tag:blogger.com,1999:blog-9324327349092259772024-02-20T11:31:19.174-08:00[Witty Blog Title]Random musings of a rabid developerJesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.comBlogger30125tag:blogger.com,1999:blog-932432734909225977.post-23498052732442018412011-11-20T23:41:00.000-08:002011-11-20T23:41:07.410-08:00smali/baksmali 1.3.0smali/baksmali 1.3.0 is out! This version now supports Honeycomb and Ice Cream Sandwich<br />
<br />
<br />
<a name='more'></a>There are a few changes and new features you should be aware of. <div><br />
</div><div><b>--api-level/-a option</b><br />
<br />
Both smali and baksmali have a new --api-level/-a option. For smali, this ensures that the dex file that smali generates will be compatible with the specified platform. For baksmali, this is primarily needed while deodexing.<br />
<br />
Specifically, when deodexing a Honeycomb or earlier odex file, you <b>must</b> specify an api level.<br />
<br />
<b>New ICS opcodes</b><br />
<br />
ICS brings a slew of new "jumbo" opcodes. These opcodes accept a 32-bit method/field/class reference, rather than the 16-bit reference that is used everywhere else. This eliminates the earlier restriction of 64k method or field references. However, there can still only be 64k referenced classes, due to other constraints in the dex file format.<br />
<br />
A list of the new jumbo opcodes can be found in the dalvik-bytecode.html document in the dalvik project in aosp (dalvik/docs/dalvik-bytecode.html). Unfortunately, the online document at <a href="http://s.android.com/tech/dalvik/dalvik-bytecode.html">http://s.android.com/tech/dalvik/dalvik-bytecode.html</a> has not been updated for ICS yet.<br />
<br />
<b>Custom inline tables for deodexing</b><br />
<br />
baksmali now allows you to specify a custom inline method table to use while deodexing. There have been some reports of versions of dalvik that are using a non-standard inline table. This will allow baksmali to be able to deodex odex files from these platforms.<br />
<br />
In order to use this feature, you must first dump the inline table from the device. I have written a small native utility to do this. You can find the source <a href="http://code.google.com/p/smali/source/browse/#git%2Fdeodexerant">here</a>. (I am not planning on distributing a prebuilt binary of this tool)<br />
<br />
You can dump the results of this tool to a file and then use the -T option to tell baksmali to use the custom inline table.<br />
<br />
<br />
As always, you can find the latest versions at <a href="http://code.google.com/p/smali/downloads/list">http://code.google.com/p/smali/downloads/list</a></div>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com5tag:blogger.com,1999:blog-932432734909225977.post-5720962559443614502011-08-22T18:13:00.000-07:002011-08-22T18:13:18.492-07:00Reporting issues in smali/baksmaliDue to an issue in the 1.2.7 release of smali/baksmali, I've noticed an interesting trend - When people find an issue in smali/baksmali, they seem to post to <insert favorite forum here>, and then proceed to completely neglect to tell me about it.<br />
<br />
<a name='more'></a><br />
For this particular issue, luckily, someone did ping me on #smali on freenode to report the issue. This was perfect, because I was able to talk to them, figure out what was going on, and reproduce the issue. And of course, to then fix it <br />
<br />
However, I didn't think this particular issue would be encountered very often, so I decided to queue up the fix for the next release. And I didn't hear anyone complaining, which seemed to back up my hypothesis that it was a rare issue that wouldn't normally be encountered.<br />
<br />
Luckily, every once in a while I do a "last 24 hours" or "last week" google search for smali and baksmali, just to see what's up. And I noticed several reports of this issue on various forums (mostly xda-dev). I'm somewhat baffled as to why they didn't file an issue or even let me know about it. I can't fix the tool if people don't let me know it's broken!<br />
<br />
Once I noticed that people were encountering the issue more frequently than I expected, I decided to go ahead and make a new release that fixes the issue.<br />
<br />
The moral of the story is simple - Please Report Bugs!JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com0tag:blogger.com,1999:blog-932432734909225977.post-37058582401154044952011-08-22T17:52:00.000-07:002011-08-22T17:52:11.028-07:00smali/baksmali 1.2.8Just a quick bugfix release to address an issue people were running into with 1.2.7, where baksmali would hang indefinitely in some cases.<br />
<br />
<a name='more'></a>Download <span id="goog_80535766"></span><a href="http://www.blogger.com/">here<span id="goog_80535767"></span></a>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com0tag:blogger.com,1999:blog-932432734909225977.post-64797940922150398432011-08-13T17:21:00.000-07:002011-08-13T17:24:49.250-07:00smali/baksmali 1.2.7smali/baksmali 1.2.7 is out, after entirely too long of a wait since the last release.<br />
<a name='more'></a><br />
This is mostly a bugfix release, although there was one new feature added.<br />
<br />
Noteworthy changes:<br />
<ul><li>Fixed a significant deodex issue, related to field accesses being resolved incorrectly. Kudos to brightidea @ xda-dev for investigating this issue!</li>
<li>baksmali will now add a comment before synthetic accessor methods that shows what field/method is actually being accessed. Kudos to jasta for suggesting this feature!</li>
</ul>And thanks to everyone else who reported the various issues that was fixed in this release.<br />
<br />
You can find the new versions at the <a href="http://code.google.com/p/smali/downloads/list">usual place.</a>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com0tag:blogger.com,1999:blog-932432734909225977.post-5051846007856619812011-08-12T21:22:00.000-07:002011-08-12T21:22:45.436-07:00Welcome!All ye of jf.andblogs.net origin, welcome!JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com0tag:blogger.com,1999:blog-932432734909225977.post-87732707989822197372010-12-23T07:14:00.000-08:002011-08-12T21:08:20.399-07:00smali/baksmali 1.2.6smali/baksmali 1.2.6 is out, now with support for gingerbread odex files. A few other minor fixes are also included. You can grab the new version <a href="http://code.google.com/p/smali/downloads/list">here</a>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com5tag:blogger.com,1999:blog-932432734909225977.post-32827906079109852112010-10-31T08:42:00.000-07:002011-08-12T21:07:09.594-07:00smali/baksmali 1.2.5 releasedJust a minor release with a few bugfixes that have gone in since the last release<br/><br/><a name='more'></a><br/><br/>Get it at <a href="http://code.google.com/p/smali/downloads/list">http://code.google.com/p/smali/downloads/list</a>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com0tag:blogger.com,1999:blog-932432734909225977.post-52804570652776192442010-08-26T17:16:00.000-07:002011-08-12T21:07:09.594-07:00smali/baksmali used for great evilI wanted to take a moment to respond to the recent buzz about using smali/baksmali to break the new Google licensing framework for the Android market.<br/><br/><a name='more'></a><br/><br/>I <strong>do not</strong> support/encourage/condone the use of my tools for any form of application piracy, copyright infringement, etc. Given the nature of the tool, it's natural that it will be used for these types of things. Some people even seem to think that piracy is the sole purpose of these type of tools. However, there are plenty of legitimate uses as well.<br/><br/>I'll admit, when I first heard the details of the new licensing scheme, my first thought was "I bet that will be easy to crack with smali/baksmali". But I can almost guarantee that the Google engineers who designed the licensing framework took this possibility into account as well. I hear they're usually pretty bright over there at Google. To think that they didn't consider it is almost absurd. What they did was to raise the bar on the effort required to pirate applications. Sure, it's still possible, but I think that overall it will have a positive effect on piracy rates.<br/><br/>I do find it interesting on an intellectual level to try and break these types of protections. Just to see if I can do it/if it's possible. There's a difference between cracking something just to see if you can crack it, and cracking something with the sole purpose of pirating applications. One is good (or at least neutral). The other is evil.<br/><br/>So in the spirit of Google's motto, I'll conclude by imploring everyone to use my tools for non-evil.JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com4tag:blogger.com,1999:blog-932432734909225977.post-26214975787626810122010-08-01T13:38:00.000-07:002011-08-12T21:07:09.594-07:00More smali/baksmali bugfixes1.2.4 is out, with a number of bugfixes<br/><br/><a name='more'></a>Get it <a href="http://code.google.com/p/smali/downloads/list">here</a><br/><br/>Changelist:<br/><ul><br/> <li><a href="http://code.google.com/p/smali/source/detail?r=784">Fixed</a> <a href="http://code.google.com/p/smali/issues/detail?id=35">an issue</a> when using baksmali on a case insensitive file system, and there are classes with names that only vary by case (thanks phd.dre and Brut.all!)</li><br/> <li><a href="http://code.google.com/p/smali/source/detail?r=764">Fixed</a> <a href="http://code.google.com/p/smali/issues/detail?id=37">an issue</a> where baksmali still printed out .line and .local directives even when the -d/--no-debug-info option is specified (thanks stari4ek!)</li><br/> <li><a href="http://code.google.com/p/smali/source/detail?r=766">Fixed</a> <a href="http://code.google.com/p/smali/issues/detail?id=39">an issue</a> in smali when in encountered a field or method that was named "null" (thanks dislam!)</li><br/> <li><a href="http://code.google.com/p/smali/source/detail?r=776">Fixed</a> <a href="http://code.google.com/p/smali/issues/detail?id=40">an issue</a> in smali when encountering a file that has blank annotation sets (thanks flipz and jjiovani!)</li><br/> <li><a href="http://code.google.com/p/smali/source/detail?r=762">Fixed</a> an issue when there is a <clinit> method with no corresponding code item</li><br/></ul>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com8tag:blogger.com,1999:blog-932432734909225977.post-91175732093246457962010-06-13T09:14:00.000-07:002011-08-12T21:07:09.594-07:00Yes, I would like some frozen yogurt with my baksmali pleasesmali/baksmali v1.2.3 is out, and now includes the ability to deodex odex files from froyo.<br/><br/><a name='more'></a><br/><br/>You can download the new versions on the <a href="http://code.google.com/p/smali/downloads/list">downloads page</a> of the <a href="http://code.google.com/p/smali/">googlecode project</a><br/><br/>In addition to the changes related to supporting froyo, the new version has a new lexer that uses much less static memory. (The old lexer had around 30mb of static data. ugh!). The new lexer is written with jflex instead of ANTLR, but it still hooks in with the ANTLR parser.<br/><br/>There was also a significant change in how "undeodexable" instructions are handled. If you want to read up on the details, go take a look at the comments in <a href="http://code.google.com/p/smali/issues/detail?id=29">issue 29</a> and <a href="http://code.google.com/p/smali/issues/detail?id=33">issue 33</a>. Thanks to Paul from modaco.com, brut.all, master.homer, flipz and JrEE2kX for the reporting/discussion/testing around this issue.<br/><br/>As usual, there are a number of little performance tweaks here and there, and a few new bug spots on the bottom of my shoes.JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com3tag:blogger.com,1999:blog-932432734909225977.post-15371241757903972862010-04-03T11:29:00.000-07:002011-08-12T21:07:09.595-07:00YABBFRYet Another Baksmali Bugfix Release. <a href="http://code.google.com/p/smali/downloads/list">v1.2.2</a> is out for your disassembling pleasure.<br/><br/><a name='more'></a>As the title suggests, this is mostly a bugfix release, fixing issues <a href="http://code.google.com/p/smali/issues/detail?id=20">20</a>, <a href="http://code.google.com/p/smali/issues/detail?id=21">21</a> and <a href="http://code.google.com/p/smali/issues/detail?id=22">22</a>.<br/><br/>I've also made some significant performance enhancements to baksmali. You can expect deodexing a full firmware to be roughly 50% faster, while performing certain operations on single files can be over 100% faster (dumping the full register info of a large dex file, for example).<br/><br/>Finally, I've added the ability to read the base BOOTCLASSPATH dependencies from odex files directly, so you shouldn't have to specify the full BOOTCLASSPATH (-c) for roms that use a non-standand BOOTCLASSPATH (like HTC Sense roms). Note however, that you will still need to add the "extra" dependencies for specific applications that need it, because those dependencies aren't stored in the odex file.<br/><br/>Get it <a href="http://code.google.com/p/smali/downloads/list">here</a>.JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com1tag:blogger.com,1999:blog-932432734909225977.post-26291702011699285062010-03-05T17:24:00.000-08:002011-08-12T21:07:09.595-07:00smali/baksmali v1.2.1 bugfix releaseI've released v1.2.1, which is mostly a bugfix release to v1.2. There are a few performance improvements and enhancements in there as well.<br/><br/><a name='more'></a>As usual, you can grab a copy on the <a href="http://code.google.com/p/smali/downloads/list">googlecode download page</a>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com4tag:blogger.com,1999:blog-932432734909225977.post-75147681452621765262010-02-21T19:27:00.000-08:002011-08-12T21:07:09.595-07:00smali/baksmali v1.2 releasedAfter <a href="http://code.google.com/p/smali/source/list?start=645">lots of hard work</a> over the last month or two, <a href="http://code.google.com/p/smali/downloads/list">smali/baksmali 1.2</a> is out!<br/><br/>The major new functionality in this release is that baksmali now supports deodexing without the help of deodexerant! It also has a new "register info" feature, to show the register types in the disassembly, and numerous minor fixes/changes/enhancements/tweaks/(and probably bugs)<br/><br/><a name='more'></a><br/><h4>Deodexing</h4><br/>In order to deodex files now, you need to have the boot class path files available for baksmali to use. By default, it looks for the 5 main framework jars in the current directory. You can of course specify additional directories to search in, add additional boot class path files, or change which boot class path files are used altogether.<br/><br/>The <a title="DeodexInstructions" href="http://code.google.com/p/smali/wiki/DeodexInstructions">DeodexInstructions</a> page has more info on how to deodex with this version. But for a quick primer, you just need to have the 5 main framework files in the current directory (core.jar, ext.jar, framework.jar, android.policy.jar and services.jar), and then specify the <em>-x</em> option for baksmali. For example:<br/><pre>baksmali -x Calculator.odex -o Calculator</pre><br/><h4>Register Info</h4><br/>Another bit of new functionality that can be very helpful is the new "register info" output for baksmali, which can be turned on with the <em>-r</em> parameter. It will analyze the registers and print some register type info before and after each instruction. There are several levels of register info output available, depending on exactly what you want to see. The default is to print the register type for any register that is used by the instruction.<br/><br/>Note that this functionality also requires that baksmali load the boot class path files - so they must be available. Here is an example of what the default register info looks like:<br/><pre>#v0=(Integer);v2=(Integer);<br/>new-array v2, v0, [C<br/>#v2=(Reference,[C);</pre><br/>The register types that are printed just before the instruction are the <strong>incoming</strong> register types, while the register types that are printed after the instruction show any changes to the registers caused by the instruction.<br/><br/>If you want to see <strong>all</strong> the register info, you can use <em>-r ALL,FULLMERGE</em> which looks something like this:<br/><pre>#v0=(Integer):merge{0x18:(Null),0x2c:(Integer)}<br/>#v1=(Conflicted):merge{0x18:(Uninit),0x2c:(Integer)}<br/>#v3=(Conflicted):merge{0x18:(Uninit),0x2c:(Char)}<br/>#v2=(Reference,[C);p0=(Reference,Ljava/lang/String;);p1=(Reference,[B);p2=(Integer);p3=(Integer);p4=(Integer);<br/>iget v2, p0, Ljava/lang/String;->count:I<br/>#v0=(Integer);v1=(Conflicted);v2=(Integer);v3=(Conflicted);p0=(Reference,Ljava/lang/String;);p1=(Reference,[B);p2=(Integer);p3=(Integer);p4=(Integer);</pre><br/><h4>Other changes</h4><br/>There are a few other miscellaneous changes as well. Make sure you take a look at the usage info for smali and baksmali. The short parameters for some of the options have changed. In particular some of the options that are mostly for debugging purposes were changed to an uppercase letter, and are now hidden by default. You can use -?? for both smali and baksmali to see the debug options.<br/><br/>baksmali also has a new -f parameter, which adds a comment with the code address before each instruction. This is useful when looking at the FULLMERGE register info, which shows the register info and code addresses for all "incoming" execution paths.<br/><h4>Things to come</h4><br/>With this release, I have added a robust code analyzer/verifier that can infer the register types and validate the instructions. I plan using this to add verification functionality in smali, so that it will optionally verify the code after assembling it. This will let you know there's a problem with the assembled code without having to push the code to a device and have dalvik complain to you about the invalid code.<br/><br/>I also want to add some way to dump/serialize the results of loading the boot class path files for baksmali, so that it can load the information it needs from the dump file, instead of reading in all 5 boot class path files every time, which should help speed it up.<br/><br/>In the longer term, I would love to be able to debug code on a device at an assembly level. This is just something that is banging around in the back of my head for now.JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com7tag:blogger.com,1999:blog-932432734909225977.post-16067467129130898232009-12-25T11:28:00.000-08:002011-08-12T21:07:09.595-07:00Have a very smali Christmas!As my Christmas present to everyone, I present smali/baksmali v1.1<br/><br/><a name='more'></a><br/><br/>This version has significant new functionality, and is much better able to handle the larger framework files that are floating around. There are a number of bug fixes as well.<br/><br/>As usual, you can grab the new versions from the <a href="http://code.google.com/p/smali/downloads/list">googlecode download page</a><br/><br/>The changes for this version include:<br/><ul><br/> <li>Added support for a new .locals directive, which can be used instead of the .registers directive, to specify the number of non-parameter registers in the method</li><br/> <li>Added support for the <code>--use-locals/-l</code> option in baksmali, to use the .locals directive instead of the .registers directive</li><br/> <li>Added support for the <code>--sequential-labels/-q</code> option in baksmali, which causes baksmali to create label names using a sequential numbering scheme, instead of basin it on the bytecode address</li><br/> <li>Added support for automatically upgrading const-string to const-string/jumbo when needed and upgrading goto or goto/16 to goto/16 or goto/32 when needed</li><br/> <li>Added support for the <code>--no-fix-string-const/-c</code> option in smali, to prevent automatically upgrading const-string instructions to const-string/jumbo</li><br/> <li>Added support for the <code>--no-fix-goto/-g</code> option in smali, to prevent automatically upgrading goto and goto/16 instructions</li><br/> <li>Added support for the <code>--no-debug-info/-b</code> option in baksmali, to prevent the output of any type of debugging info (.local, .line, .parameter, etc.)</li><br/> <li>misc. bugfixes</li><br/></ul>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com9tag:blogger.com,1999:blog-932432734909225977.post-15166005126498972712009-12-22T18:45:00.000-08:002011-08-12T21:07:09.595-07:00Nook updatesNow that the <a href="http://www.nookdevs.com/">nook softroot</a> is out, I'd like to take a moment to explain the structure of the updates, the nook bootup process in general, and how the softroot came about.<br/><br/><a name='more'></a><br/><br/>The "softroot" that was released is simply an unsigned update file that copies a custom init.rc to the device that enables adb (only over wifi, for now). Unsigned update, you ask? Yes. The nook accepts unsigned updates. The structure of an update is fairly simple. Basically, it's a .tar.gz file that contains at least 3 files: Firmware_Update/update.script, Firmware_Update/md5sums and Firmware_Update/sha1sums.<br/><br/>The update.script file is a script that is run by turboboot - a proprietary bootloader/recovery environment that is used in the nook. Documentation for turboboot can be found easily with $favorite_search_engine. The turboboot environment is a minimal linux+busybox environment. It allows you to run arbitrary system commands from the update.script, so you can really do just about anything you want in an update.<br/><br/>For the md5sums and sha1sums file, the only requirement is that running md5sum -c md5sums and sha1sum -c sha1sums from the Firmware_Update directory will succeed. So you need to have at least 1 valid checksum in each file.<br/><br/>That's basically it for the update. You can add whatever else the update needs of course. To run the update, place the .tar.gz file on the root of the external sdcard, named bravo_update.dat, power off the nook, and then after powering it back on, immediately press the page up/page down buttons. It should then check for the update file, unpack it, validate the md5sums and sha1sums files, and then run the update.<br/><br/>If you've already taken a look at the 1.1.0 or 1.1.1 update files that have been released recently, you've probably noticed that they aren't just a plain .tar.gz, but they have what appears to be a signature header at the beginning of the file. As far as I can tell, when a new update is available, the nook system software will download the signed update to the 3rd partition of the internal sdcard, verify it and then strip off the signature header and place the plain .tar.gz in the root of the 2rd partition of the internal sdcard. It then sets a value in a specific place on the sdcard that tells turboboot to check the internal sdcard instead of the external sdcard for the update. And then when the nook reboots, turboboot finds the unsigned update file and applies it.<br/><br/>If you want to poke around in the 1.1.1 update, the file is available for download from <a href="http://www.barnesandnoble.com/nook/support/?cds2Pid=30195">Barnes and Noble's Nook FAQ page</a>. Once you download the update, you'll need to strip off the signature header before extracting it:<br/><br/><code>tail -c +174 signed_bravo_update.dat > signed_bravo_update.tar.gz</code><br/><br/>Finally, a bit more about the nook's start-up in general. The nook appears to boot directly from the sdcard. There is around 200mb of space at the end of the sdcard, past the end of the last partition, where various things are stored: bootloader, kernel, etc. These all appear to be stored at specific locations relative to the end of the sdcard.<br/><br/>The primary bootloader is stored at an offset relative to the end of the sdcard. For non-sdhc cards, this is at endofsd-9k. And there is additional bootloader code stored at endofsd-281k.<br/><br/>The kernel is stored at endofsd-4377k, and interestingly, the last full firmware update that was applied on the nook is stored at endofsd-135449k. And finally the kernel and ramdisk for the turboboot environment is stored at endofsd-139545k.<br/><br/>The ramdisk for the turboboot environment contains an interesting script named initboot.script that handles the update process. One of it's features is that if the nook fails to successfully boot 3 times in a row, then it will perform a "fallback", and will wipe out the OS partition and re-apply the full update that is stored in the ending area of the sdcard.<br/><br/>It was when I was first looking at the initboot.script file that I realized that the nook accepted unsigned updates. Once I realized that, I pinged fellow nookdev robbiet480 (because he has a nook, and I don't yet :/) and after a few tries, was able to build an initial proof-of-concept update that he ran on his nook and proved that we could in fact get an unsigned update to run.<br/><br/>From there, the other nookdev members took it from the proof-of-concept that I did to the actual softroot that is available now. Great job guys!<br/><div id="_mcePaste" style="overflow: hidden;width: 1px;height: 1px">$favorite_search_engine</div>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com6tag:blogger.com,1999:blog-932432734909225977.post-2411818195597181082009-11-07T18:30:00.000-08:002011-08-12T21:07:09.596-07:00smali/baksmali v1.0It's 1.0. Yep, sure is. Go get it <a href="http://code.google.com/p/smali/downloads/list">here</a>.<br/><br/>Why should you get the new version you ask? Well, it's been spiffified, of course. The new spiffiness includes:<br/><ul><br/> <li>The label syntax has changed from "label:" to ":label". This makes the lexer happy. And a happy lexer uses less memory and runs faster. Yay for happy lexers</li><br/> <li>Added a -r/--rewrite-labels option for smali, which will rewrite the smali files you give it, changing any labels in the old format to the new format.</li><br/> <li>deodexerant now opens the odex files as read-only, so you can run deodexerant directly on the odex files in /system, without having to remount, or copy them to /data.</li><br/> <li>You can give baksmali an apk file now, so you don't have to manually extract out the classes.dex file</li><br/></ul><br/>And one final note: deodexerant will *not* work on 2.0 devices. Once the 2.0 source is released, I should be able to build a new deodexerant binary that will work on 2.0.<br/><br/>Update: There are now also wrapper scripts available for smali/baksmali, so you can call "baksmali ..." instead of "java -jar baksmali.jar ..."<br/><br/>Update 2: Now that the Eclair source is out, I've added a new build of deodexerant to the downloads page that works with 2.0JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com22tag:blogger.com,1999:blog-932432734909225977.post-366134969705130512009-09-10T16:32:00.000-07:002011-08-12T21:07:09.596-07:00slay those pesky odex files!odex files getting you down? Now you can do something about them!<br/><br/><a name='more'></a>For the first time ever, with <a href="http://smali.googlecode.com">smali/baksmali 0.96</a>, and some help from <a href="http://code.google.com/p/smali/wiki/DeodexInstructions">deodexerant</a>, you can get rid of that optimized dex odor from your rom.<br/><br/>It slices! It dices! It turns odex files into .smali files, capable of being assembled into normal classes.dex files! And all for the low low price of $0.00. Can't afford $0.00 right now? That's ok! I guarantee I can get you low monthly payments of $0.00 for 12 months (offer good while supplies last)<br/><br/>Order now! Operators are <a href="http://code.google.com/p/smali/downloads/list">standing by</a> (they can also help with <a href="http://code.google.com/p/smali/wiki/DeodexInstructions">deodex instructions</a>)JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com8tag:blogger.com,1999:blog-932432734909225977.post-13693332132643389892009-08-29T09:38:00.000-07:002011-08-12T21:07:09.596-07:00A new releaseof <a href="http://smali.googlecode.com">smali/baksmali</a>. (no, not a new rom :p)<br/><br/><a name='more'></a>smali/baksmali v0.95 is out. The major change in this version is a re-implemented/optimized version of dexlib, as well as changes in smali/baksmali to work with the new dexlib. Also, I've optimized baksmali, so it should run much much quicker now (up to 4x quicker). smali should also be a bit quicker, but nothing you'll probably notice.<br/><br/>As far as new functionality goes, baksmali will now output registers that are mapped to method parameters using a p<n> syntax, instead of the normal v<n> syntax. i.e. p0 is the first method parameter (or the "this" reference, for non-static methods). This makes it much less of a pain when you need to increase the number of registers in a method. Previously, you would have to go through and bump up the number for all of the parameter registers, but with the p<n> syntax, you won't have to. If you don't want this functionality, you can use the -p command line argument to turn it off.<br/><br/>The new versions can be found at the <a href="http://code.google.com/p/smali/downloads/list">project download page</a>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com1tag:blogger.com,1999:blog-932432734909225977.post-59686644040251288452009-08-20T12:15:00.000-07:002011-08-12T21:07:09.596-07:00Calling it quitsI've decided to call it quits on the whole rom making thing. As I'm sure you've noticed, I haven't done much work on the rom lately anyway, not to mention the way overdue CRC1 rom that I never got out. I'll still be around of course, working on smali and hanging out on freenode. But I just don't have enough interest in doing rom work anymore to continue.<br/><br/><a name='more'></a>I sincerely appreciate all the kind words about my roms, and I want to again thank everyone who has donated.<br/><br/>If you're looking for a new rom to try, I recommend cyanogen's <a href="http://forum.xda-developers.com/showthread.php?t=537204">stable</a> or <a href="http://forum.xda-developers.com/showthread.php?t=539744">experimental</a> roms. He's put a lot of work into them, and they have many more features than mine do/did.<br/><br/>Also, If anyone wants to grab my build environment and maintain a "JF-style" build on their own, go for it :) If you need any pointers on how to update the build environment as new official releases come out, feel free to ask (There really isn't all that much involved in it).JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com129tag:blogger.com,1999:blog-932432734909225977.post-60183965452647907002009-07-19T06:08:00.000-07:002011-08-12T21:07:09.597-07:00OTAs in JFv1.51If you are on the US or EU versions of JFv1.51, you'll likely get an OTA update notification soon, if you haven't already. I had accidentally left in the otacerts.zip file, which allows OTAs to be downloaded and verified.<br/><br/>For now, there are a few possible work arounds.<br/><br/>1. delete the otacerts.zip file manually. The file is at /system/etc/security/otacerts.zip But keep in mind if you use this method your phone will continuously re-down the OTA and try to verify it, which is bad on your bandwidth usage and your battery life<br/><br/>2. Replace /system/build.prop on your phone with the one from the ADP1 version of JFv1.51<br/><br/>3. chmod 000 the OTA file in /cache (unconfirmed solution, but is likely to work)<br/><br/>Or you can ignore the popups for now, and wait for my new release based on the new update (CRC1). I promise to "disable" the OTA mechanisms in a better way than simply deleting otacerts.zip :). I should be able to get something out soon (next few days)<br/><br/>In the meantime, even if you accidentally hit "accept" on the popup, it won't be able to install the update when it gets into recovery, so you aren't at risk of losing root access.JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com64tag:blogger.com,1999:blog-932432734909225977.post-7627457861865701512009-06-22T19:09:00.000-07:002011-08-12T21:07:09.597-07:00It's alive!Alive I tell you, Alive! <a href="http://code.google.com/p/smali/downloads/list">baksmali</a> lives!<br/><br/><a name='more'></a>baksmali is the companion disassembler to smali, the dex assembler that I <a href="http://jf.andblogs.net/2009/06/07/smali-v09/">released recently</a>. I've released a new version of smali as well, with a number of improvements/fixes. This is the v0.91 release of both tools.<br/><br/>As you might imagine, baksmali is able to disassemble a dex file into a set of text files, one for each class, containing the methods, field, annotations, debug info and dalvik bytecode for that class.<br/><br/>And then smali can munch on those text files and assemble them into shiny new classes.dex file, which you can repackage back into an apk/jar/zip file and throw on your android-based phone.<br/><br/>And most importantly, you can make magic happen when you make changes between the 2 steps.<br/><br/>Both tools are released as executable jar files. You can invoke them and get the usage information with<br/><br/><code>java -jar smali-0.91.jar<br/>java -jar baksmali-0.91.jar</code><br/><br/>Unfortunately, the current incarnation of the tools are somewhat porcine. If you try to assemble/disassemble a large dex file and get an out-of-memory type error, try running with -Xmx512m, which increases the max java heap size to 512MB<br/><br/><code>java -Xmx512m -jar smali-0.91.jar [options]<br/>java -Xmx512m -jar baksmali-0.91.jar [options]<br/></code><br/><br/>The main project page for smali/baksmali is <a href="http://smali.googlecode.com">http://smali.googlecode.com</a>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com14tag:blogger.com,1999:blog-932432734909225977.post-45484251039796697992009-06-07T15:19:00.000-07:002011-08-12T21:07:09.597-07:00smali v0.9!After lots and lots of work, the first release of <a href="http://smali.googlecode.com">smali</a> (dex assembler) is out finally!<br/><br/><a name='more'></a>I'll try and get a wiki going on the googlecode page with some documentation, as far as usage and syntax goes. But for now, you can look at the examples and tests to see the syntax. There are tests for every opcode, so you should be able to find the syntax you need.<br/><br/>Next up, I'll get started on baksmali, the disassembler companion to smali.JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com8tag:blogger.com,1999:blog-932432734909225977.post-37854465416210007602009-05-25T07:50:00.000-07:002011-08-12T21:07:09.597-07:00JFv1.51 Quickfix for Multitouch Browser issueThe browser in the US and UK versions of JFv1.51 have an issue where multi-touch doesn't work. This issue does not affect the ADP version. You can fix this issue yourself easily.<br/><br/><a name='more'></a>(In terminal emulator)<br/><pre>su<br/>mount -o rw,remount /system<br/>rm /system/app/Browser.odex<br/>reboot</pre><br/>Alternatively, I've also created a tiny update.zip file that you can apply in recovery mode to fix the problem. This is just a quickfix update that deletes the Browser.odex file.<br/>[download id="22"]<br/><br/>Just to be clear: if multi-touch in your browser is working, you don't need to do anything. If it is not working, you can follow the terminal emulator instructions above, or apply the update.zip quick fix.<br/><br/>Note: I've updated the US and UK downloads in the v1.51 release blog post to an updated version that has this fix. I didn't update the version number since this is a small problem that is easily fixed. I believe jf updater has been updated with the fixed builds as well.JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com62tag:blogger.com,1999:blog-932432734909225977.post-81225130956035390782009-05-24T06:34:00.000-07:002011-08-12T21:07:09.598-07:00JFv1.51 images are out!I've released a new set of images based on the US/ADP/EU versions of CRB43.<br/><br/><a name='more'></a><br/><br/>update: The US and UK builds have been updated to fix the MT browser issue. New build enviroments and CHANGES for them too. See <a href="http://jf.andblogs.net/2009/05/25/jfv151-quickfix-for-multitouch-browser-issue/">this post</a> for more info<br/><br/>----<br/><br/>There are no major changes from my JFv1.50 release, except a few new kernel modules.<br/><br/>As usual, I don't guarantee whether you'll have to wipe or not. My advice: Make a backup with nandroid then try the update without a wipe. If you run into problems, do a wipe or restore your backup.<br/><br/>They should be available on the JF Updater app. Or you can download and install them manually<br/><h6>[download id="13"] (672003d8164c37c3c088a80294ac86b2)<br/>[download id="14"] - 98a66d11f90fdb7ab4f2442a329d2519<br/>[download id="15"] - 85f4212794c142a65967a47a4d099304<br/>[download id="16"] - 6012fce5b60225e47e2de5762492a457<br/>[download id="17"] - 5af8c9ab3c2f818869bbb35ec1ed8628<br/>[download id="18"] - ba5b9c0488cce07ac1931623c74b6161<br/>[download id="19"]<br/>[download id="20"]<br/>[download id="21"]</h6>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com280tag:blogger.com,1999:blog-932432734909225977.post-43610395463324989242009-05-01T07:37:00.000-07:002011-08-12T21:07:09.598-07:00When is your 1.5 coming out?The answer is... wait for it.. wait for it..<br/><br/><strong>NOW!</strong><br/><br/><a name='more'></a>If only i <a href="https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=1014600">had a nickel</a> for every time someone asked me that over the last week or two :p<br/><br/>This build is based on the <a href="http://www.htc.com/www/support/android/adp.html">official HTC ADP1.5 image</a>. <a href="http://www.htc.com/www/support/android/adp.html"></a>You should install the 1.5 radio from that HTC page before using this build.<a href="http://www.htc.com/www/support/android/adp.html"><br/></a><br/><br/>Some of the important changes/features<br/><ul><br/> <li>I'm now including a much better Superuser app, courtesy of zinx</li><br/> <li>I've rebuilt the kernel and enabled netfilter support, to allow tethering.</li><br/> <li>You can see/download both paid and protected apps in the market</li><br/> <li>Luke's <a href="http://lukehutch.wordpress.com/2009/01/25/get-multi-touch-support-on-your-t-mobile-g1-today/">multitouch kernel patch and browser patch</a> have been ported to the .27 kernel and 1.5 browser.</li><br/> <li>Support for the "apps to sdcard" hack. There's not a separate "apps to sdcard" build. You can use the same build whether you are doing apps to sdcard or not. Thanks to haykuro for explaining how he's done this in his images, so I can make my builds compatible. See <a href="http://android-dls.com/wiki/index.php?title=Haykuro_Apps_to_SD">here</a> for a tutorial on setting this up.</li><br/> <li>I'm not including a recovery image in this build. I ran into some snags getting all of the recovery features to work with the 1.5 recovery image, so I'll forgo that for now. If you don't have a spiffy recovery image yet, I've packaged up the most recent recovery image from JFv1.43 as a separate update.</li><br/></ul><br/>If you are already running a "rooted" build (ADP/mine/haykuros/the dudes/whatever), it is available for download via the <a href="http://code.google.com/p/jfupdater/">jf updater</a> app (available on the market). If you don't see it, open up the app settings, and choose either ADP1 or "All the updates" for the "System Mod Version" setting.<br/><br/>Or you can always download the update.zip file below and install it yourself. If you have a G1 and this is your first time installing a community firmware, you'll need to follow any of the multitudes of tutorials out there for rooting your phone.<br/><br/>You can try to upgrade without a wipe, but if you run into any problems (for example, su/Superuser doesn't work), you'll need to perform a wipe. If you don't want to wipe, take a nandroid backup before upgrading, so you can revert back to the backup.<br/><br/><strong>[download id="9"]</strong><br/><br/><strong>[download id="10"]</strong><br/><br/><strong>[download id="11"]</strong><br/><br/><strong>[download id="12"]<br/></strong>JesusFrekehttp://www.blogger.com/profile/12165265467569950886noreply@blogger.com402