Difference between revisions of "Potential Projects"

From CDOT Wiki
Jump to: navigation, search
Line 182: Line 182:
  
 
Create an HTTP facility (transport, not rendering) for desktop apps that shares cache and cookies with the browser -- basically export the firefox http engine. A dbus-exported [http://svn.mugshot.org/dumbhippo/trunk/client/linux/src/hippo-dbus-http.c proof-of-concept] exists which which shows some basic structure. A better solution would support more http features, avoid the dbus daemon for large files, and use Firefox instead of libcurl. One challenge is making this work when Firefox is not running. Resources: RH online desktop team.
 
Create an HTTP facility (transport, not rendering) for desktop apps that shares cache and cookies with the browser -- basically export the firefox http engine. A dbus-exported [http://svn.mugshot.org/dumbhippo/trunk/client/linux/src/hippo-dbus-http.c proof-of-concept] exists which which shows some basic structure. A better solution would support more http features, avoid the dbus daemon for large files, and use Firefox instead of libcurl. One challenge is making this work when Firefox is not running. Resources: RH online desktop team.
 
 
 
== Localized Search in Firefox Search Box ==
 
 
Firefox features a search box that can be customized to work with different search engines (e.g., Google, Yahoo!, Wikipedia). Most websites have their own search facilities, but the location of the search box on the page varies considerably. Create a mechanism for Firefox to pick up a website's search mechanism and make it accessible to the user through the Firefox search box, so that users have a consistent place to look for site-specific search. This might use a mechanism similar to [http://www.rssboard.org/rss-autodiscovery feed autodiscovery].
 
 
Resources: myk
 
  
 
== Universal Firefox on a USB Key ==
 
== Universal Firefox on a USB Key ==

Revision as of 11:26, 24 September 2007

Introduction

This is a list of potential projects that need people. If you'd like to work on one of these, move the chosen project to the Project List and create a page for your work based on the Sample Project template.

Projects

Sample Project

This is a sample project stub. You can use the template for Sample Project in order to create a project page for one of the stubs below. This is how you 'sign-up' for a project.

NOTE: if someone has already created the project page, speak to this person and see if you can join them. If so, simply add your name to the Project Leader(s) page. Otherwise, you can become a contributor later.

Mozilla Source and Symbol Server

Create a source code server to work with Mozilla's current symbol server. This will allow users to do full debugs of nightly release builds with access to remote symbols and full source code.

Port canvas3d to non-Windows platforms

The canvas3d extension is a binary C++ component that adds a 3D OpenGL drawing context to the HTML canvas element. Currently canvas3d works only on Windows. Port the code to Linux and/or Mac OS X. Porting involves implementing a pbuffer class for each new platform. Some familiarity with OpenGL, or a willingness to learn it, is required.

Resources: Cathy, Vlad.

Add OpenID support to Bugzilla

Many open source projects rely on bugzilla for bug tracking, and open source developers use different instances, forcing them to have multiple logins. Complete OpenID support in bugzilla.

References: see bug 294608

Full-Text History Search extension

Firefox3 probably won't ship with support for full-text history search, even though url-bar autocomplete (with urls/titles) is a really great feature/idea. Write an extension in order to add this feature.

References: http://skrul.com/blog/projects/xpclucene/

Mozilla Tree Visualization

Use the canvas element in order to create a generic front-end for data visualization of the Mozilla source. Many types of information about the Mozilla project can be keyed to the source tree. For example, file change-rates, code-coverage, bug activity per module, checkins per module, etc. The visualization will be a heatmap, showing certain types of activity in the tree. Doing this visualizaiton using canvas3d would allow for a third axis, and changes over time to be shown.

Mozilla Data Visualization Back-ends

Create a data-source for the Mozilla Tree Visualization project. This means data mining cvs.mozilla.org, bugzilla.mozilla.org, bonsai.mozilla.org and preparing the data for use in a heatmap visualization.

Improve Mozilla Application and Platform Documentation

Not enough is understood about the various communities and companies who develop with Mozilla, as opposed to those who work on the Mozilla code itself. A proper survey needs to be conducted in order to find out which APIs are being used and how, what is hard, what could be improved, etc. The Mozilla Platform Application Developers is interested in changing this. Their goal is to work with the Mozilla development community to survey, understand, and document the current state of the art. In this project you will work with the mozpad community to connect with projects/companies in orbit around Mozilla and do interviews with their developers. You will then document this information.

Localize Mozilla

Join a localization team (see http://developer.mozilla.org/en/docs/Localization) and help localize Mozilla into a language other than English, for example, Chinese.

One recent suggestion is to do en-CA and fr-CA localizations for Canadian users.

  • References: Mic, Alix

My Favicons extension

Write an extension to allow the user to pick his/her own favicons for sites, based on a URL regexp. The image should get stored as part of the user’s profile, so that it doesn’t have to live in the same place on the user’s disk. Ideally, it should also be possible to export the regexp -> icon mapping as a regexp -> data URI mapping that can be imported into another profile.

References: http://blog.vlad1.com/archives/2007/08/10/151/

Tbeachball - Quantifying Mozilla's Responsiveness

This project will add instrumentation to Mozilla in order to determine how much time is spent away from the main event loop, and therefore from the user's mouse/keyboard input. When the user does something, how long does it take for Mozilla to act on that stimulus? One thing that sometimes interferes with interface responsiveness is spending too long away from the main event loop, perhaps in layout or some other intensive computation. This leads to new events from the user, such as mouse clicks or keypresses, not being noticed and acted upon. This could also be expanded in order to include the triggering of log points and specific dialog openings, etc. with a view to understanding what the long-running events actually are.

References: http://shaver.off.net/diary/2007/08/25/tbeachball/ and http://shaver.off.net/misc/latency-tracing-patch.txt

Mozilla Developer Virtual Appliance

Create a Linux-based virtual machine with all the necessary build, development, and testing tools necessary for working on Mozilla. This virtual appliance would be targeted at developers not familiar with the Mozilla toolchain and wanted to get started.

Add support for more compilers to distcc

An earlier project added multi-compiler support to distcc, and then MSVC support. The framework is now in place to support even more compilers. Add support for another compiler, perhaps Java, C#, or something else.

Necessary Skills: C, Python, Makefile

Port distcc to MSYS

The recent work to add MSVC support to distcc works using the cygwin build environment. However, since this project was completed, Mozilla has switched to MSYS (see the [http://developer.mozilla.org/en/docs/Windows_Build_Prerequisites Mozilla Win32 build docs). Port distcc so that it works with the new MSYS environment.

References: Cesar and Tom

Add Multi-File transfer to distcc

The recent work to add MSVC support to distcc only allows for a single file to be sent back from build slaves. This means that extra debugging info (e.g., .pdb files) cannot be sent with binaries. Modify distcc so that it can support debug Mozilla builds.

References: Cesar and Tom

Add password managers for various platforms

Previous work was done on adding OS X Keychain integration to Firefox. This work was focused on tightly binding C++ and Mac API calls. Since then more work has been done to modularize the password manager in Firefox. Various platform "keychains" need to be integrated with the new system. See bugs 309807, 106400, and 371000. See also this e-mail message.

Convert password storage to a SQLite database

Currently, Firefox stores logins in a text file in the user's profile (signons2.txt). The format is simple, but inflexible. Storage of other browser data, such as cookies and form history, has steadily been moving towards using SQLite-backed databases so moving password storage to the same kind of storage would be good. Some work was started in bug 288040.

Support for OpenID

OpenID is an open, decentralized, free framework for user-centric digital identity. It's built on existing web technologies so that it can work without special support in the browser. However, tighter browser integration could allow for increased security and easier management. See bug 356853. Needs definition as to exactly what should be done and how. Could be initially implemented as an extension.

Refactor master password out of NSS

Users can enable a "master password" in their browser, which securely encrypts all their stored passwords. The current implementation can be awkward to use and has some limitations, see bug 322617. Fixing this would involve having password manager being more involved with the cryptography operations -- deriving a key from a passphrase with PKCS#5, and using NSS and PKCS#11 to encrypt/decrypt entries.

Add Offline Support to an open source web app

Firefox 3 will support offline abilities, such that web developers can write their apps so they work even when no network is present. Good headway has been made already porting Zimbra. Pick another web app and add offline support, for example: Moodle.

Thunderbird SMTP Auto-Sensing

Modify Thunderbird so that it uses the correct SMTP server for your current network and IP. In other words, if you are at home, use your home ISP's SMTP, but if you use a school wireless network, switch to the school's SMTP. These various SMTP configurations should be manually controlled by the user, that is, you don't have to try and figure out which SMTP to use. Rather, it should be configurable in an options dialog.

Firefox Session Saver Extension

Write a simple (i.e., nothing more than a simple "save session") extension to leverage the existing Session Store API in order to allow the user to save his/her currently open tabs and restore them. This is similar to what Firefox does when it unexpectedly crashes and gives the user the option to restore a previous session. NOTE: there are extensions that do this already, but many of them are more feature rich than need be. See also this blog post.

Thunderbird Picture Preview

Write an extension for Thunderbird that allows the user to start a slideshow preview of all images in the current message. The feature would work like this: when a mail arrives with pictures attached, a button or some other UI is added to the message that allows the slideshow to begin. When clicked, the images are shown one after the other, properly scaled to fit in the window. Consider using Canvas to do the custom graphics, and take a look at this existing Thunderbird extension for ideas.

Cross-Platform Mozilla Build Farm

Using the accumulated knowledge of the buildbot and distcc projects, and using the physical resources of Hera, create a cross-platform distributed build system suitable for testing patches to Mozilla.

Bugzilla component-watching

It's common for a developer to want to "watch" all the bugs that are filed in a certain Bugzilla component. Mozilla uses a convoluted and painful system of synthetic accounts to simulate this capability, and we'd all really prefer to be able to do it more directly and with less opportunity for error.

Reference: Bug 76794

Automated Testing

Programmatically driving the browser to perform human-like tasks repeatedly. The programmer will have to create a framework for driving browser components in JavaScript or other scripting language to repeatedly test various pieces of the application. Ideally, the successful completion of this task will see a test or suite of tests that can be invoked by command line to start the browser, run the test suites and then exit and record the results for display or analysis.

Reference: Work with Mozilla's Rob Campbell and Jay Patel from the QA Team. See also http://labs.mozilla.com/2007/09/coscripter/

Litmus Extension

We wanted to have an extension that would be able to serve users testcases to try and allow them to submit results to http://litmus.mozilla.org. We could expand the feature set of the extension to do a lot of other cool things with Bugzilla and Website Reporter integration. So we need some ideas there. If you're interested, you could consult the work of David Hamp-Gonsalves, who created the Buggy Bar extension.

"Avoid loading the same page twice" Extension

Create an extension to Firefox so that when a bookmark is clicked, and that site is already open in any tab in any window, that tab/window is brought to the front rather than loading the page again.

Related tech and skills: XUL, JavaScript


Firebug "linting" for portability problems

Lots of web developers use Firebug and Firefox for building their applications, but we want those apps to work well in other browsers as well. If Firebug knew about JS or CSS patterns that could cause problems in other browsers, it would make it much easier to have those applications work in all browsers.

Add-on update helper tools for developers

As Firefox 3 gets ramped up, there are nearly 3000 add-ons that need to get updated to some degree. Write tools (web or XULRunner) to look in add-ons and find things that developers will need to update. You'll want to work with the Mozilla documentation and evangelism teams to figure out what you're looking for and what to recommend to users.

Make FUEL work with Firefox 2

Mozilla's FUEL (Firefox User Extension Library) is a JavaScript Library designed to help developers build extensions using terminology and interfaces that are more familiar to them. Currently it works with Firefox 3, but it would be helpful to have it work with Firefox 2 as well. Work toward having the existing FUEL Tests pass for Firefox 2.

Benchmark SVG

Benchmark SVG performance and look for areas of the code that could be optimized.

Resources: #mozpad

XULRunner Application Packaging

Help to develop an automated packaging system for XULRunner applications.

Resources: plasticmillion, mfinkle, #mozpad

Expose DBUS to XPCOM

The desktop bus (DBUS) is used to communicate between desktop components on a Linux (or Unix) system. If this was exposed to XPCOM objects for bidirectional communication then tighter integration between the desktop and the browser could be achieved. There has been some previous work in this area. Resources: RH online desktop team.

Embeddable HTTP Engine

Create an HTTP facility (transport, not rendering) for desktop apps that shares cache and cookies with the browser -- basically export the firefox http engine. A dbus-exported proof-of-concept exists which which shows some basic structure. A better solution would support more http features, avoid the dbus daemon for large files, and use Firefox instead of libcurl. One challenge is making this work when Firefox is not running. Resources: RH online desktop team.

Universal Firefox on a USB Key

Create a portable USB installation of Firefox that will work on all of Windows, Linux, and Mac. Currently there are binaries for doing this with Windows/Linux. You'll need to get binaries for the app on each platform, figure out how to share a common profile, and get it all on a single USB key. Some existing work in this area has been done, and could be built upon.


'Compact databases' feature for Mozilla calendaring applications (Lightning, Sunbird)

Add the possibility to compact the database for local storage calendars based on SQLite using the VACUUM command. At the moment the local database for events and tasks will not get smaller even if you delete your tasks and events. This has serious performance implications for people, who work a lot with their calendar.

See also Bug 352976

Resources: ctalbert, daniel in #calendar on irc.mozilla.org