404 Error Page Status

Feature Description

We want Firefox to override server-supplied 404 error pages in certain cases.  The Firefox 404 page should provide the user with tools we think are more useful to the user for resolving the situation, than the server supplied page.  This includes alternate ("did you mean") URLs derived from the Places database and some pre-loaded search links/controls.  This is being tracked in bug 482874.


Final work was delayed while I worked on about:support/Troubleshooting Information.  Since then I've rewritten the Places part of the patch in JavaScript as requested in the most recent code review.  I still have some L10N issues to address and some general touch up.

Next Steps

    Pinky: "Gee Brain, what do you want to do next week?"
    The Brain: "The same thing we do every week, Pinky—try to land this patch!"

  • The feature is controlled by a pref so it can be turned off.
  • Webmasters who don't want their 404 error pages to be overridden may have to add padding to their 404 error pages.  However, since IE and Google Chrome are already overriding 404 error pages using a similar size test, webmasters already need to do this.
  • We want to provide a way for the user to see the original 404 error page, but that's not in this patch.

404 Error Page Status

Feature Description

We want Firefox to override server-supplied 404 error pages in certain cases.  The Firefox 404 page should provide the user with tools we think are more useful to the user for resolving the situation, than the server supplied page.  This includes alternate ("did you mean") URLs derived from the Places database and some pre-loaded search links/controls.  This is being tracked in bug 482874.


Found and worked around an SQLite bug (Bug 514291).  Did some performance analysisis, suggestions query tweaking, etc.

Latest patch is attached to the bug and review is requested.

A try server build is available
Next Steps
  • If the reviews come back positive, get it landed.
  • If the reviews raise issues, address them ASAP.
  • The feature is controlled by a pref so it can be turned off.
  • Webmasters who don't want their 404 error pages to be overridden may have to add padding to their 404 error pages.  However, since IE and Google Chrome are already overriding 404 error pages using a similar size test, webmasters already need to do this.
  • We want to provide a way for the user to see the original 404 error page, but that's not in this patch.

404 Error Page Status

Feature Description

We want Firefox to override server-supplied 404 error pages in certain cases.  The Firefox 404 page should provide the user with tools we think are more useful to the user for resolving the situation, than the server supplied page.  This includes alternate ("did you mean") URLs derived from the Places database and some pre-loaded search links/controls.  This feature is being tracked in bug 482874.


Almost ready to go.  Notable progress this week:
  • The docshell changes have r+ from bz with just a few changes, most of which I've already made.
  • Axel has provided some feedback on localization.
  • Johnath has provided feedback on the XHTML.
  • Marco has provided some feedback about the Places changes.  Notably he points out that query performance could be an issue on large databases.
  • David Dahl has provided me with a "max_places" test database.  The bad news is the URL suggestions query can take ~30 seconds to run against this database.  The good news is that a very simple length-check guard in the query can make a big difference.
Next Steps
  • Address  the localization changes Axel raised, and then get formal L10N review.
  • Address the performance issues in the URL suggestions query.  (Probably just a LIMIT-based approach for the initial patch)
  • Get a formal review of Places changes (Dietrich did an informal review a few weeks ago).
  • Get a final UI review.
  • Make sure all the parts of the patch have r+ from the various parties.
  • Land the thing.

404 Error Page Status

Feature Description

We want Firefox to override server-supplied 404 error pages in certain cases.  The Firefox 404 page should provide the user with tools we think are more useful to the user for resolving the situation, than the server supplied page.  This includes alternate ("did you mean") URLs derived from the Places database and some pre-loaded search links/controls.


The newest patch is almost ready for review.

This patch is mostly focused on making the code right, but there are a couple of functional changes:
  • The Firefox 404 error page is now pref-controlled.  There's a pref to turn it on and off, and a pref to limit which server-supplied 404 pages we'll override.  Currently we will only override server 404 pages of 512 bytes or less.
  • We're dropping the background image for now.

Next Steps

  • Complete the patch and get it in for review.
  • Blog about it.


  • This feature is being tracked by Bug 482874 - Provide a friendlier/more useful alternative when the user encounters a 404 error page.

Edit: Added Notes section.

Building Firefox under Eclipse/CDT, or I gotta have my identifier completion


I've been using Eclipse as a C++ IDE for Mozilla development, which apparently puts me in a distinct minority among Mozilla developers.  Since I'm neither an Emacs nor a Vim user, I needed an alternative.  Frankly, Eclipse is not all that great as an editor, but having an IDE with halfway decent identifier completion and code browsing goes a long ways towards making up the difference.  It's not all that hard to set up an Eclipse project for mozilla-central (i.e. Firefox), but it's not exactly obvious either.  The purpose of this blog post is to explain one way to do so.

Disclaimer and Caveats

I'm not claiming that this is the optimal way to set up mozilla-central under Eclipse.  In fact, I'm not even claiming that it's right.  All I'm saying is that this approach is relatively simple and seems to actually work.  For expediency, these instructions make some assumptions.  First, they assume that you are developing under OS X.  I believe that they are easily adaptable to Linux and Windows+cygwin (or Windows+mozilla-build), but I don't know for sure.  Second, the instructions assume that you are already able to build mozilla-central from the command-line.  If you don't already know how to do that, then you don't want to start here.

Disclaimer #2

As a final disclaimer, this approach is relatively new to me.  I've mostly been using an older version of the CDT under Eclipse 3.0 and I'm not even going to mention what I had to do to get indexing to work properly.  This new approach is a lot simpler, and I'm reasonably confident that it will work well for sustained use.


I originally wrote these instructions in text-only form.  I've since added some marked-up screenshots for many of the steps.  These screenshots are attached at the end of the post.


And last, but not least, thanks to Mike "firetoad" Collins for trying out a slightly earlier version of these instructions.

Now the instructions.

Install Eclipse/CDT

  • Obviously you need Eclipse.  I will offer some advice, but I'm largely assuming that you can figure this part out for yourself.  These instructions assume that you are using Eclipse 3.5, better known as Galileo.  You need Eclipse with the CDT, the "C/C++ Development Tooling".  The easiest way to get this is to go to http://www.eclipse.org/downloads/ and download the pre-packaged Eclipse IDE for C/C++ Developers, which is what I've done.  I have an existing Eclipse installation, but rather than trying to upgrade, I've just got them side-by-side, which doesn't seem to be a problem as long as you set up separate workspaces for them.
  • By default Eclipse runs with a 256M heap, which is too small to reliably index the mozilla-central tree.  I've been running with 512M without any problem, although I've not really pushed it very hard yet.  There ought to be a way to configure the Eclipse heap size through the UI, but if there is, I haven't found it yet.  In the meantime, I've been launching it from the command-line, like so:
    • /Applications/eclipse-galileo/Eclipse-galileo.app/Contents/MacOS/eclipse -vmargs -Xmx512M &
  • Your path will differ, of course.  Also note that the path above is a little bit unusual because I've renamed the actual Eclipse application so it won't conflict with my other (older) Eclipse installation.
  • By default, Eclipse engages "Scalability Mode" for large files, which means it turns off most of the cool features.  I've increased the threshold to 15,000 lines (nsDocShell.cpp is about 11,000 lines of code, for example).  To change this setting, choose Eclipse >> Preferences, then type "scalability" into the search box.

Create a new project

  • Start Eclipse (but see the note above about making sure you're running it with enough memory).
  • Ctrl-click in the Project Explorer, choose New >> C++ Project
    • this will open the C++ Project dialog
  • In the C++ Project dialog,
    • In the Project name text field, enter the project name
      • I'm using firefox-eclipse
    • Uncheck Use default location
    • In the Location text field, enter the full path for the project source code.
      • I'm using /Users/cbartley/Dev/firefox-eclipse/src
    • Under Project type, choose Makefile project >> Empty Project
      • Note that other project options can cause Eclipse to create a default makefile, which you don't want.
    • Click Finish
  • You should now see the project firefox-eclipse in the Project Explorer.

Configure the project

  • Ctrl-click on firefox-eclipse in the Project Explorer, and choose Properties.
    • This will bring up the Properties for firefox-eclipse dialog.
  • In the Properties for firefox-eclipse dialog,
    • Select the Resource panel
      • Note the path after Location in the Resource panel.
        • This is the same path you entered a few steps above, but if you're like me, you've probably forgotten it already.
    • Select the C/C++ Build panel
      • Select Builder Settings inside the C/C++ Build panel
        • Uncheck Use default build command
        • In the Build command text field, enter:
          • bash -l -c 'make -f client.mk $0 $1' -b
            • this hairy looking command invokes a bash shell which in turn invokes the actual make command.  The dash-small-L option tells it to behave like a login shell.  This sidesteps some environment problems that I've run into when running make directly from Eclipse.  I won't elaborate on the $0, $1, or -b, except to tell you to be careful that you put the quotes in the right place!
        • In the Build directory text field, enter the path to the project source
          • I'm using /Users/cbartley/Dev/firefox-eclipse/src
            • Note that this the same path found in the Location text field on the Resource panel. I usually just Cmd-C copy it from there any more, to reduce the possibility of a typo
      • Select the Behavior tab inside the C/C++ Build panel
        • In the Build (Incremental build) text field, delete all and just leave the field empty.
    • Expand the C/C++ Build tab to show the tabs for its subpanels.
      • Select the Settings sub-panel
        • Select the Mach-O parser under Binary Parsers
    • Expand the C/C++ General tab to show the tabs for its sub-panels
      • Select the Index sub-panel
        • Check Enable project specific settings
        • Under Select indexer
          • Choose Full C/C++ Indexer (complete parse)
          • Check Index unused headers as C++ files
    • Click Apply
  • The project is now set up, but there's no source code yet.

Get the Source Code

  • Eclipse has created the directory for the project, but it's empty.  I've created my project at /Users/cbartley/Dev/firefox-eclipse/src.  Following standard convention, I want to check the mozilla-central project out to a directory named src, the same one as above, in fact.  The problem is that Mercurial complains if the directory already exists.  Checking the source out first doesn't help, because then Eclipse complains that the directory already exists.  I'm going to check mozilla-central out to a different location and then manually merge them.  Note that Eclipse has created some files inside the src directory that we want to preserve.
  • cd into the project directory
    • I do cd /Users/cbartley/Dev/firefox-eclipse/src
  • cd out one level
    • cd ..
      • I'm now in /Users/cbartley/Dev/firefox-eclipse
  • Now type the following commands
    • hg clone http://hg.mozilla.org/mozilla-central src-aside  # get the source code
    • mv src-aside/.h* src                                      # move mercurial files into src
    • mv src-aside/* src                                        # move regular files into src
    • rmdir src-aside                                           # remove the now superfluous src-aside directory

Set up the .mozconfig file

  • Create a basic .mozconfig file for the project; I'm assuming you've already got a usable one some place.
    • For example, I do cp ../mozilla/src/.mozconfig src
  • Make sure the "-s" option is NOT on for make
    • Normally make displays the complete command line for building each file.  When invoked with the "-s" option, however, it only displays the name of the file being built.  This is a problem for building under Eclipse, since Eclipse parses the build output to figure out where the header files are.
    • In my existing .mozconfig, I have the following line:
      • mk_add_options MOZ_MAKE_FLAGS="-s -j4"        # before
    •     I delete the "-s", leaving
      • mk_add_options MOZ_MAKE_FLAGS="-j4"           # after

    Build the project under Eclipse

    • Make sure that firefox-eclipse is selected in the Project Explorer.
    • Select the Console tab in the bottom pane so you can see the build output
    • Select Project >> Build Project from the menu
    • Wait

    Build the project under Eclipse, again

    • We want to do Clean rebuild under Eclipse; This extra step seems to be necessary for Eclipse to find the IDL-generated header files.  Don't ask me why.
    • Make sure that firefox-eclipse is selected in the Project Explorer
    • Select the Console tab in the bottom pane
    • Select Project >> Clean... from the menu
      • Make sure that Start a build immediately is checked
      • Click OK
    • Wait

    Make sure indexing is started

    • After the build completes Eclipse should start Indexing automatically.  A progress indicator should appear in the status bar on the lower right.
    • If indexing hasn't started, you can invoke it manually by:
      • Ctrl-click on firefox-eclipse in the Project Explorer, and choose Index >> Rebuild
    • Indexing takes something like four hours on my machine.  Eclipse remains usable in the meantime, you just won't have access to the more advanced code-completion and browsing features.

    After Indexing is completed

    • Cmd-Shift-R is kind of like the Awesome Bar for source files.
    • Cmd-Shift-T is kind of like the Awesome Bar for types, functions, etc.
    • Ctrl-Space invokes identifier completion.
    • Just typing identifier. or identifier-> will show a list of member variables and functions if you pause for a second.
    • F3 will take you to the declaration of an identifier
    • F3 over a #include will open the header file.
    • Hovering over a macro will show you its expansion.
    • Actually, many of the features will work in full (e.g. Cmd-Shift-R) or in part (e.g. identifier completion) before indexing is completed.

    How do you surf the firehose?

    How do you surf the firehose?

    I'm not sure that "drinking from the firehose" is the right metaphor for Mozilla.  I still claim it's more like having people constantly shooting you with water pistols from every different direction.

    There's not much method to my madness, but here's how I try to keep up with the firehose:

    I'm subscribed to four newsgroups in Thunderbird:
    • mozilla.dev.planning
    • mozilla.dev.apps.firefox
    • mozilla.dev.platform
    • mozilla.dev.tree-management
    I don't actually read them any more, with the exception of mozilla.dev.planning which I now have set up to deliver straight to my main Mozilla email account.  (Because Mike Beltzner told me to.  "There's hardly any traffic on that list anyway" I recall him saying.)  Since conventional email is hardly used at Mozilla, this seems to work OK.

    I also have bugzilla notifications going to my main email account.  I don't know what I was thinking.

    I'm subscribed by email (my personal email) to the Firebug Group on Google Groups.  In hindsight this was a mistake, but I haven't fixed it yet.  I'm mostly skimming the thread subjects at this point, but I at least have some inkling of what's going on there.  I check this a couple of times a day when I check my personal email.

    I read planet.mozilla.com every couple of days.  Anymore I mostly skim it, reading just the occasional post.  I really should try to check it every day, since I find it pretty useful to the extent that I do read it.

    I peruse reddit and news.ycombinator.com (Hacker News) several times a day, and slashdot every now and then.  If a blog post doesn't make it to reddit or Hacker News, I probably won't see it.  I don't feel like I'm missing a lot of stuff by doing this, but on the other hand I do feel like I have to wade through a lot of crap to find the stuff I do want to read.  (Ha, Sturgeons Law: 90% of everything is crap.  Curtis's corollary to Sturgeons Law: On the Internet, 90% of the rest is crap too.)

    I don't do twitter, largely because it seems to scratch an itch I don't have.

    I have three hours or so of phone meetings every week, not counting the weekly Mozilla.org meeting.  I take these meetings fairly seriously since I'm working remotely.  I don't take notes, but I do attempt to actually pay attention.  I'm prone to zoning out even in meetings I'm physically in, and it's worse when I'm dialed in remotely.  My secret weapon?  I usually surf pictures on Flickr while I'm on the phone.  This works surprisingly well since it doesn't require any major effort from the verbal part of my brain.  In the future I should probably take notes at least some of the time.  Like any time Mike Beltzner is saying something important.

    I don't use a feed reader.  In the past I've used Bloglines and Google Reader.  I abandoned Google Reader because I always hated its infinite scrolling model.  (I don't remember why I hated it; I like infinite scrolling in other cases) .  I do have a master list of blogs I read.  Despite the dozens of blogs on this list, I only read a few regularly these days, and none of them are technical blogs anymore.

    And of course Mozilla lives and breathes IRC.  I'm using Colloquy as my IRC client.  It's not wonderful, but it works OK.

    Tag, you're it.

    Robcee's original post (complete with standard questions which I totally ignored) was designed to be viral.  I can't beg off this part, since I put him up to it.  Seriously, though.  I'm only asking because I think I might learn something.

    What should Mozilla look for in an automated review system?

    In last week's All Hands when talking about possible process improvements, Mike Shaver asked if there was any interest in some sort of automated review system.  Google's Mondrian is probably the most well known review system.  Review Board seems to be the best known open source review system.  I've never used Review Board, but as an ex-Googler, I have used Mondrian.

    I think Mozilla would really benefit from a good review system.  My review system experience is limited to Mondrian, which is, of course, proprietary to Google.  I don't know how Review Board and the other open source systems measure up, and, as I'm about to get to, I think Mondrian itself has some notable deficiencies.  My overall point is that if we decide to adopt a review system, we should choose carefully.  I'll try to offer some guidance about how to do that.

    Quick Overview of Mondrian

    Mondrian has a dashboard view for outstanding changelists but to me the essence of Mondrian is its UI for reviewing a single changelist, and that's what I'm going to talk about here.  At the highest level you have a list of files modified in the changelist.  You can click on each file and get a diff of that file against the baseline in a side-by-side view.  You can easily see which lines were added, modified, or deleted.  OK, everybody is familiar with side-by-side diffs these days.  Here's the cool part though: A reviewer can click on any line and leave an inline note or comment.  Afterwards, the reviewee can go through the diff and respond to the comments.  The way this often works in practice is the reviewer leaves an instruction in the comment and the reviewee responds by making the requested change and replying to the comment with "Done."  In fact, there's a "Done" button as well as a "Reply" button on the comment just to make this last part easy.

    Once the reviewee has updated the changelist to the reviewer's satisfaction, the reviewer will approve the change and the reviewee can submit it to the source control system.

    What's not to like?

    Mondrian's review UI is great for simple line-by-line reviews, things like:
    • "You don't need to check for null here."
    • "You should check for an error here."
    • "Fix indentation" 
    Sometimes this level of line-by-line scrutiny is really useful.  For example, exception-free C++ code often requires the programmer to check virtually every single function call for errors.  This is hard for even the best programmers to get 100% right.  But let's be clear here.  What Mondrian is really good for, all the time, is line-by-line nitpicking of code.  And frankly, line-by-line nitpicking of code is already pretty easy without some fancy tool like Mondrian to make it extra easy.

    Mondrian's review comment system really seemed to encourage a style where there was a one-way flow of instructions from the reviewer to the reviewee: "Do this.  Do this.  Do this." and the reviewee replies with "Done.  Done.  Done."  Sometimes this is appropriate, but oftentimes it isn't.

    Of course the reviewee always has the option of clicking "Reply" instead of "Done".  You could have a whole thread of comments and replies if you wanted to.  But given the limitations of the UI, that would be kind of like communicating with short messages written on post-it notes.  And not regular sized post-it notes either, but rather those extra tiny ones.

    So Mondrian not only encouraged a review focus on individiual lines, it also tended to encourage a one-way flow of information from reviewer to reviewee which could easily degrade into a one-way flow of commands.

    What would I want in a "good" review system?

    It may seem like I'm arguing that a review system should actively discourage line-by-line review, and that's not actually the case.  I think that review style is often useful, and pretty much any review system, good or bad, will support it.

    There are really two fundamental things that I do want out of a review system.
    • The system should go out of its way to support a bi-directional flow of information between the reviewer and the reviewee.  In the extreme, it should provide a means of carrying on an actual conversation.  This could be supported within the review system itself, but even a simple means of escalating a conversation to email (or even IRC) might be a big help.
    • The system should support higher level discussions about the code under review.  Actually I think it should go so far as to encourage this kind of information flow.  You can sort of do this with Mondrian, but you are usually better off just going to email.

    Some general guidance

    I've put a fair amount of thought into how I'd like an ideal review system to work.  Mostly I've been thinking in terms of concrete features, but that's of limited utility if what you want to do is choose between existing review systems rather than writing a review system yourself.

    So I'm trying to figure out some general guidance.  I think what I'm trying to say in this post more than anything is that affordances matter.  Mondrian, for example, seems to afford the line-by-line scrutinizing, nitpicking approach to code reviews.  It also seems to afford a model where the reviewer simply gives instructions to the reviewee and the reviewee just automatically carries them out.

    Mondrian offers some minimal affordance for discussing (as opposed to simply commenting on) a particular line of code, but it could do a lot more.  Notably it does not seem to offer any real affordance for discussion at the design level, which has always seemed to me like a serious omission.

    A simple concrete example

    Here's one simple way that Mondrian could be improved which I hope will illustrate my point about affordances.  As described above, Mondrian comments have two buttons, "Reply" and "Done".  It could offer other choices as well, so you might have:

    "Reply", "Done", "Done w/comment", "Defend", etc.

    These latter two functions could easily be done with "Reply", but if you give them their own buttons, you explicitly tell the reviewee that there are other options that can be considered here.  In particular, in this case, they give the reviewee permission to say "I had a reason for doing this the way I did, would you please consider my reasoning?".

    My larger point here is that a review system's UI strongly shapes the code review process, and it can shape it in good ways or bad ways.  As a result, we want to think not just about what we want out of a code review system, but also what we want for the code review process itself.

    Super-short Summary

    1. A good review system needs to support two-way information flow, and it should probably go so far as to actively encourage it.
    2. A good review system should support review at a higher level than simply line-by-line, and it should probably go so far as to actively encourage it.
    3. Affordances matter.


    • This post describes Mondrian as of about nine months ago when I last used it.  It may have received a major upgrade in that time, I don't know.  Mondrian may also have had useful features that I didn't know about -- not all of its features had good discoverability.
    • Probably the most common code review case at Google is a reviewer and reviewee who are peers, have known each other for months if not longer, and who sit near each other, often in the same room.  Mondrian's limitations are a lot less important in this scenario, since a lot of out-of-band communication can happen the old fashioned way, by talking face-to-face.
    • In some circumstances you can have reviewers and reviewees who work in different locations and who have never met in person.  This is not uncommon at Google, and it is in fact very common at Mozilla.  In this scenario, the way the review system works becomes much more important.  
    • The obvious way to get around Mondrian limitations is to fall back to email.  I ultimately started emailing concerned parties requesting informal reviews before requesting formal reviews through Mondrian.  Mondrian can still be used in this case -- it can see pretty much any outstanding changelist.  Nevertheless, by making an informal request in email, you could get a high-level review of a changelist without getting mired in low-level details.  And since this was through email rather than Mondrian comments, you could hold an actual conversation about it.  It turned out my team's superstar intern had independently started doing the same thing.  I jokingly referred to this as "doing a review before the review".
    • This post might lead you to believe that I'm a Mondrian-hater.  Actually, I think Mondrian is a very good tool, I just feel like they quit working on it before they were done.

    The Firebug tabs-on-top UI refactoring has landed

    Note: Since I was getting lots of comment spam I've closed comments here.  If you have any comments or suggestions, feel free to file them on the Firebug Issues page or drop us a line in the Firebug discussion group in the Tabs on Top thread.

    The Firebug tabs-on-top UI refactoring has finally landed in Firebug 1.4 alpha.  You can download it as firebug-1.4.0a17.xpi (or later) from http://getfirebug.com/releases/firebug/1.4.  This change is pretty close to the one I described and prototyped in February (see Improving the Firebug UI) and which in turn follows my original proposal from late last year (see Firebug UI: Tabs on the Top).  These blog posts describe the reasoning behind the change so I won't rehash it in full.

    The old layout had a toolbar at the top of the Firebug UI, and a tabbed "panel bar" below it.  Sometimes a tabbed "side panel bar" would show up to the right of the main panel bar.  This change essentially takes the toolbar and the side panel bar (when it appears) and puts them inside the active panel.  This puts the panel tabs at the top of the Firebug UI (hence the name "tabs-on-top") and panel specific controls underneath.  Several controls that are not panel-specific (the main Firebug menu, the search box, and the detach window and close window buttons) have been moved to the tab strip, so they are effectively in the same location as before the change.  The idea is that UI elements that are specific to a panel look like they are inside that panel.  This description probably makes the change sound more complicated that it really is.  A screenshot will better communicate how it looks.

    Firebug UI with the tabs-on-top layout.  The Net panel is selected.

    Even the screenshot is a poor substitute for actually downloading the extension and trying it out, so let me encourage you to do that.

    There are still some outstanding issues, before this change can really be called complete.
    • The location of the main panel Options menu is less than ideal.  I think the right thing to do here is to merge the panel options with the menus on the panel tabs, but I haven't had time to prototype it yet.
    • The Inspect button really belongs on the top toolbar.  I've held off on relocating it because I think it will look too much like the label of a panel tab.  There are a couple of options here.  We could change the styling to make it look more like a button.  We could replace the label with an icon, or we could change the tab styling so even unselected tabs show a tab outline.  Probably any of these solutions would work, we just need to figure out what the best one is.
    • The position of the debugger buttons can jump drastically under some conditions.  Normally the debugger buttons only appear when the Script panel is active.  However, if JavaScript execution is paused, the debugger buttons always appear, regardless of the tab.  With the new layout this means their position can jump if you switch to another tab which doesn't have a side panel bar.  I think it might make sense to move them to the left side of the toolbar, which will eliminate this problem.  This seems like it might be a fairly controversial change, so we really need to explore it separately.

    What are "futures", and why should you care?


    Last week we were discussing, among many other things, ways to speed up Firefox during startup.  One obvious option was to move more of our I/O off of the main thread.  This in turn involves making more code asynchronous, and asynchronous code is simply harder to manage.  Mike Shaver mentioned something about "futures" as a possible way to handle the extra complexity, and then the discussion moved on to something else.

    I'm not exactly an expert, but I've not just used futures, I've written my own implementations in JavaScript and even Object Pascal (in hindsight I'm not sure the latter was a good idea, but it was certainly an interesting exercise).  Futures seem esoteric, but they really shouldn't be -- the idea is really quite simple.  In this post I'll try to explain what futures are and how they can be used to make asynchronous programming easier.

    What exactly is a future anyway?

    In the simplest form, a future works like an IOU.  I can't give you the money you've asked for right now, but I can give you this IOU.  At some point in the future, you can give me the IOU and I'll give you the money -- if I have it.  If I don't have it yet, then you can wait until I do.  I get paid on Friday.

    Alternatively there's the dry cleaner metaphor.  You drop your clothes off on Monday and the clerk gives you a ticket that you can use later to reclaim your clothes after they've been cleaned.  The clothes will be ready on Tuesday morning, but if you show up too early, you'll have to wait.  On the other hand, if there's no hurry, you can just do other stuff on Tuesday and show up on Wednesday with a reasonable expectation that they'll be ready when you arrive.  You'll just hand your ticket over, collect your clothes, and be on your way.

    A future is similar to the IOU (or the dry cleaning ticket).  It gives you a way to represent the result of a computation that has not yet completed, and it allows you to access that result once it becomes available.  So you can call a function which starts some asynchronous process but doesn't wait for it to finish.  Nevertheless the function can return you a useful result: a future which can be used to claim the real result later.

    Of course if you ask for the result too soon, you'll have to wait.  On the other hand, if the result becomes available before you want it, then it will wait for you.

    A simple example

    Here's an example of what this might look like in pseudo-JavaScript:

    function doStuff() {
      var cleanClothesFuture = dryCleaner.dropOff(dirtyClothes);
      var cleanClothes = cleanClothesFuture.get();  // block if the result is not ready yet

    Compare this to the traditional way we'd handled this in JavaScript, using a callback:

    var cleanClothes = null;

    function doStuff() {
      dryCleaner.dropOff(dirtyClothes, function (clothes) { cleanClothes = clothes; });

    These examples are not one hundred percent semantically identical, but they should be close enough to illustrate the point.  I contend that the first function is easier to write, easier to read, and easier to reason about.  I also contend that the difference isn't enough to get excited about.  It's when things get more complicated that futures become really useful.

    A more complicated example

    Imagine that I have a web page that sends an AJAX request to a server and then displays the results in an IFRAME -- and furthermore does it automatically on page load.  I have to wait for both the AJAX request to return data and for the IFRAME to finish loading -- only then can I display the results.  This can be done fairly simply using callbacks:

    function showData(dataUrl, iframeUrl) {
      var data = null;
      var iframeBody = null;
    tryToShowData { if (data && iframeBody) { showDataInIframe(data, iframeBody); } }
      requestDataFromServer(dataUrl, function (response) {data = response.data;  tryToShowData() });
      insertIframeBody(iframeUrl, function (iframeDoc) {iframeBody = iframeDoc.body; tryToShowData() });

    Now, imagine the same thing done with futures:

    function showData(dataUrl, iframeUrl) {
      var dataFuture = requestDataFromServer(dataUrl);
      var iframeBodyFuture = insertIframeBody(iframeUrl);
      showDataInIframe(dataFuture.get(), iframeBodyFuture.get());

    Again, these two examples are not semantically equivalent -- notably there's no blocking in the first example.  Now let's imagine that we had a way to turn an ordinary function into a new function which takes futures as arguments and which returns a future in turn.  As soon as all the future arguments became available, the base function would be called automatically -- and once the base function completed, its result would be accessible through the future returned earlier.  I'll call this capability "callAsap": call a function as soon as possible after all of its future arguments become available.  Using callAsap(), the previous example might be rewritten as:

    function showData(dataUrl, iframeUrl) {
      var dataFuture = requestDataFromServer(dataUrl);
      var iframeBodyFuture = insertIframeBody(iframeUrl);
      showDataInIframe.callAsap(dataFuture, iframeBodyFuture);

    In this case we don't care about the return value for showDataInFrame.  This example is much closer in behavior to the earlier callback-based example.  In fact, the callAsap() method would be implemented with callbacks underneath, but they would all be nicely abstracted away under the hood.

    One of the nice things about callAsap() is that it can nicely handle cases where you are waiting on more than one future.  Imagine that you've asynchronously requested data from two different servers:

    function showData(dataUrl1, dataUrl2, iframeUrl) {
        var dataFuture1 = requestDataFromServer(dataUrl1);
        var dataFuture2 = requestDataFromServer(dataUrl2);
        showDataInIframe.callAsap(dataFuture1, dataFuture2, iframeBodyFuture);

    This segues nicely into the next topic: Arrays of futures.

    Arrays of futures

    Imagine if you have not one, or two, or three futures, but rather an arbitrary number of futures.  What we'd really like to have is a way to take an array of futures and produce from it a single future for an array of concrete values.  Something like:

    function showData(dataUrlArray, iframeUrls) {

      // The "dataFutureArray" is a concrete array of futures.
      var dataFutureArray = requestDataFromServers(dataUrlArray);

      // The "dataArrayFuture" is a future for a concrete array of concrete values.
      var dataArrayFuture = Future.createArrayFuture(dataFutureArray);

      showDataInIframe.callAsap(dataArrayFuture, iframeBodyFuture);

    What this example might look like rewritten in callback style is left as an exercise to the reader.

    An advanced example

    OK, now for a more elaborate example.  Imagine a function which retrieves the first page of Google search results for a particular query, and then goes through and re-orders the results based on its own ranking system.  Furthermore, imagine that this ranking is computed based on the contents of each web page.  We'll need to to make requests to many different servers for many different web pages.  This will be fastest if we issue all the requests at once.

    function search(query) {

      // Take a concrete search result and return a future to a
      // [searchResult, ranking] pair.

      function requestWebPageAndRanking(searchResult) {
        var webPageFuture = requestWebPage(searchResult.url);
        var rankingFuture = computeRankingFromContent.callAsap(webPageFuture);
        return Future.createArrayFuture([webPageFuture, relevanceFuture]);


      // Take a concrete array of search results and return a future to
      // an array of [searchResult, ranking] pairs, sorted by ranking.

      function requestSearchResultsSortedByRanking(searchResultArray) {

    rankingArrayFuture = Future.createArrayFuture(
          [requestWebPageAndRanking(searchResult) for (searchResult in searchResultArray)]
        return sortArraysByKeyIndex.callAsap(
    rankingArrayFuture, 1);

      // Request search results, re-rank them, and then display them.
      var searchResultArrayFuture = requestGoogleResults(query);
      var sortedRankingArrayFuture =

    In all fairness, this is not as simple as a synchronous blocking implementation.  Keeping your arrays of futures and futures of arrays straight is a little bit taxing.  Imagine what a callback model might look like, however, with callbacks inside callbacks.  One advantage of using futures is that you can often write traditional blocking code and then in straightforward fashion translate that code into asynchronous code using futures.


    Notes, in no particular order

    • The examples may look like JavaScript, but they are, in fact, pseudo-code.  The implementation of some of the helper methods, psuedo-code or otherwise, are left to the imagination.
    • I have completely glossed over error handling, including such interesting topics as exception tunneling, fallback values (nulls, empty arrays, NullObjects), not to mention timeouts and retries.  If this sound scary, it's because error handling in any kind of async code is a difficult topic.  Futures don't make the situation any worse, and might make it better.
    • The name "callAsap" is my invention, although I'm certain the underlying idea has been invented independently many times.  Also note that callAsap() and Future.createArrayFuture() are fundamentally quite similar.
    • Java futures (java.util.concurrent.future) use a blocking get() method like the one in the first example.  I don't actually know how you could do a blocking get in conventional single-threaded JavaScript, which is the whole genesis of callAsap().  Practical JavaScript futures need to be "listenable futures" which raise events when they resolve.  The methods callAsap() and Future.createArrayFuture() can then be implemented using this capability.  Client code can then use these methods to avoid writing explicit callbacks.
    • The re-ranking Google search results example is contrived, but it's based on a similar project I did a few years ago.  In that project I used callbacks, and it was quite painful.

    Improving the Firebug UI

    Refactoring the Firebug UI

    My main project for the last few weeks has been doing some Firebug UI work that I’ve been thinking about for a while.  Since I’ve just been moving UI elements around (to make the layout more logical), I’ve been referring to this effort as “UI refactoring”.

    So here’s the deal.  It’s not enough that I think this redesign is an improvement, or even that the Firebug working group has been supportive.  We need to know what you, the Firebug user, think.

    Before I get started, I should let you know that I’ve been working on the Firebug 1.4 branch, which is still undergoing active development.  The changes I’m proposing are fairly simple in the sense that they are really only a reorganization of existing UI elements.  However, if you have not yet seen 1.4, be aware that there are other significant UI changes, notably John J. Barton’s work on the activation UI, kpdecker’s multi-file search capability, and Hans Hillen’s work on accessibility.

    What’s wrong with the current Firebug UI?

    The basic UI layout for Firebug today consists of a toolbar across the top, a main “panel bar” underneath, and an optional “side panel bar” to the right.  Each panel bar consists of a set of tabbed panels, only one of which is visible at a time.


    This is a simple design, and seems pretty logical.  There’s a problem, however.  The contents of the toolbar can change substantially depending on which panel is selected in the main panel bar.  For some panels this is maybe not such a big deal.  At the other extreme, consider the Net panel.  When the Net Panel is selected, it adds seven buttons to the toolbar which control what information is displayed inside the Net panel’s content area.

    For me, using these buttons, with this placement, feels even weirder than the screenshot looks.  They would feel much more natural if they were adjacent to the content area that they affect.

    Furthermore, the side panel bar is only visible when either the HTML tab or the Script tab is selected.  And, of course, it’s a different set of panels that show up in the side panel bar depending on whether it’s the HTML panel, or the Script panel.  Logically, you could think of the HTML panel as having a smaller panel bar embedded inside it which in turn contains several sub-panels specific to the HTML view.  You can think of the Script panel similarly.  The other panels simply don’t have any sub-panels they need to display.

    Basically, the problem is that clicking a tab changes not only the main panel below the tab strip, but also the toolbar above it, and oftentimes the side panel bar to the right as well.  You click on a tab, and the whole UI mutates around it.  Even though the current UI layout makes superficial sense, I contend that the toolbar – or at least a big chunk of it – belongs inside the main panel bar.  I also think you can make an argument that the side panel bar, when it appears, should look like it’s inside the currently selected panel.

    Another way to think of it: The tabs – Console, HTML, CSS, Script, DOM, and Net – should appear at the top of the Firebug UI where the panel-specific toolbar elements appear now.  Since these main panels are the major organizing principle of the Firebug UI, it makes sense that they should be more prominent.

    I’m clearly not the first person to come to this conclusion since Firebug Issue 222 makes a similar argument.

    This is the layout I have working now.  I’ve been referring to this as the “tabs-on-top” layout.

    In this design, not all of the toolbar has been relocated.  The main Firebug menu that hangs off the Firebug button is in the top-left just as it always has, and the search elements and Firebug window control elements are still at the top right.  Although the side panel bar doesn’t really appear to be inside the main panel, it does at least appear subservient. 

    There are some smaller changes as well.  Notably the tabs are now “stand-up” style –

    , rather than “hanging” style – .

    The handling of the main panel Options button and the Inspect button still leave something to be desired.  Nevertheless, I think this is a more logical layout, although you probably have to use it to really appreciate the improvement.

    What now?

    As I mentioned above, you really need to try the out the new layout to see if it’s really does work better.  To that end, I’ve attached an experimental version of the Firebug extension to the post, so you can download it and try it out.  I’ve tested it fairly thoroughly on Windows and to a lesser extent on Mac and Linux.  Also, since it’s built on top of the 1.4 branch, which is still under active development, you probably won’t want to leave it installed for too long.  However, it should be good enough for you to take it for a test drive.