I imagine a challenge would be drools generation of bytecode via asm etc. Not sure how that works on phone with no dalvik compiler. If rules precompiled (keep think smaller for dalvik), and class generation not happening, it might just work. Sent from my phone. On 16/04/2010, at 3:16 PM, Justin King <justin.matthew.king(a)gmail.com&gt; wrote:

I was under the impression that the problem might lie in the rules being compiled at runtime into java bytecode (and not the required dalvik)? I'm attempting to port over a framework that uses both Drools and Javassist and I was thinking both libraries are going to cause the same problems. Am I on the right track here? On Fri, Apr 16, 2010 at 9:52 PM, Salaboy <salaboy(a)gmail.com&gt; wrote: -- Regards, Justin King PhD Candidate Faculty of Information and Communication Technologies Swinburne University of Technology http://www.ict.swin.edu.au/ictstaff/justinking

I thought the RETE tree should be no problem, as its just some in memory structure. I was thinking the right hand side of the rules would be the issue as java classes are generated for those. My thoughts were the generated java bytecode would not be compatible with Android. On Mon, Apr 19, 2010 at 1:25 PM, Mauricio Salatino <salaboy(a)gmail.com&gt;wrote: -- Regards, Justin King PhD Candidate Faculty of Information and Communication Technologies Swinburne University of Technology http://www.ict.swin.edu.au/ictstaff/justinking

On 28/12/2011 09:27, profversaggi wrote: Kris has shown jBPM works on android with a small change to the classloader. I have heard of people making Drools work on Android, but there is nothing formal. So at this stage I'd say "maybe". Give it a try, find out what the problems are and get patches back to the project to help. Mark

Oh.. that's cool.. I will take a look on that.. On Mon, Jan 2, 2012 at 9:58 AM, profversaggi <profversaggi(a)gmail.com&gt; wrote: --  - CTO @ http://www.plugtree.com  - MyJourney @ http://salaboy.wordpress.com  - Co-Founder @ http://www.jugargentina.org  - Co-Founder @ http://www.jbug.com.ar  - Salatino "Salaboy" Mauricio -

Here's what I got working and when it failed .... it was just the basic 'hello world' kind of stuff: Worked: final KnowledgeBuilder kbuilder = KnowledgeBuilderFactory.newKnowledgeBuilder(); Died: // this will parse and compile in one step kbuilder.add(ResourceFactory.newClassPathResource("HelloWorld.drl", DroolsTest.class), ResourceType.DRL); LogCat: 01-02 13:13:17.684: E/AndroidRuntime(537): java.lang.RuntimeException: Unable to start activity ComponentInfo{com.versaggi.android.droolsandroidtest/com.versaggi.android.droolsandroidtest.DroolsAndroidTestActivity}: java.lang.NullPointerException That's about as far as I could get. I think this DOC will probably help out .... https://github.com/cgreenhalgh/droolstest/blob/8ea1a06d9b1b38a40b1add7f98.... .. I had to re-compile the CompositeClassLoader.java file and then reinsert it into the drools-api-5.1.0.M2.jar file to get past this issue: "A small modification is needed in org.drools.util.CompositeClassLoader. Apparently, the Android JVM doesn't allow null as the parent classloader so I changed that to super(CompositeClassLoader.class.getClassLoader()) in the constructor, and that works fine." However I think that his other modifications aren't completely integrated into the drools-compiler-5.1.0.M2.jar or the drools-core jar files he provides just yet. ********************************* On Mon, Jan 2, 2012 at 2:11 PM, Wolfgang Laun <wolfgang.laun(a)gmail.com&gt;wrote:

No .. the flavor of what I was interrogating of his code and docs seemed to indicate that both the parsing and compiling of the DRL could be done in one step ... which is what I'd prefer to do given the application I'm looking at it for .... ****************************** On Mon, Jan 2, 2012 at 2:40 PM, Esteban Aliverti <esteban.aliverti(a)gmail.com > Just out of curiosity, did you try adding a binary package instead of a > DRL? > > Best Regards, > > XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX > > Esteban Aliverti > - Developer @ http://www.plugtree.com > - Blog @ http://ilesteban.wordpress.com > > > > On Mon, Jan 2, 2012 at 2:30 PM, Matthew Versaggi <profversaggi(a)gmail.com&gt;wrote: > >> Here's what I got working and when it failed .... it was just the basic >> 'hello world' kind of stuff: >> >> Worked: >> final KnowledgeBuilder kbuilder = >> KnowledgeBuilderFactory.newKnowledgeBuilder(); >> >> Died: >> // this will parse and compile in one step >> kbuilder.add(ResourceFactory.newClassPathResource("HelloWorld.drl", >> DroolsTest.class), ResourceType.DRL); >> >> LogCat: >> 01-02 13:13:17.684: E/AndroidRuntime(537): java.lang.RuntimeException: >> Unable to start activity >> ComponentInfo{com.versaggi.android.droolsandroidtest/com.versaggi.android.droolsandroidtest.DroolsAndroidTestActivity}: >> java.lang.NullPointerException >> >> That's about as far as I could get. >> >> I think this DOC will probably help out .... >> >> https://github.com/cgreenhalgh/droolstest/blob/8ea1a06d9b1b38a40b1add7f98.... >> .. >> >> I had to re-compile the CompositeClassLoader.java file and then reinsert >> it into the drools-api-5.1.0.M2.jar file to get past this issue: >> >> "A small modification is needed in org.drools.util.CompositeClassLoader. >> Apparently, the Android JVM doesn't allow null as the parent classloader >> so I changed that to super(CompositeClassLoader.class.getClassLoader())in the constructor, and that works fine." >> >> However I think that his other modifications aren't completely integrated >> into the drools-compiler-5.1.0.M2.jar or the drools-core jar files he >> provides just yet. >> >> ********************************* >> >> >> >> On Mon, Jan 2, 2012 at 2:11 PM, Wolfgang Laun <wolfgang.laun(a)gmail.com&gt;wrote: >> >>> >>> >>> On 2 January 2012 13:58, profversaggi <profversaggi(a)gmail.com>>> >>>> I've run across the work of a very bright man in the UK who did it as >>>> part of >>>> an integrated systems he was working on ... and I got his code working >>>> up to >>>> the point where it died loading the DRL rules. >>>> >>> >>> His rules? Your rules? Did you try the simple "Hello World" rule? >>> >>> rule HW >>> when >>> then >>> System.out.println( "Hello world" ); >>> end >>> >>> -W >>> >>> >>>> I think this might shed some light on things but it's fast getting into >>>> true >>>> software engineering and I'm just not there yet .... >>>> >>>> Here is his codebase: >>>> https://github.com/cgreenhalgh/droolstest >>>> >>>> >>>> -- >>>> View this message in context: >>>> http://drools.46999.n3.nabble.com/Drools-on-android-tp723184p3626516.html >>>> Sent from the Drools: Developer (committer) mailing list mailing list >>>> archive at Nabble.com. >>>> _______________________________________________ >>>> rules-dev mailing list >>>> rules-dev(a)lists.jboss.org >>>> https://lists.jboss.org/mailman/listinfo/rules-dev >>>> >>> >>> >>> _______________________________________________ >>> rules-dev mailing list >>> rules-dev(a)lists.jboss.org >>> https://lists.jboss.org/mailman/listinfo/rules-dev >>> >>> >> >> _______________________________________________ >> rules-dev mailing list >> rules-dev(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/rules-dev >> >> > > _______________________________________________ > rules-dev mailing list > rules-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/rules-dev > >

Once it starts getting into the DX of the DRL rules to feed into Drools it's gets quickly past my experience level .... I had hoped for a clean do it all in one place kind of thing but it doesn't look that way..... ********************** On Mon, Jan 2, 2012 at 2:46 PM, Wolfgang Laun <wolfgang.laun(a)gmail.com&gt;wrote:

On 02/01/2012 12:58, profversaggi wrote: There was a change that kris had to do for the CompositeClassLoader, I'm not sure if it was committed or not - where android insists on a parent classloader. Maybe try that in your local branch. Mark