.laptop.org/git/users/mstone/puritan
>>>
>>>with conf/distro == debian and conf/debian/distro == sid.
>>
>>Ahem, does this mean that you did not in fact use a Debian system but 
>>some homecooked lookalike?
>
>I test sugar in chroots so as to be able to efficiently generate 
>reproducible results across multiple distros. My debian chroots are 
>constructed with debootstrap, as is apparent in the "debian.mk" 
>Makefile in source code that I mentioned.

Sounds interesting (as I would expect from you!)  I do still, however, 
recommend you to mention any such "specialties" about your environment 
when filing bugreports.  Yes, some developers may react in the line of 
"get lost, you don't use a pristine system!", but others (like me) will 
still appreciate your input but will be helped to know that some issues 
perhaps are triggered only under odd circumstances.

Finding out _later_ about oddities of an environment is almost certain 
to cause grumpy developers, no matter if it turned out to be relevant 
for the particular issue dealt with: you feel cheated if not openly 
provided all potentially relevant data.


Kind regards,

  - Jonas

-- 
* Jonas Smedegaard - idealist & Internet-arkitekt
* Tlf.: +45 40843136  Website: http://dr.jones.dk/

   [x] quote me freely  [ ] ask before reusing  [ ] keep private
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.alioth.debian.org/pipermail/debian-olpc-devel/attachments/20091014/0654bfa9/attachment-0001.pgp>