[IAEP] Fwd: [sugar] Pippy VS Develop

[Edward Cherlin]

This post to the Sugar list raises some issues that we should address.
There are patches for major bugs not being reviewed, and there are
major bugs not being fixed. That's apart from the fact that more
developers are needed to work on major Activities.

---------- Forwarded message ----------
From: Jameson Chema Quinn <jquinn at cs.oberlin.edu>
Date: Mon, May 19, 2008 at 5:12 AM
Subject: Re: [sugar] Pippy VS Develop
To: Marco Pesenti Gritti <mpgritti at gmail.com>
Cc: Sugar List <sugar at laptop.org>

On Mon, May 19, 2008 at 4:51 AM, Marco Pesenti Gritti
<mpgritti at gmail.com> wrote:
>
> I'm trying to form my opinion about inclusion of Pippy and Develop in
> Sucrose 0.82.
>
> Here is my understanding:
>
> * Pippy is well maintained and tested but it's supposed to be just a
> stop gap solution. For the future we will be focusing on Develop.
> * Develop is not very mature and tested yet. Jameson is the only
> developer and he is not sure if he is going to be able to keep doing
> it.
>
> I'm sure I didn't get this quite right, so please step in and correct me :)
>
> Marco

You are pretty much right. I would add one point: that currently, the
main flaws in develop are all, really, flaws in glucose (and/or
fructose, see below):

* Cannot open an activity bundle from standard journal, only run it.
(patch exists)
* Cannot have two versions of an activity bundle installed at once
(dev and stable) while debugging - esp. necessary for working on
Develop itself. Also, you are forced to churn the activity.info
version number (upwards or downwards, it doesn't matter) every debug
cycle, because "same version" silently fails to install.
* Activity bundle format is not well-enforced (no warnings), so many
activities have noncompliant bundles; this can lead to data loss in
Develop. (this is what I meant by including fructose above)
* Datastore is unreliable. This led to a major dataloss bug in the
past which "magically fixed itself", and to a separate minor one in
the present (sometimes, save fails non-silently; activity complains,
gives choice of halting exit, and later saves properly) which has not
been resolved. Without a reimplemented datastore, a la tomeu's
quickie, this does not lend confidence.

These are really the only high-priority outstanding issues with
develop as it stands (aside from translation and broader testing). And
the process of code review does not appear to be working for me, I
have a patch for the first of these issues which has languished for
months. From where I'm sitting, it would be nice if checkin capability
on joyride were more liberally granted (with some specific community
norms about self-review), and the strict third-party review
requirement were a filter when cherrypicking changes FROM joyride, not
into it.

As to whether I will be able to continue: hopefully, this will be
resolved within a couple of weeks.

Jameson

_______________________________________________
Sugar mailing list
Sugar at lists.laptop.org
http://lists.laptop.org/listinfo/sugar




-- 
Edward Cherlin
End Poverty at a Profit by teaching children business
http://www.EarthTreasury.org/
"The best way to predict the future is to invent it."--Alan Kay