Speeding up jQuery’s each function

In my previous post, “Investigating JavaScript Array Iteration Performance“, I found that among a selection of different array iteration methods, jQuery’s each function was the slowest. It’s worth mentioning again that these investigations are pretty academic – array iteration and looping speed is unlikely to be the source of performance problems compared to actual program logic, DOM manipulation, string manipulation, etc. I just found it interesting to poke into how things work in different browsers. That said, with the recent release of jQuery 1.4 emphasizing performance so much, I wanted to see what if anything could be done to speed up each (which is used inside jQuery all over the place), and whether it would made much of a difference.

Again, the details are after the jump.
Read the rest of this entry »

Investigating JavaScript Array Iteration Performance

The other day I was working on some JavaScript code that needed to iterate over huge arrays. I was using jQuery’s $.each function just because it was simple, but I had heard from a bunch of articles on the web that $.each was much slower than a normal for loop. That certainly made sense, and switching to a normal for loop sped up my code quite a bit in the sections that dealt with large arrays.

I’d also recently seen an article on Ajaxian about a new library, Underscore.js that claimed to include, among other nice Ruby-style functional building blocks, an each function that was powered by the JavaScript 1.5 Array.forEach when it was available (and degrading for IE). I wondered how much faster that was than jQuery’s $.each, and that got me to thinking about all the different ways to iterate over an array in JavaScript, so I decided to test them out and compare them in different browsers.

This gets pretty long so the rest is after the jump.
Read the rest of this entry »

JSONView 0.4 with content negotiation

JSONView 0.4 was just approved on addons.mozilla.org. This one is mostly a bugfix release from 0.3, but I couldn’t let it go out without one new feature, so I finally added support for content negotation. This means that you can go to the new JSONView options dialog (in the Addons menu) and enable sending the “application/json” content type with your HTTP Accept header. Some applications (like some Ruby on Rails websites) are set up to return JSON if you ask for it via the HTTP Accept header, and XML or HTML otherwise. This means thay you can now specify a preference for JSON, which should help when you’re in that situation. I specifically made sure it works with CouchDB, which is an exiting document-based database with a JSON API. Before this option, CouchDB would return plain text to browsers, but now you can make it send JSON that’ll be nicely formatted by JSONView. Note that the option to modify the HTTP Accept header isn’t on by default because many people won’t need it, and it has the potential to screw up some websites that don’t expect browsers to ask for JSON. That said, I’m sure this will be helpful in a bunch of situations.

JSONView Options

On the bugfix side, I fixed the issue I mentioned in my post about JSONView 0.3 where values that were 0 or false wouldn’t show up at all as a result of some overzealous error-checking. I also hunted down an infuriating bug where JSON documents would sometimes report as invalid when Firebug was on. This took a while to track down, but apparently Firebug slows down the loading of the document enough to mess with how I was reading a stream (and I had to be careful not to re-trigger an earlier bug that could cause Firefox on OSX to crash.) Anyway, it’s fixed now, and everybody should see more reliable document rendering, Firebug or not. For those living on the edge, JSONView is also marked as compatible up to Firefox 3.7a1pre, so you should have no trouble installing it on Minefield or the upcoming Firefox 3.6 alpha.

One thing that might stand out is that the new options dialog might not be localized for your language. JSONView’s localizations are handled by the generous volunteers at Babelzilla, which is great for me, but not as great when I want to rush out a release with important bugfixes. In this case I chose to ship 0.4 before all the new localizations had arrived, but I’ll add them all in again for the next release once the translators have caught up.

So grab JSONView 0.4 from addons.mozilla.org, or just update your extensions if you’ve already got it installed. As usual, feel free to send me feedback as comments on this post, and file bugs and feature requests at Google Code.

Swoopo Profits Greasemonkey Script – Entertainment Shopping

In the last few weeks I’ve become increasingly obsessed with the evil genius that is Swoopo.com. Swoopo is a penny-auction site – users buy bids for $0.60, and each bid placed on an item increases the price by $0.12. The cost of bids and the amount they increase the price of the item vary depending on the type of auction and the country you’re in. Swoopo does a lot to make it harder to win, though. For example, if the bid is placed within 10 seconds of the end of an auction, the closing time of the auction is extended by 10 seconds or so, so people can have last-second-sniping wars for as long as they want. They also offer a “BidButler” service that will automatically bid for you, and of course if two users in an auction are using it, the BidButlers just fight until they’ve used up all the money they were given. Swoopo’s operation is like cold fusion for money – they make insane amounts of cash off their users, and they only have to drop-ship the item to one user so there’s theoretically very little operating cost (they already have the money from selling bids, and they don’t need to maintain inventory). They’re shameless enough to even have auctions for cash, gold, and even more bids! Because everyone in the auction is paying to participate, even if the winner gets some savings on the item, Swoopo makes far, far more on the sunk bids – sometimes 10x the price of the item in pure profit.

Jeff Atwood (of codinghorror.com) has written about Swoopo multiple times, and some techies have even tried to game the system, but it hasn’t worked. I was introduced to Swoopo through Jeff’s blog but I hadn’t thought about it forever, and for some reason it came up again recently. After looking at it a bit, I was just floored by how they’ve managed to set up such a perfect money-generating system. The company that runs Swoopo is called “Entertainment Shopping”, which I guess is supposed to be a suggestion that it’s like gambling (where it’s “fun” to lose money) though they really, really don’t want to be regulated as gambling. I don’t personally find gambling (or bidding on Swoopo) to be that fun, but I do find it entertaining to watch the astronomical profits tick up as more and more suckers toss money into an auction. So I built a little Greasemonkey script that’ll add the estimated profit to Swoopo above the price of an auction, updating in real time as people place bids.

Example screenshot of Swoopo Profits

It took quite a bit of work to sniff out the prices from the page (I suspect they make it hard to scrape on purpose), but I’ve checked it out and the script works pretty well on current and recent auctions on all of Swoopo’s different sites (US, Canada, UK, Germany, Austria, and Spain). It won’t work on some of their older auctions, where the rules were slightly different (and bid costs were different, too). The basic formula looks like this:

((currentPrice - bidAmount) / bidAmount) * bidCost + currentPrice - worthUpTo

I’m calculating it with all the fairness to Swoopo I can muster. I calculate the number of bids based on the current price and the amount each bid moves the price (bidAmount) times the cost of bids (bidCost). The winner still has to pay the current price, so I add that in, but I subtract what Swoopo says the item is “worth up to” since they probably have to pay around that to drop-ship it to a customer. As the example screenshot shows, this leads to examples like an iMac selling for $364.75 (plus another $392.40 in bids for the winner), but a total pure profit of $9,827.98 for Swoopo. Exciting! I’ll readily admit that my calculation is not always 100% accurate. There are a number of things I don’t take into account – I assume shipping is a wash, so it’s not included. I assume Swoopo’s paying the full retail “worth up to” price when they’re probably not. I count bids as all costing the same even though they might have been won at a “discount” via a bid auction. In cases where I can’t figure out some numbers I default them to hardcoded values, which might be wrong. I also don’t take into account “Swoop it now”, which lets bidders buy the item for its full price minus the money they’ve sunk into bids, effectively getting out of the auction entirely. This would reduce Swoopo’s profits but it isn’t recorded anywhere so I can’t factor it in. So take the number with a grain of salt – it’s entertainment.

Grab the script and start poking around swoopo.com. Hopefully you’ll have as much fun as I have with it.

As a side note about my Greasemonkey scripts, I’ve retired the “Amazon Super Saver Snooper” script. Amazon has changed the way the API used to look up Super Saver eligability works, and I can’t get at that data anymore. More importantly, Amazon now puts “Prime” logos next to Prime-shippable items, so the script isn’t necessary anymore.

Updates to JSONView, XBList, and TopProcess

I’ve finally gotten around to doing some maintenance updates on three of my most-used bits of software (JSONView, XBList, and TopProcess), all in the last couple weeks. Now that they’re all approved and live, I thought I’d summarize what’s changed.

JSONView 0.3

JSONView 0.3 is now available at addons.mozilla.org. This fixes an error that was showing up when really large JSON files were being displayed, and adds a feature that displays empty arrays and object on one line instead of on two (and doesn’t display the expand/collapse button for them). I also put in some preliminary support for JSONP, based on a patch submitted by Gabriel Barros. The catch is that it only works with content served with the “application/json” MIME type, which is actually not the correct MIME type for JSONP – it should be “text/javascript” or another JavaScript type. Unfortunately I haven’t been able to figure out how to get Firefox to let me handle “text/javascript” the way I do with “application/json” – it seems to be special-cased or something. If any Firefox gurus are reading and know a solution, please let me know! I was holding onto this release for a while hoping to fix that, but I decided that releasing something was better than nothing, so it’s out there and you can play with it. You can see an example JSONP response with the callback highlighted here. I did notice a bug with the 0.3 release right after it was approved – if a value is 0 or false, it shows up as blank. I’ll have that bug fixed with the next release, which should be very soon. It’s great to see that JSONView has become so popular, with over 50,000 downloads and about 12,000 regular users.

XBList 3.2.4

Halo 3: ODST was released a couple weeks ago, and Bungie redesigned their site a bit to include ODST info in your Service Record. This included changing the pages just enough that XBList got confused trying to figure out emblems for people who had never played Halo, and started showing the wrong emblems for those people (only if you prefer Halo emblems to Xbox emblems, which isn’t the default). Not a huge bug, but I took the opportunity to clean up a bunch of stuff in XBList and put out a little release. In addition to the Halo emblem fix, I fixed a bug where your settings could get corrupted and you wouldn’t be able to start XBList. I also cleaned up the menus and settings, consolidating the Halo links into one item, and removing the option to choose a custom notification sound or turn off debug logging. The debug log is much easier to get to if you’re having trouble – previously I had people dig through their Application Data folders for the log, but now there’s an item in the Help menu that opens the debug log directly. Paired with some much more detailed logging, I should be able to fix future problems much more easily. I also changed the system tray icon to bring XBList to the front on a single click rather than a double click, since that feels much more natural in Windows 7. Lastly (and probably not too importantly for most people), I’m storing emblems in your local Windows profile now, instead of the roaming profile. For the few people who use XBList and have roaming profiles enabled, this should save you some sync time. Anyway, you should have been prompted to update when I pushed the update two weeks ago, but if you haven’t gotten it yet you should download and install it now. I still have plans for a major overhaul of XBList, but it’s low on my list of priorities – XBList is still pretty popular, but not as much as it used to be, with maybe 5,000 active users.

TopProcess 1.4

This last update has been a long time coming. Ever since I installed Internet Explorer 8, the TopProcess sidebar gadget has been randomly crashing. I’m not sure what changed in Internet Explorer to make it screw up every so often, but it does. However, it only crashes after running for a few days, so for months I’ve been tweaking the code a bit, then waiting until it crashes, then tweaking some more, rinse, repeat. I’ve finally nailed it down to the point where crashes are very rare, and then I added some code that automatically resets the gadget when it crashes. So you should never see it mess up again. There is also a gadget log file in the gadget’s install directory that I’m using to store errors, so it should be easier to troubleshoot in the future. Lastly, Jean-Pierre van Riel contributed a patch that added IO tracking to TopProcess, so there is now a third way to view your processes. This view shows you the total IO usage (combined bytes in and out, per second) which seems to cover both disk and network access. Get the update from me or from Windows Live Gallery and be sure to rate the gadget. I’m excited to see that it’s been downloaded over 100,000 times from WLG (plus who knows how many downloads from my site).