On 6 Dec 06, at 8:16 PM 6 Dec 06, Brett Porter wrote:

I generally agree with what John has said in the thread so far. I think this is a good thing: a nice out of the box installation

Define nice? If it's an installation that is different then something standard that users typically get then it is a not a good thing. I can just see the threads now:

"I put my global stuff in the /etc/maven/settings.xml and blah blah blah"
"In /etc/maven/settings, mine is in $MAVEN_HOME/conf/settings.xml"
"I don't have a $MAVEN_HOME/conf"
"What version are you using?"
"I'm just using what came with Redhat/BSD/Solaris"

So instead of having our standard platform neutral single directory structure, we have each distro doing whatever they want. If it was a package that dropped stuff into the standard way we are doing things great. If not then I'm am so strongly -1 to any such work I can't be emphatic enough. It will fracture out user base completely. I frankly don't care about distros, I care about our users and their experience and their ability to communicate with other users regardless of what platform they are on. Having different setups because Redhat/Solaris/ BSD feel like doing things different is not a good thing.

For the enterprise I see an agent that can be installed on any OS and pulls down our software from a Maven repository and lays down a structure and pattern that is consistent across any machine for the sake of providing a common infrastructure so that our users can communicate effectively with one another regardless of what platform they are using.

, but you can still roll your own if needed. I think it's important that we break this down into the requirements like these listed below and work through them, using the prototype as a starting point. I agree we can't apply it 'as is', which I think was well known, but some of the concepts are features we'd like to add to Maven anyway so I don't see any problem working through it.

I think we have far more pressing concerns then worrying about. This is definitely not a priority for the vast majority of our users. Releasing 2.0.5, 2.1, and plugins are definitely of more interest I would imagine.


The key requirement from me would be that Maven continues to operate in the same way (warts and all). So, while we may be considering unifying versions and other such things to install and execute Maven, Maven builds should, by default, work the same.

It cannot possibly be the same if it's installed differently on different platforms. It's just another form of vendor lock-in. If Maven, Continuum, Archiva and the stack of our applications had to be navigated differently and used slightly differently then something we can certainly provide in a platform neutral way then I think that's one of greatest disservices we could do to our users. Everyone has lived without RPMs and I don't think we've ever had a single request for one.

If it is something where a package was made and it looked the same as we do now that's great. If it created a barrier where someone who was using Maven on Windows and then moves to Redhat and installs Maven as they are accustomed to and some default system version kicks in because it's in the path first and produces completely unexpected output then that's a bad thing.

I think this is a key issue we are going to need to work through, but it also seems like one that isn't the primary concern from the requirements below. These are phase 1: build and install Maven on Fedora. Phase 2 is use Maven to build and install other things on Fedora, which is where the other options and the contention comes in, right?

So, next steps... should we start fleshing this out on a wiki somewhere? Are there any non-conententious parts of the patches that can be proposed in JIRA now?


Anyone can knock themselves out. I care about users on Linux system but I frankly don't care about LHS for Java as I don't believe it provides a good experience for our users and never could. Given that the requests for RPM, Debian package, or Solaris package are zero I think it's safe to say this is not a great concern to our user base.

Jason.

Thanks,
Brett

On 07/12/2006, at 7:02 AM, Carl Trieloff wrote:

John Casey wrote:

Let's start with: What are the requirements?

Regards,

John


yes that is good - I would not see the patches set as something to commit as is but a prototype to see if it could be done. we now think it can - and would do it in a way that works for maven.

The key parts are:

-  boot strap via build from source
-  OS boot strap-able, with no network, this is more than just mvn -o
-  meet fedora package guidelines

Can break those our one by one, unless Deepak or Rafi fill them first.

Carl.




---------------------------------------------------------------------
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