Difference between revisions of "OpenGrok"

From CDOT Wiki
Jump to: navigation, search
(File History)
(OpenGrok)
 
(6 intermediate revisions by 3 users not shown)
Line 1: Line 1:
== Introduction ==
+
==Introduction==
OpenGrok is the source browser in use by the OpenSolaris communityIt uses Java Servlets running on Java5
+
OpenGrok is a source code searcher similar in idea to MXR.  It is very advanced and written in Java.  It uses Apache Lucene to build a search database for a source code tree.  This tool is perfect for the new mercurial repository as it has built in hg repo parsing.  There is only version number searching capability I think: you can't use, for example, FIREFOX_3_1b1_RELEASE in the History box.  If there is interest in Mozilla adopting this program, I'd be willing to write support for itI'd also like to customize the look and feel for the Mozilla web sites.  I have a list of things to do in my script.
  
 +
[[OpenGrok/script|The Script]]
  
==Idea's==
+
==Setup==
Opengrok has 2 major componenetsOpengrok.jar is a system independant Java Application that will index source code, which can then be viewed with its own internal client, or sent onto a Java Servlet for internet usage. Source.war is the servlet that actually serves up the information.  <insert information when i feel like doing so>
+
Setting this up is fairly easy if you know your way around a unix shellYou will need the following tools:
== Features and Benefits ==
+
* JDK 1.5+
==== File History ====
+
* Servlet container - JBoss, Glassfish, Tomcat, Jetty, etc.
OpenGrok maintains code version history.  I am currently unsure of whether it looks at the CVS versioning information or whether it compares files between indexes. There is also options for many diffs and comparing different, non-concurrent versions.[http://cvs.opensolaris.org/source/history/on/usr/src/uts/sun/io/eri.c Example]
+
* Exuberant Ctags
 +
* Mercurial
 +
* Basic unix programs: sed, zip, unzip, nice, bash
  
After a little looking at the OpenSolaris OpenGrok, it seems that the version histories are more Wiki style than CVS style, so they are arbitrary
+
I am assuming you have set up all the required applications installed and configured.
  
 +
You are going to need to download the opengrok binary release tarball from [http://opensolaris.org/os/project/opengrok/ OpenGrok].  Untar it (tar zxf opengrok-0.7.tar.gz) and make note of the directory you put it in.  Next you will configure the following variables in my script:
 +
<pre>
 +
#Mercurial Repository to use
 +
REPO=http://hg.mozilla.org/mozilla-central <-- Change to your HG Repo
  
[http://blogs.sun.com/chandan/entry/releasing_opengrok_0_3_subversion Blog about OpenGrok] that explains file history better.  This means that it is not really the best choice to use the CVS history information for indexing mozilla code because the history tracker will just use the newest version of the file, not the branch that it is in, thus, It would be a wise course of action to have either one opengrok webpp for each branch or have one opengrok with all versions in it.
+
#Base directory for contents
 +
BASEDIR=/home/jhford/mozilla <-- This is the base directory for the OpenGrok data files and the HG repo
  
==== Extendable ====
+
#This is the name of the data directory for OpenGrok
If there are files which cannot be analysed by the included analysers, OpenGrok allows the user to develop their own analysers, I am unsure of whether or not there is a JavaScript analyser allready, maybe even an XPCOM/XUL analyser could be written, but I don't know the first thing about XPCOM/XUL so this maybe allready implemented in another language. 
+
DATADIR=data-dir <-- This is the folder *name* for opengrok's data
  
==== In File Linking ====
+
#This is the name of the source directory for OpenGrok
The indexer hyperlinks all symbols to the file and line number of their declaration.  If the symbol is in multiple files it shows a list of all files which have it.
+
SRCDIR=src-dir <-- This is the folder *name* for opengrok's copy of HG code
  
==== Integrates Well ====
+
#OpenGrok install path - Where is opengrok.jar?
The header and footer are completely customizable. OpenGrok could be integrated with the current MDC potentially. The header and footer are implemented through a variable which holds the copmlete html of the header and footer.
+
OGDIR="/home/jhford/opengrok/opengrok-0.7" <-- this is where the OpenGrok tarball files ended up. It is likely where you ran "tar zxf opengrok-0.7.tar.gz"
  
== Example Implementation ==
+
#OpenGrok source.war from opengrok tarball
[http://cvs.opensolaris.org/source/ OpenSolaris]
+
OGCONTEXT="grokzilla" <-- this variable defines the grokzilla in http://china:8080/grokzilla
 +
OGWAR="${OGCONTEXT}.war" <-- Likely, you will never need to change this
  
== Installation Procedure ==
+
#Your application server's .war deployment directory
==== Notes ====
+
WEBAPPDIR="/home/jhford/tomcat/apache-tomcat-6.0.18/webapps" <-- This is where your app server's war deployment folder is or where you want the resulting war file
I will be listing notes for my own reference, feel free to use them
+
<pre>
  
==== Steps ====
+
Then run the script and notice that you have opengrok! It will likely show up at http://localhost:8080/grokzilla
* Install Ubuntu Linux as a server from the Ubuntu DVD
 
* Install the following packages
 
  - ctags -> exuberant-ctags
 
  - sun-java5-jdk
 
  - tomcat5 tomcat5-admin tomcat5-webapps
 
    - Installs GNU Java which must be removed
 
  -make (using 3.81, may need 3.80)
 
  -CVS
 
* Run "sudo update-alternatives --config java" and "apt-get remove --purge java-gcj-compat"
 
* Change the tomcat5 config (/etc/default/tomcat5) to point to the new default Java runtime (/usr/lib/jvm/java-1.5.0-sun/) and disable the security manager, until that can be figured out
 
* Setup a Mozilla code directory (mkdir /var/mozilla)
 
* Follow guide at the [http://www.opensolaris.org/os/project/opengrok/manual OpenSolaris Project Guide]
 
* Enjoy!
 
 
 
 
 
==== Important Info ====
 
* /var/lib/tomcat5 is the default $CATALINA_HOME
 
* /etc/default/tomcat5 has some configuration files
 
* My SRC_ROOT is /var/mozilla
 
* My DATA_ROOT is /var/tmp/opengrok_data
 
* My run.sh uses a ctags at /usr/bin/ctags not the default
 
* File permissions need to be changed [http://tomcat.apache.org/tomcat-5.0-doc/security-manager-howto.html this may help]
 
* Setting up the Admin webapp [http://www.onjava.com/pub/a/onjava/2003/06/25/tomcat_tips.html here]
 
 
 
 
 
* [http://www.opensolaris.org/jive/thread.jspa?threadID=8934&tstart=30 Usefull topic for history info]
 

Latest revision as of 22:37, 12 November 2008

Introduction

OpenGrok is a source code searcher similar in idea to MXR. It is very advanced and written in Java. It uses Apache Lucene to build a search database for a source code tree. This tool is perfect for the new mercurial repository as it has built in hg repo parsing. There is only version number searching capability I think: you can't use, for example, FIREFOX_3_1b1_RELEASE in the History box. If there is interest in Mozilla adopting this program, I'd be willing to write support for it. I'd also like to customize the look and feel for the Mozilla web sites. I have a list of things to do in my script.

The Script

Setup

Setting this up is fairly easy if you know your way around a unix shell. You will need the following tools:

  • JDK 1.5+
  • Servlet container - JBoss, Glassfish, Tomcat, Jetty, etc.
  • Exuberant Ctags
  • Mercurial
  • Basic unix programs: sed, zip, unzip, nice, bash

I am assuming you have set up all the required applications installed and configured.

You are going to need to download the opengrok binary release tarball from OpenGrok. Untar it (tar zxf opengrok-0.7.tar.gz) and make note of the directory you put it in. Next you will configure the following variables in my script:

#Mercurial Repository to use
REPO=http://hg.mozilla.org/mozilla-central <-- Change to your HG Repo

#Base directory for contents
BASEDIR=/home/jhford/mozilla <-- This is the base directory for the OpenGrok data files and the HG repo

#This is the name of the data directory for OpenGrok
DATADIR=data-dir <-- This is the folder *name* for opengrok's data

#This is the name of the source directory for OpenGrok
SRCDIR=src-dir <-- This is the folder *name* for opengrok's copy of HG code

#OpenGrok install path - Where is opengrok.jar?
OGDIR="/home/jhford/opengrok/opengrok-0.7" <-- this is where the OpenGrok tarball files ended up. It is likely where you ran "tar zxf opengrok-0.7.tar.gz"

#OpenGrok source.war from opengrok tarball
OGCONTEXT="grokzilla" <-- this variable defines the grokzilla in http://china:8080/grokzilla 
OGWAR="${OGCONTEXT}.war" <-- Likely, you will never need to change this

#Your application server's .war deployment directory
WEBAPPDIR="/home/jhford/tomcat/apache-tomcat-6.0.18/webapps" <-- This is where your app server's war deployment folder is or where you want the resulting war file
<pre>

Then run the script and notice that you have opengrok!  It will likely show up at http://localhost:8080/grokzilla