Joern Nettingsmeier schrieb:
> Andreas Hartmann wrote:

[...]

>> IMO we shouldn't change our release procedure. For many users,
>> downloading is much easier than checking out from SVN.
>> We shouldn't underestimate the clear signal of a downloadable tarball:
>> "Here's something we tested, not some random development version".
> 
> well, a version tagged as RC is *not* a random development version.

You and me know that, but does the average user know it?


>> I can imagine that many people won't give Lenya a try unless this
>> statement is made.
> 
> you certainly have a point, and for many open-source projects, a
> "one-click install" package makes a lot of sense: it's easy for users,
> allows for quick testing, and gives confidence in the maturity of the
> product.
> 
> but with a complex servlet engine like lenya, i think things are different:
> 
> * getting it to run at all requires some java knowledge (classpaths,
> version problems etc.)
> * lenya is a server application, so we can assume our users (who will be
> *admins* of a server deployment) have slightly more expertise and less
> fear of technical issues than your usual end-users.
> * getting binary java packages to run is sometimes harder than building
> them yourself - or at least problems with your local environment are
> often easier to spot during building.

We won't release any binaries (IIUC), so the build process will still
be necessary with the tarballs.


> hence, i think binary packages are not worth the hassle.

Yes, I agree.

> therefore, if users get in touch with the source anyway, i think it's a
> small step to using an svn client once a year.

Hmm, I'm hesitant. I guess there are good reasons why the ASF projects
provide tarballs.

What do the others think?

-- Andreas


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

Reply via email to