Jonathan Lalou's Blog

Archive for the ‘General’ Category

I wrote the following notice on Vincent NAEGELEN‘s profile on LinkedIn:

Vincent joined the Prime Brokerage team as an intern. He quickly integrated himself to the new project “from scratch” I was setting up, and learned various technologies of JEE world, such as GWT, Hibernate, Spring, etc.Vincent kept on working within my team after his internship, on a brand new and innovative application based on Mule ESB, giving satisfaction to all his colleagues. I strongly recommand Vincent for his technical skills and functionnal knowledge.

Case

I have the following block in my Mule config file:

[xml]<servlet:inbound-endpoint path="authenticationService" address="http://localhost:1234">[/xml]

. Even though this block matches XSD constraints, it is illegal from a functionnal point of view.

I get the following stacktrace:

[java]Offending resource: mule-config.xml; nested exception is org.springframework.beans.factory.BeanDefinitionStoreException: Unexpected exception parsing XML document from class path resource [lalou/jonathan/mule-config.xml]; nested exception is org.mule.config.spring.parsers.processors.CheckExclusiveAttributes$CheckExclusiveAttributesException: The attributes of Element servlet:inbound-endpoint{address=http://localhost:1234, name=.authenticationService:inbound.157:inbound-endpoint.158, path=authenticationService} do not match the exclusive groups [address] [ref] [path][/java]

Explanation and Fix

The exception is meaningful: for the tag <servlet:inbound-endpoint>, only one among the three following attributes is needed: path, ref and address.

To fix the issue, keep only the relevant attribute.

Case

I try to deploy a Mule ESB configuration, using this XML:

[xml]<?xml version="1.0" encoding="UTF-8"?>
<mule xmlns="http://www.mulesoft.org/schema/mule/core" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:pattern="http://www.mulesoft.org/schema/mule/pattern"
xsi:schemaLocation="http://www.mulesoft.org/schema/mule/core
http://www.mulesoft.org/schema/mule/core/3.1/mule.xsd
http://www.mulesoft.org/schema/mule/pattern
http://www.mulesoft.org/schema/mule/pattern/3.1/mule-pattern.xsd
">
<pattern:simple-service name="authenticationService"
address="http://localhost:1234/authenticationService"
component-class="lalou.jonathan.esb.components.AuthenticationComponent"
type="direct" />
</mule>[/xml]

I get the following error:

[java]Ignored XML validation warning
org.xml.sax.SAXParseException: schema_reference.4: Failed to read schema document ‘http://www.mulesoft.org/schema/mule/pattern/3.1/mule-pattern.xsd'[/java]

Extended Stacktrace

[java]2011-11-22 16:10:25,375 WARN  xml.XmlBeanDefinitionReader         – Ignored XML validation warning
org.xml.sax.SAXParseException: schema_reference.4: Failed to read schema document ‘http://www.mulesoft.org/schema/mule/pattern/3.1/mule-pattern.xsd’, because 1) could not find the document; 2) the document could not be read; 3) the root element of the document is not <xsd:schema>.
at com.sun.org.apache.xerces.internal.util.ErrorHandlerWrapper.createSAXParseException(ErrorHandlerWrapper.java:195)
at com.sun.org.apache.xerces.internal.util.ErrorHandlerWrapper.warning(ErrorHandlerWrapper.java:96)
at com.sun.org.apache.xerces.internal.impl.XMLErrorReporter.reportError(XMLErrorReporter.java:380)[/java]

Notice that Mule ESB files are similar to classic Spring files. Of course, I first checked the pointed XSD was actually reachable.
Anyway, this error should be raised when your application, for any reason -firewall, proxies, network interruption-, cannot access the remote site where the XSD is hosted.

Fix

• Copy the XSD to a local folder
• Create a file spring.schemas
• Make it available in the classpath in META-INF.
• Add the following line
• [java]http\://www.mulesoft.org/schema/mule/pattern/3.1/mule-pattern.xsd=WEB-INF/classes/mule-pattern.xsd[/java]

  The pattern is: missing resource (beware of escaping colon) = path in classpath of the local XSD

• Rebuild, pack and run!

A Short String

I have a table of which one field is a CLOB. Let’s say I have to insert one record with a short text. The following command is allowed:

[sql]INSERT INTO jonathan_table VALUES (1, ‘hello world!’);[/sql]

A Bigger Text

Error ORA-01704

Now, my text is larger, let’s say 5000 characters. When I launch the same query, I get the following error:

[sql]ORA-01704: string literal too long[/sql]

Indeed, Oracle/SQL*Plus have a limit on CLOB inserts: 4000 bytes.

Workaround

To pass through the limit on canonical SQL, you’ll have to use a PL/SQL procedure. The following command will be successful for any text larger than 4000 bytes, but shorter than 32767:

[sql]DECLARE
bigtext1   VARCHAR2 (32767);
BEGIN
bigtext1 := lpad(‘X’, 32000, ‘X’)
INSERT INTO jonathan_table VALUES (1, bigtext1);
END;[/sql]

An Even Bigger Text

Errors ORA-06550 and PLS-00103

You guess it: beyond this limit of 32 KB, an error occurs. So the following script:

[sql]DECLARE
bigtext1 VARCHAR2 (42000);
BEGIN
bigtext1 := lpad(‘X’, 42000, ‘X’)
INSERT INTO jonathan_table
VALUES (1, bigtext1);
END;
[/sql]

raises such an error:

[sql]Error at line 1
ORA-06550: line 5, column 4:
PLS-00103: Encountered the symbol "INSERT" when expecting one of the following:

. ( * % & = – + ; < / > at in is mod remainder not rem
<an exponent (**)> <> or != or ~= >= <= <> and or like like2
like4 likec between || multiset member submultiset
The symbol ";" was substituted for "INSERT" to continue.[/sql]

Fix this issue

I searched a lot to find an easy fix to go beyond the limit of 32KB. My point was that with Java for instance there is no limit of 32KB. In the same way, with TOAD I was able to update the record with many mega bytes of text, via the clipboard. After further search, I learnt that the 32KB barrier was a SQL*Plus limitation on actual strings, but the patterns insert into ... select ... from were not affected.
Here is the idea:

• create a temporary table
• split the text into blocks shorter than 32KB
• insert the blocks into the temporary table
• perform a first insert with a null CLOB
• update the record using a select on the temporary table (yet you can insert the actual value since previous step)

Here is an example:

[sql]DROP TABLE tt_jonathan_table;
CREATE GLOBAL TEMPORARY TABLE tt_jonathan_table
(
ID NUMBER(10),
pdlsuffix CLOB
) ON COMMIT PRESERVE ROWS;

TRUNCATE TABLE tt_jonathan_table;

DECLARE
bigtext1 VARCHAR2 (32767);
bigtext2 VARCHAR2 (32767);
BEGIN
bigtext1 := lpad(‘X’, 32000, ‘X’)
bigtext2 := lpad(‘Y’, 32000, ‘Y’)
INSERT INTO tt_jonathan_table
VALUES (1, bigtext1);

INSERT INTO tt_jonathan_table
VALUES (2, bigtext2);

INSERT INTO jonathan_table
(id, myClobField)
VALUES (jonathan_seq.NEXTVAL, NULL);

UPDATE jonathan_table
SET myClobField = (SELECT CONCAT (rls1.myClobField, rls2.myClobField)
FROM tt_jonathan_table rls1, tt_jonathan_table rls2
WHERE rls1.ID = 1 AND rls2.ID = 2)
WHERE myClobField is null;
END;
/

TRUNCATE TABLE tt_jonathan_table;
[/sql]

Case

Sometimes, you need test the Log4J’s loggers are called with the right parameters. How to perform these tests from with JUnit?

Let’s take an example: how to test these simple class and method?

[java]public class ClassWithLogger {
private static final Logger LOGGER = Logger.getLogger(ClassWithLogger.class);

public void printMessage(Integer foo){
LOGGER.warn(&quot;this is the message#&quot; + foo);
}
}[/java]

Example

Define an almost empty Log4J appender, such as:

[java]public class TestAimedAppender extends ArrayList&lt;String&gt; implements Appender {
private final Class clazz;

public TestAimedAppender(Class clazz) {
super();
this.clazz = clazz;
}

@Override
public void addFilter(Filter newFilter) {
}

@Override
public Filter getFilter() {
return null;
}

@Override
public void clearFilters() {
}

public void close() {
}

@Override
public void doAppend(LoggingEvent event) {
add(event.getRenderedMessage());
}

@Override
public String getName() {
return &quot;TestAppender for &quot; + clazz.getSimpleName();
}

@Override
public void setErrorHandler(ErrorHandler errorHandler) {
}

@Override
public ErrorHandler getErrorHandler() {
return null;
}

@Override
public void setLayout(Layout layout) {
}

@Override
public Layout getLayout() {
return null;
}

@Override
public void setName(String name) {
}

public boolean requiresLayout() {
return false;
}
}[/java]

Then create a TestCase with two fields:

[java]public class ClassWithLoggerUnitTest {
private ClassWithLogger classWithLogger;
private TestAimedAppender appender;
…
}
[/java]

In the setup, remove all appenders, create an instance of our appender, and then add it to the logger related to the class which we want to test:

[java] @Before
public void setUp() throws Exception {
final Logger classWithLoggerLogger = Logger.getLogger(ClassWithLogger.class);
classWithLoggerLogger.removeAllAppenders();
appender = new TestAimedAppender(ClassWithLogger.class);
classWithLoggerLogger.addAppender(appender);
appender.clear();

classWithLogger = new ClassWithLogger();
}[/java]

Then write the following test. The code is documented:

[java] @Test
public void testPrintMessage() throws Exception {
final String expectedMessage = &quot;this is the message#18&quot;;

// empty the appender
appender.clear();
// check it is actually empty before any call to the tested class
assertTrue(appender.isEmpty());

// call to the tested class
classWithLogger.printMessage(18);

// check the appender is no more empty
assertFalse(appender.isEmpty());
assertEquals(1, appender.size());
// check the content of the appender
assertEquals(expectedMessage, appender.get(0));

}[/java]

Conclusion

This basic example shows how to perform tests on logger, without overriding the original code or using mocks. Of course, you can improve this basic example, for instance in discriminating owing to the log level (INFO, WARN, ERROR, etc.), use generics, and even any other fantasy ;-).

I wrote the following notice on Karim CHORFI‘s profile on LinkedIn:

Working with Karim is a pleasure: always available, smiling, with a sense of humor. This is for the human profile ; concerning his professionnal abilities, let’s mention his double skills -finance and IT- make him able to handle various problems, as complex technically as functionnaly.

I recommand Karim warmly.

I wrote the following notice on David CHAU‘s profile on LinkedIn:

David gathers two main qualities: first, he is an actual developer, I mean: a guy with a passion for technologies, Java, frameworks, able to deal of Android on his iPhone, to argue pros and cons on the JARs his depends on, or advise on complex technical points. Besides, he is a nice, sociable, and open-minded boy, working with whom is a pleasure.

Case and Topology

RMI services are deployed on UAT, exposed via a F5, at the following address: t3://my-f5-frontal.my.domain.extension:7090
The actual servers are my-first-node.my.domain.extension and my-second-node.my.domain.extension.

The client application is deployed in a remote location, on a QA server.
The ports are open between QA and UAT, and we can ping and use telnet with no issue on QA.

Anyway, when I launch the client application from QA, I get the following error:

[java]2011-10-31 06:41:03,277 INFO support.DefaultListableBeanFactory – Destroying singletons in org.springframework.beans.factory.support.DefaultListableBeanFactory@79e304: defining beans [jonathanServiceClient]; root of factory hierarchy
Exception in thread "main" org.springframework.beans.factory.BeanCreationException: Error creating bean with name ‘smartServiceClient’ defined in class path resource [com/lalou/jonathan/rmi-client-spring.xml]: Invocation of init method failed; nested exception is org.springframework.remoting.RemoteLookupFailureException: JNDI lookup for RMI service [rmiServices] failed; nested exception is javax.naming.CommunicationException [Root exception is java.net.ConnectException: t3://my-f5-frontal.my.domain.extension:7090: Bootstrap to my-f5-frontal.my.domain.extension/111.222.012.123:7090 failed. It is likely that the remote side declared peer gone on this JVM][/java]

Explanation and Fix

Owing to my understanding, here is the point: when the client tries to connect to F5, it presents itself with its name, and the F5 returns the name of the actual server. If both client and server are not on the same domain (“domain” as network domain, no link with Weblogic domain), then the DNS resolution may fail.

To fix this issue, you have to follow one or both of the following points: everything depends on your local topology.

WebLogic: “Listen Address”

Modify the “Listen Address” in WebLogic administration console, from home:  Servers > MyFirstNode/MySecondNode > Configuration > General > Listen Address > update it

By “update” the “Listen Address”, I mean providing the complete name of the machines, including the domain extension.
eg: my-first-node.my.domain.extension and my-second-node.my.domain.extension, rather than my-first-node and my-second-node (or, even worse, localhost).
You can also provide an IP, cf. WebLogic documentation on Oracle’s website.

Of course, you can decide to set it directly in WebLogic’s config.xml.

Caution! This option may also be set via the command line running WebLogic, using the flag -Dweblogic.ListenAddress=... Therefore, take care to be consistent between the content of console/config.xml and the command line option.

Hosts

On client side, check the content of hosts file. Usually, you can found it at /etc/hosts (or C:\WINDOWS\system32\drivers\etc\hosts on Windows XP).
Assuming your machine is myClientMachine with an IP 123.123.123.123 and a domain extension remote.domain, then your hosts file should look like:

[java]127.0.0.1 localhost
123.123.123.123 myClientMachine[/java]

Update it to:

[java]127.0.0.1 localhost
123.123.123.123 myClientMachine myClientMachine.remote.domain[/java]

Lamine SIDIBE honored me in writing the following notice on my profile on LinkedIn:

“I was working at BNP Paribas when Jonatthan joined the Prime Brokerage team. His skills allowed him to work in a very demanding environment. It was very pleasant to work with him for his human and profesional qualities. Jonathan is able to find efficient technical solutions for complex functional business requirements.” October 25, 2011

1st Lamine SIDIBE, J2EE Consultant, BNP PARIBAS
worked directly with Jonathan at BNPParibas

Jacky LALOU honored me in writing the following notice on my profile on LinkedIn:

“Jonathan is a detail-oriented manager who watches the balance sheet like a hawk without ever losing sight of the strategic objective.” October 5, 2011

1st Jacky Lalou, Actuarial analyst, BNP Paribas Assurance
was with another company when working with Jonathan at BNPParibas