org.apache.ldap.common.message.SearchRequestImpl needs to implement toString()

• Key: DIRSERVER-583
• Type: Improvement
• Status: Closed
• Resolution: Fixed
• Priority: Major
• Assignee: Emmanuel Lecharny
• Reporter: Norval Hope
• Component/s: ldap
• Affects Version/s: 1.0-RC1

Description

The other message classes implement friendly toString()s, but the search request doesn't meaning the default INFO level is of very little use for debugging. I set the priority as major because this severely impacts debugging system behaviour, and certainly will be noticed by new users installing RC1 (and may prevent them for providing the raw material to help in raising JIRA issues).

Some example output (notice the RECEIVED message is a generic java.lang.Object.toString()):

[21/02/2006 11:06:42] [/155.35.171.101:2351] SENT: BindResponse
        Ldap Result
            Result code : (ResultCodeEnum[SUCCESS=0]) success
            Matched DN : 'null'
            Error message : 'null'

[21/02/2006 11:06:42] [/155.35.171.101:2351] RECEIVED: org.apache.ldap.common.message.SearchRequestImpl@132ae7
[21/02/2006 11:06:42] [/155.35.171.101:2351] WRITE: Search Result Entry
        Object Name : 'cn=ETA1,dc=example,dc=com'
        Attributes
Attributes

Comments

[Norval Hope]

Here is a method implementation which I think should do the trick :


    public String toString()
    {
        StringBuffer sb = new StringBuffer();

        sb.append( " SearchRequest\n" );
        sb.append( super.toString() );
        sb.append( " baseDn : '" ).append(baseDn).append("'\n");
        sb.append( " filter : '" );
        filter.printToBuffer(sb);
        sb.append("'\n");
        sb.append( " scope : " ).append(scope.toString()).append('\n');
        sb.append( " typesOnly : " ).append(typesOnly).append('\n');
        sb.append( " sizeLimit : " ).append(sizeLimit).append('\n');
        sb.append( " timeLimit : " ).append(timeLimit).append('\n');
        sb.append( " derefAliases : " ).append(derefAliases).append('\n');
        sb.append( " attributes : " );

        boolean first = true;

        for (Iterator it = attributes.iterator(); it.hasNext();)
        {
            final String s = (String)it.next();

            if (!first)
                sb.append(", ");
            sb.append('\'').append(s).append('\'');
        }
        sb.append('\n');

        return sb.toString();
    }

[Alex Karasulu]

One word: patch. :). Give a committer a patch he can apply and well things get done faster. Take a look at svn diff.

[Alex Karasulu]

Oh and a test case with the patch is also a good idea. We don't want to introduce things like NPE into the code just to print out strings.

[Alex Karasulu]

Not a bug but a improvement I think.

[Norval Hope]

method provided as a patch and, as suggested, added some defensive coding checks to avoid all possible NPEs

[Norval Hope]

improved coding slightly after investigating precise coding of StringBuffer.append(Object)

[Emmanuel Lecharny]

Sorry guys, it's a valid concern, but it's already implemented :)

just set the following logger in log4j.properties file :
org.apache.directory.shared.ldap.codec.TwixDecoder=DEBUG

and you will get :

[07:55:42] DEBUG [org.apache.directory.shared.ldap.codec.TwixDecoder] - Decoding the PDU :
[07:55:42] DEBUG [org.apache.directory.shared.ldap.codec.TwixDecoder] - 0x30 0x3A 0x02 0x01 0x03 0x63 0x35 0x04 0x15 0x6F 0x75 0x3D 0x53 0x65 0x61 0x72 0x63 0x68 0x2C 0x20 0x6F 0x3D 0x49 0x4D 0x43 0x2C 0x63 0x3D 0x55 0x53 0x0A 0x01 0x00 0x0A 0x01 0x03 0x02 0x01 0x00 0x02 0x01 0x00 0x01 0x01 0x00 0x87 0x0B 0x6F 0x62 0x6A 0x65 0x63 0x74 0x43 0x6C 0x61 0x73 0x73 0x30 0x00 0x00 0x00 0x00 0x00
[07:55:42] DEBUG [org.apache.directory.shared.ldap.codec.TwixDecoder] - Decoded LdapMessage : LdapMessage
    message Id : 3
    Search Request
        Base Object : 'ou=Search,o=IMC,c=US'
        Scope : base object
        Deref Aliases : deref Always
        Size Limit : no limit
        Time Limit : no limit
        Types Only : false
        Filter : '(objectClass=*)'

[Emmanuel Lecharny]

Will be fixed in RC1

[Emmanuel Lecharny]

Ok, applied in trunks !