From Sugar Labs
< 0.86
Revision as of 06:22, 22 May 2009 by Sascha silbe (talk | contribs) (→‎Proposal Goals: merge some more of the "old items")
Jump to navigation Jump to search

Team Home   ·   Join   ·   Contacts   ·   Resources   ·   FAQ   ·   Roadmap   ·   To Do   ·   Meetings

Sucrose Development

Sucrose 0.85.x is an unstable development series intended for testing and development purposes. Sucrose uses odd minor version numbers to indicate development status, so this unstable 0.85.x series will eventually become the 0.86 stable release.



Date Task Notes
2009 Jun 01 Release goals proposal
Jun 29 New modules proposal
Jul 29 Sucrose 0.85.1 Tarballs Due
Jul 30 Sucrose 0.85.1 Development Release
Sucrose 0.85.2 Tarballs Due
Sucrose 0.85.2 Development Release
Sucrose 0.85.3 Tarballs Due
Sucrose 0.85.3 Development Release
Aug 20 Sucrose 0.85.- Tarballs Due
Aug 21 Sucrose 0.85 Beta 1 (0.85.-) Feature, API, String freeze
Sucrose 0.85.5 Tarballs Due
Sucrose 0.85 Release Candidate 1 (0.83.5)
Sucrose 0.85.6 Tarballs Due
Sucrose 0.85.6 Release Candidate 2 (0.85.6)
Sep 17 Sucrose 0.86 Tarballs Due Hard code freeze
Sep 18 Sucrose 0.86 Final Release!

Glucose Development Team/Release/Modules

Fructose Development Team/Release/Modules

Glucose Dependencies

Fructose Dependencies

  • pyabiword
  • hulahop

Proposal Goals

Switch to a standard WM

Switch to a standard compliant window manager (possibly Metacity)

New toolbar widget


  • tabs support (open popup windows in tabs, saving of tabs history, standard behavior of Browse should not change -> no open tab by default)
  • better naming of files to be uploaded (change temp name to something based on the title)
  • export for offline viewing (Web page - HTML only, Web page - Complete)
  • creating of web pages (highlighting support in Write, Activity with special HTML based features(can happen outside of the official cycle))
    • bookmarks (global bookmarks, at the moment we only have session bookmarks and the autocompletion functionality)

Tags in the Journal

  • auto completion for already existing tags, tag clouds
    • its implemented in Library activity as well

More Accelerators (short cuts)

  • make sure we use the accelerators where possible, get discussion about which modifiers to use for which settings as early as possible going

Printing support

Search in home view

  • the search is recently builds disabled


Flash activities


  • tagging buddies to build up relations, tagging can happen by a teacher tagging a class or the learner can tag himself

List views

  • in the mesh view a list view of the access points
  • switching to use gtk-tree-view for the lists (journal, activity) - this has accessibility support already

Mesh View

  • use buddy color to seed the position to get a more stable positioning in the mesh view

Ad-hoc networking

  • as an alternative to the mesh


Keyboard control panel extension

  • (Sayamindu would prefer to keep it as a seperate module, since everyone may not choose to go with XKB)

Dictionary support in the shell

  • link to email here

CP - Language in native language

  • link to ticket here


  • de-couple the portfolio from the base TA

Library activity

Alsroot 16:16, 21 May 2009 (UTC)

API work

  • backwards compability
  • (http://wiki.sugarlabs.org/go/Development_Team/sugar-port will take care for the activities needs for a working activity on 0.82 for example) see "Decoupling of Sucrose" section Alsroot 16:16, 21 May 2009 (UTC)
  • (?) start moving to pure dbus interface for glucose
    • dbus method to resume/open activities(instead of sugar-toolkit API)
      • something simple like dbus_object.activate(uid) or
    • etc.

Decoupling of Sucrose

The major idea is to have tough core(with stable release cycle) <=> dbus-API/sugar-toolkit-API <=> unlimited count of activities that uses core functionality and do not follow 6 months release cycle(which could be overmuch for activity).

It could looks like:

  • core - glucose, six months(or so) release cycle, w/o any activities only API
  • bridge - sugar-port for example, between all(in ideal) already deployed sugars and activities i.e. it provides backwards compatibility(so the same activity code will work on all sugars) and at the same time provides features from newest sugar(so the same activity code will use last sugar's features)
  • world - the rest of sugar world i.e. fructose/honey (but now there is no differences between them) that use core directly, if all deployed sugars have the same API for desired functionality(for example in case of preselected mime type, ObjectsChooser has different API for 0.82-0.86), or use bridge otherwise.
    imho another point to have activities outside of core release cycle - activities have more shorter release cycle then core has

And of course deployers can form any sets from these components

Alsroot 16:14, 21 May 2009 (UTC)

Version support for datastore

See Version_support_for_datastore/Proposal.

Datastore features that could benefit Library

These features can benefit Activities/Library activity. At present(for 0.82/0.84) Library uses "rich" datastore client which fetches all Journal objects to make short list of objects(by collapsing all versions to one item), moreover it has to unzip .xo bundles to get bundle_id value. So its not good.

With all these features implemented Library can use "thin" datastore client and not fetch all objects, just make proper request for datasotre.find() to get sorted and collapsed objects by portions (by using offset/limit find() arguments).

Get collapsed find()'s resultset

Let user call datastore.find() to get list of collapsed objects(all versions by one item).

That should mean that any datastore object should have "object id":

  • activity_id for objects that were generated by activities (already implemented)
  • bundle_id for .xo bundles in the Journal
  • uid for other objecets (already implemented)

History of collapsed object

Let datastore clients browse history of collapsed objects:

  • for activity_id objects - sort by timestamp (already impelemented)
  • for bundle_id objects - sort by activity_version(or so) field

Sort find()'s resultset

Let datastore clients sort objects by any field(not only predefined like uid, activity_id etc).

Sort by buddies count


Add support for the current version of Rainbow so

  1. security flaws in Browse (which are the norm rather than the exception) don't allow an attacker to take over the machine and
  2. programming mistakes cannot affect other activities (like deleting all Journal contents).


Add some automated testing to catch bugs earlier and ease integration work:

Old items

Proposed modules