Hi

I agree with the infrastructure people, I assume there are plenty of people suggesting to install plenty of useful, fancy tools. The main goal of my mail was to see if other people also were missing something like OpenGrok. And if so, we will have to work out how and where to set it up, and maintain it. And we seem to be making some progress on that.

On the techincal side, OpenGrok does not strictly need "file://" access.
But I think performance wise, it would be best.
I have tried to run OpenGrok on the JMeter project which I had checked out using "svn checkout http://";. But then I was not able to search the Subversion history, and when I were viewing the subversion history through the web browser, then the OpenGrok web application running on my server was actually talking http to the subversion repository to get the history. So for best performance, and to limit the load on the subversion server, I think it would be best to access the subversion repository using "file://"

I think that using "svnsync" to create a read-only mirror of subversion on the machine where subversion was running, would be one possible solution
to achieve "file://" acceses on the server where OpenGrok is running.
I am not sure if it is allowed to do a "svnsync" of the Subversion repository. I tried it from my server, but got an error message from the subversion server. I more or less expected that, as I am not sure if the infra team want to add extra load to the subversion servers by allowing other people to mirror it. But perhaps an Apache machine would be allowed to mirror the subversion repository.


Regards
Alf


Henning Schmiedehausen wrote:
If it needs file:// access to svn, you can not install it on an Apache
Solaris zone. It would need to run on the svn server itself and given
the flamew^Wdiscussion on board and the bar that Justin set for the Java
interest group, I am pretty sure that it is out of the question to get
this in the short run.

If you really are interested and want to help out, please look at
http://cwiki.apache.org/INFRA/ and consider joining infrastructure,
infrastructure-private (or the proposed and yet to be created
infrastructure-dev) mailing lists.
The infrastructure people are quite reluctant to set up new services,
mainly because they fear (rightly so, given the track records of past
volunteers) that someone requests or sets up a service and the
disappears, leaving the maintenance to the few infrastructure people
that have been around continuously for a long time.

Personally, I believe that it is important that infra knows about
'users' (and I that use case, I consider the committers and members of
the ASF also "users of the infrastructure") that would like to have new
services that we do not provide yet. Not every wish might be fulfilled,
though.

        Best regard
                Henning


On Sat, 2007-09-01 at 13:29 -0700, Nathan Bubna wrote:
Hmm.  I'll put "learn how to set up OpenGrok on Velocity's Solaris
zone" on my things-i-would-like-to-find-time-to-do-someday list.  Of
course, if someone more familiar with OpenGrok and/or installing
services on a Solaris zone wanted to help me set it up for Velocity,
then we'd have some example for other interested projects to work off
of. ;)

On 9/1/07, Will Glass-Husain <[EMAIL PROTECTED]> wrote:
(off topic for Jakarta, I know)

Velocity has it's own Solaris zone.  We could set something up on an
experimental basis for Velocity, and if someone in Jakarta got excited about
it they could request a zone/set theirs up based on the Velocity config.

WILL

On 9/1/07, Nathan Bubna <[EMAIL PROTECTED]> wrote:
I would LOVE to have something like this for Apache projects.   Looks
much more useful than ViewVC.  I particularly like that the source
code view would be able to plug right into PMD reports to make those
more useful too.

On 9/1/07, Alf Høgemark <[EMAIL PROTECTED]> wrote:
Hi

I have a number of times missed an an easy to use web interface for
searching through all Jakarta source code and subversion change logs,
and to also being
able to see line number and subversion change log history for a
particular file.

The OpenGrok tool ( http://www.opensolaris.org/os/project/opengrok/ )
seems to me to be very useful in that respect.
So I would like to suggest that OpenGrok is set up to search and index
the Subversion repository at http://svn.apache.org/repos/asf/
OpenGrok seems to be a lot more useful than what is currently available
using a web browser to point to http://svn.apache.org/repos/asf/

I have set up a prototype test for running OpenGrok 0.5 on the JMeter
source, to see how it looks and works.
You can test the prototype at
http://www.kanonbra.com/opensource/jakarta/jmeter/searchable_src/
Here are some links you can test to see various views :

http://www.kanonbra.com/opensource/jakarta/jmeter/searchable_src/search?q=&defs=&refs=&path=&hist=HttpSampler2
http://www.kanonbra.com/opensource/jakarta/jmeter/searchable_src/search?q=&defs=&refs=&path=HttpSampler2&hist=
http://www.kanonbra.com/opensource/jakarta/jmeter/searchable_src/history/src/protocol/http/org/apache/jmeter/protocol/http/sampler/HTTPSampler2.java
http://www.kanonbra.com/opensource/jakarta/jmeter/searchable_src/xref/src/protocol/http/org/apache/jmeter/protocol/http/sampler/HTTPSampler2.java
Note that the changelist numbers on my prototype are incorrect, the
reason is that I have used "svk" to mirror the JMeter part of the
subversion repository at http://svn.apache.org/repos/asf/

I think it would be most useful if OpenGrok index was available for all
the Jakarta projects.
For example if I am doing some testing of JMeter, get an exception with
a stack trace, and I want to look at a specific file, with line number
subversion changelog history, from the commons httpclient.
You can have a look at http://src.opensolaris.org/source/ to see how
OpenGrok looks when it has indexed many projects.


Do other people miss a searchable index of all the source code of
Jakarta, and an easy way to browse source code and subversion history,
both of their own and other's projects ?

Any suggestions on how to proceed to get an OpenGrok index of the
Jakarta subversion repository available from the jakarta website ?


Some thoughts on how it could be set up :
OpenGrok needs a servlet container to run, and ideally local file access
to the subversion repository, to increase speed, and minimize load on
subversion server.
svnsync could be used to synchronize the repository to the machine where
OpenGrok is running.
I think it would be enough to update the index once or twice per day.

Comments appreciated

Regards
Alf Hoegemark
JMeter committer



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


--
Forio Business Simulations

Will Glass-Husain
[EMAIL PROTECTED]
www.forio.com

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to