Hi Chris,

This is great news!  I'm very happy to see all of the various enhancements that 
you've worked on.  

The next step is to figure out how to smoothly integrate these changes as well 
as any future changes you make into the distribution.  There are two issues:

(1) Configuration management.  One approach would be to create a SVN branch for 
you to work in.  That way you can do whatever you want, commit your changes, 
and we will have a way to review and merge your changes and offer suggestions 
and help.  An alternative would be for you to work on the main branch.  (Note 
that SVN branching is extremely simple but merging is not.  Do you have prior 
experience with SVN branching/merging?) 

(2) Issue management.  We are striving for all changes to the code base to be 
reflected by at least one Jira issue.  One way we do this is by requesting that 
all SVN commits be accompanied by a log message that refers to the Jira 
issue(s) associated with the changes in the commit.  To implement this, you 
would need to create Jira issues describing each of your fixes prior to their 
associated commits. 

How does this sound to you? If reasonable, then the next step would be for me 
to set up SVN and Jira accounts for you. 

Cheers,
Philip



----- Original Message -----
From: Chris Thomson <[EMAIL PROTECTED]>
Date: Monday, October 23, 2006 2:30 am
Subject: [HACKYSTAT-DEV-L] Various Hackystat fixes/enhancements I'm working on
To: [email protected]

> Hi,
> 
> We are planning to use Hackystat (hopefully from later this week) 
> to 
> collect some research data (general overview of our planned work is 
> to 
> be found at the links in my signature).
> 
> I've been given permission to hand back any fixes that I make to 
> the 
> existing code base as I do them, so this email is firstly to give 
> an 
> overview of what I've already done, and secondly to ask about the 
> best 
> way to pass the changes back?
> 
> A build including my fixes can be found at 
> http://ext.dcs.shef.ac.uk/~u0058/, which we will be using as our 
> live 
> server, but please don't use it to post (significant amounts of) 
> data to 
> it. I may delete non-sheffield accounts occasionally.
> 
> So a brief overview of the fixes so far:
> 
> - [Feature] Server Non Standard installation location
> In Sheffield I needed to install hackystat to a server with no 
> admin 
> access to Tomcat, and on a different path (in particular the 
> servlets 
> are only available in /servlet). I found that to make this happen I 
> needed to make a few changes to the Kernel, Installer, and the 
> build files.
> 
> (also a draft page of documentation at) 
> http://ext.dcs.shef.ac.uk/~u0058/docbook/ch06s05.html
> 
> A (nice) feature of this is that HackyInstaller now defaults to the 
> server from which it was downloaded from, with no further editing...
> 
> I'm still testing this feature as I can't think of an easy way to 
> identify all the hard coded paths - It think I found them all, but 
> I 
> guess there may be others that I have not found!
> 
> - [Feature] Sensor installation in a read-only environment
> These changes allow sensors to be installed into a read-only 
> environment. The changes allow the initialisation of the sensor 
> properties, however no attempt is made to ensure the sensor exists.
> 
> This is a general change so that any sensor could allow 
> configuration 
> even if the directory is read-only. However each sensor installer 
> will 
> need to implement the new canInstall method to indicate if the 
> directory 
> is read only. If it does not you may get a file write error 
> depending on 
> the implementation. So far I've only sorted out the Eclipse sensor.
> 
> - [Bug] Eclipse sensor will not install into alternative plug in  
> directories
> The installer checks to see if .eclipse_product exists, but this is 
> not 
> in alternative plugin directories. I've changed this to check for 
> the 
> plugins directory.
> 
> - [Bug] HackyCore_Kernel -> Pagewriter.java error reporting
> Enhanced the error reporting so that this class is debugable. 
> Corrected 
> errors in the logger code so that it actually works!
> 
> - [Bug] HackyInstaller does not read the context properly from the 
> config file
> 
> When loading settings from the config file the Model does not copy 
> these 
> into the system properties list (it does if you edit them in the 
> common 
> settings however) as a result if you change the context, when you 
> first 
> start up HackyInstaller it fails the initial ping, in 
> Notification.java. 
> I've added this functionality into hackyCore_Installer->Model.java, 
> but 
> is perhaps needs refactoring as is a rough copy.
> 
> - [Bug] Logs only write to the Catalina log directory.
> 
> - [Bug] Settings only read from the Catalina settings directory / 
> or a 
> known war directory.
> 
> Remember I've already fixed these bugs, so there is no need to work 
> on them!
> 
> Thanks,
> 
> Chris
> 
> -- 
> Chris Thomson
> Sheffield Software Engineering Observatory
> University of Sheffield, UK
> 
> http://observatory.group.shef.ac.uk
> http://www.shef.ac.uk/dcs/research/groups/vt/research/observatory.html 
> 

Reply via email to