[SOLVED] Scipting assistance to search for servlets in xml file

onthevirg71 · 07-18-2011, 12:33 PM

Hello all,

We're migrating from Tomcat on Linux to Glassfish on Windows and I'm trying to make it easier to find the servlets that we don't currently have implicitly mapped in the Tomcat web.xml files for contexts. What I've done is copy all our contexts to a temp directory and written a script so that I can loop through each context, find the files that are not mapped inside a package then check if the java file is a servlet or not. From there I store each java file that matches, strip off the extension and then search in the context web.xml for a match then output to the screen.

The problem that I'm running into is that I know there are class files that should be returned and yet they are not being displayed. I'm not a Linux guy, so I'm just looking for a little help to see if I'm missing something obvious. I've attached the code for the script below. Thanks.

Code:

# for all contexts
for context in `ls "/var/tomcat/temp/script/" | grep -v servletmap.sh`
do
	echo $context
	
	# for all java files without a package
	for file in `ls "/var/tomcat/temp/script/$context/src/java/" | grep .java`
	do
		#check to see if it is a servlet.
		if egrep -q "(protected|public) void do(Get|Post)" /var/tomcat/temp/script/$context/src/java/$file
		then
			echo
		else
			class=`echo $file | sed s/.java//`
			
			#check to see if it is declared in web.xml
			if grep -q $class /var/tomcat/temp/script/$context/web/WEB-INF/web.xml
			then
				echo "    $class"
			fi
		fi
	done
done

Tinkster · 07-18-2011, 01:39 PM

Hi, welcome to LQ!

Now if you could show us sample data to go with your snippet...

Cheers,
Tink

onthevirg71 · 07-18-2011, 05:23 PM

Thanks Tinkster.

I changed up the script a little to have some additional output (removed protected portion of the egrep). Also changed the servlet names to protect the innocent. Apologies for all the white space.

Code:

contextname

















    Servlet1

    Servlet2




    Servlet3
    Servlet4
    Servlet5

In these cases I know they are all correct, because servlets 1-5 are protected methods, not public, even though they are in the xml. The problem I'm having is that there are servlets that qualify under egrep portion, yet aren't declared in the xml and I don't get those as results.

grail · 07-18-2011, 08:26 PM

What we need to see is the input .java file that has the protected and public methods that are not being captured by the egrep?

Also, I think you could simplify your code a little:

Code:

for FILE in /var/tomcat/temp/script/*/src/java/*.java
do
	if ! egrep -q "(protected|public) void do(Get|Post)" $FILE
	then
		CLASS=${FILE##*/}
		CLASS=${CLASS%.java}
		XML=${FILE%src*}web/WEB-INF/web.xml
		
		if grep -q "$CLASS" $XML
		then
			echo "    $CLASS"
		fi
	fi
done

I didn't see any reason to put in the superfluous echo which has created your original output. If you have a valid
reason then simply add an else to the above and you should get the same desired affect.

onthevirg71 · 07-19-2011, 10:18 AM

Thanks for the tip on the script grail.

Here's the sample servlet code snippet for a servlet not being captured by the script that is not contained in a web.xml file:

Code:

public class AP_CostCenter extends HttpServlet {

    protected FileUtility fu = new FileUtility();

    @Override
    public void doGet(HttpServletRequest req, HttpServletResponse res)
            throws IOException {
        res.setContentType("application/vnd.ms-excel");
        PrintWriter out = res.getWriter();

grail · 07-19-2011, 03:20 PM

Well based on what I see so far I am not sure I understand the logic?

The script says that if the following regex is not in the file:

Code:

(protected|public) void do(Get|Post)

As the snippet you have shown does have this entry:

Code:

public void doGet(HttpServletRequest req, HttpServletResponse res)

Matching portion in red means this file will be skipped.

onthevirg71 · 07-19-2011, 04:07 PM

Thanks grail,

That's where I was going wrong. Moving the exclusion to the grep statement searching the web.xml files gave me the results that I was hoping to get. I appreciate your help on this.