This morning I was alerted that there was a new rev of the Kindle firmware coming - 2.0.3. It was, of course, a post on the user community MobileRead that did it. It's a testament to the user community (and to the Kindle's popularity) that there're fans rabid enough to tell eachother when a new firmware is out. Amazon has typically been very cryptic about what's actually in the firmware - they don't post release notes - but also very forthcoming in what's in it by posting the sources used.
Now, these aren't the sources to the Kindle's proprietary software, just the underlying open source OS changes. Still, clues can be divined from this, even if it's tasseomancy-style.
Downloading the latest source code release from Amazon,
The gplresults.tar inside - this is the tarball for the release - was dated 4/14 for a drop downloaded this morning.
Initial cursory comparison shows the dates on 31 files changed from 03/06/2009 for the 2.0.2 (309510017) release vs the 04/14/2009 for the current. I'll get around to looking at the linux-2.6.22-lab126 changes in more depth, like I did last time in I don't know Linux - Kindle update 2.0.2 (309510017).
Other than the date, + indicates increase in filesize from 2.0.2, - decrease in filesize:
+ alsa-lib-1.0.13_patch.tar.gz,
- alsa-utils-1.0.13_patch.tar.gz,
- busybox-1.7.2.tar.bz2,
- dosfstools-2.11.tar.bz2,
- e2fsprogs-1.38_patch.tar.gz,
- klibc-1.5.tar.bz2,
- linux-2.6.22-lab126.tar.bz2,
- procps-3.2.7_patch.tar.gz,
- taglib-1.5.tar.bz2,
+ uboot-1.3.0-rc3.tar.bz2,
- udev-112.tar.bz2,
202 - 203 Comparison Report
Twitter's REST timelines (public, friends, users) return created_at dates in a way C# can't natively parse, so they have to be reformatted. Apparently, this format is whatever comes out of Ruby. Here's my solution, elegance not included.
public static string ReformatTwitterDate(string twitterDate)
{
// twitter date: "Sat Apr 11 05:22:54 +0000 2009"
// DateTime.Parse()'able date: "Sat, Apr 11 2009 05:22:54 +0000"
MatchCollection c = Regex.Matches(twitterDate, @"\w+");
return string.Format("{0}, {1} {2} {7} {3}:{4}:{5} +{6}", c[0], c[1], c[2], c[3], c[4], c[5], c[6], c[7]);
}
... or (without regexp) ..
public string ReformatTwitterDate2(string twitterDate)
{
string[] c = twitterDate.Split(' ');
return string.Format("{0}, {1} {2} {5} {3} {4}", c[0], c[1], c[2], c[3], c[4], c[5]);
}
And yes, they know the Search API's date format is different.
Jesse Vincent over at Massively Parallel Procrastination has done something very impressive - conversion of pdfs and epubs directly on the Kindle2 via his program Savory
The modification uses igorsk's update maker modification and a modified version of Calibre to do the changes.
I can't wait to try it out and see how it goes!