[jboss-jira] [JBoss JIRA] (DROOLS-4662) Resources added through KieHelper get interpreted using the default character set

[Mario Fusco (Jira)]

    [ https://issues.redhat.com/browse/DROOLS-4662?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13937346#comment-13937346 ] 

Mario Fusco edited comment on DROOLS-4662 at 12/18/19 10:53 AM:
----------------------------------------------------------------

I tried on master the following test case setting the file.enconding as you suggested but it works for me

{code}
    @Test
    public void testI18NKieHelper() throws UnsupportedEncodingException {
        String testString = "\u2200x. \u2203y. x < y";

        String rules =
                "rule \"foo\"\n" +
                        "  when\n" +
                        "    not String()\n" +
                        "  then\n" +
                        "    insert(\"" + testString + "\");\n" +
                        "  end\n";

        KieServices kieServices = KieServices.Factory.get();
        KieHelper kieHelper = new KieHelper();

        Resource rulesResource = kieServices.getResources().newByteArrayResource(rules.getBytes("UTF-8"), "UTF-8");
        rulesResource.setTargetPath("src/main/resources/rules.drl");
        kieHelper.addResource(rulesResource);

        KieBase kieBase = kieHelper.build();

        KieSession kieSession = kieBase.newKieSession();
        kieSession.fireAllRules();

        String out = (String) kieSession.getObjects().iterator().next();
        System.out.println(out);

        assertEquals(out, testString);
    }
{code}

and prints

∀x. ∃y. x < y

I also tried the opposite with this slightly modified test case

{code}
    @Test
    public void testI18NKieHelper() throws UnsupportedEncodingException {
        String testString = new String("\u2200x. \u2203y. x < y".getBytes( "latin1" ));

        String rules =
                "rule \"foo\"\n" +
                        "  when\n" +
                        "    not String()\n" +
                        "  then\n" +
                        "    insert(\"" + testString + "\");\n" +
                        "  end\n";

        KieServices kieServices = KieServices.Factory.get();
        KieHelper kieHelper = new KieHelper();

        Resource rulesResource = kieServices.getResources().newByteArrayResource(rules.getBytes("latin1"), "latin1");
        rulesResource.setTargetPath("src/main/resources/rules.drl");
        kieHelper.addResource(rulesResource);

        KieBase kieBase = kieHelper.build();

        KieSession kieSession = kieBase.newKieSession();
        kieSession.fireAllRules();

        String out = (String) kieSession.getObjects().iterator().next();
        System.out.println(out);

        assertEquals(out, testString);
    }
{code}

e.g. setting both the encoding of string and rule resource to latin1 and this also works adn  prints as expected

?x. ?y. x < y

As a final note consider that in all cases the KieHelper class is only intended for internal use and it is not part of our API.



was (Author: mfusco):
I tried on master the following test case setting the file.enconding as you suggested but it works for me

{code}
    @Test
    public void testI18NKieHelper() throws UnsupportedEncodingException {
        String testString = "\u2200x. \u2203y. x < y";

        String rules =
                "rule \"foo\"\n" +
                        "  when\n" +
                        "    not String()\n" +
                        "  then\n" +
                        "    insert(\"" + testString + "\");\n" +
                        "  end\n";

        KieServices kieServices = KieServices.Factory.get();
        KieHelper kieHelper = new KieHelper();

        Resource rulesResource = kieServices.getResources().newByteArrayResource(rules.getBytes("UTF-8"), "UTF-8");
        rulesResource.setTargetPath("src/main/resources/rules.drl");
        kieHelper.addResource(rulesResource);

        KieBase kieBase = kieHelper.build();

        KieSession kieSession = kieBase.newKieSession();
        kieSession.fireAllRules();

        String out = (String) kieSession.getObjects().iterator().next();
        System.out.println(out);

        assertEquals(out, testString);
    }
{code}

and prints

∀x. ∃y. x < y

I also tried the opposite with this slightly modified test case

{code}
    @Test
    public void testI18NKieHelper() throws UnsupportedEncodingException {
        String testString = new String("\u2200x. \u2203y. x < y".getBytes( "latin1" ));

        String rules =
                "rule \"foo\"\n" +
                        "  when\n" +
                        "    not String()\n" +
                        "  then\n" +
                        "    insert(\"" + testString + "\");\n" +
                        "  end\n";

        KieServices kieServices = KieServices.Factory.get();
        KieHelper kieHelper = new KieHelper();

        Resource rulesResource = kieServices.getResources().newByteArrayResource(rules.getBytes("latin1"), "latin1");
        rulesResource.setTargetPath("src/main/resources/rules.drl");
        kieHelper.addResource(rulesResource);

        KieBase kieBase = kieHelper.build();

        KieSession kieSession = kieBase.newKieSession();
        kieSession.fireAllRules();

        String out = (String) kieSession.getObjects().iterator().next();
        System.out.println(out);

        assertEquals(out, testString);
    }
{code]

e.g. setting both the encoding of string and rule resource to latin1 and this also works adn  prints as expected

?x. ?y. x < y

As a final note consider that in all cases the KieHelper class is only intended for internal use and it is not part of our API.


> Resources added through KieHelper get interpreted using the default character set
> ---------------------------------------------------------------------------------
>
>                 Key: DROOLS-4662
>                 URL: https://issues.redhat.com/browse/DROOLS-4662
>             Project: Drools
>          Issue Type: Bug
>    Affects Versions: 7.10.0.Final, 7.20.0.Final, 7.28.0.Final
>         Environment: Windows. Can be reproduced on other OSes using "-Dfile.encoding=latin1".
>            Reporter: Max Zerzouri
>            Assignee: Mario Fusco
>            Priority: Major
>         Attachments: Test.java
>
>
> When a resource is added using {{KieHelper.addResource}}, any specified encoding is ignored when it is added to the {{MemoryFileSystem}} by {{KieFileSystemImpl.write(Resource)}} (effectively assuming the {{InputStream}} is already encoded in UTF-8).
> Additionally, when the resource is read again by {{AbstractKieModule.getResource}}, it constructs a {{Resource}} object with no encoding specified, whereas it should probably specify an encoding of UTF-8, since it seems that sources are considered to be UTF-8 already at least in the case of the {{MemoryKieModule}}.
> I've attached a fairly minimal test case where the rules bytes are passed in UTF-8 with the encoding explicitly specified. This currently produces an error when the default file encoding is not UTF-8 (eg, by passing {{-Dfile.encoding=latin1}} as a JVM argument, or by using the default arguments on Windows).



--
This message was sent by Atlassian Jira
(v7.13.8#713008)