Hi Martin,
Martin van den Bemt wrote:
As far as I can see, most committers don't have the time anymore to do cactus
work or moved to other
things in life.
That's true. Vincent Massol, Cactus founder and main developer, is
working on other projects. A couple of years ago, Nicolas Chalumeau and
myself joined the project. Unfortunately, Nicolas passed way, and I
slowly faded away, mostly for personal reason (I haven't contributed too
much for any OSS project and the last months).
What I get from other people is that Cactus is still something that deserves
work being done to keep
it alive and get it up-to-date, but to do that we need more volunteers (new
blood) and volunteers
within Jakarta to get those volunteers aboard / start applying patches.
Yes, I agree. I tried to at least keep the contributors alive, but am
lacking behind the work (for instead, a have a starred conversation
regarding a patch from William Ferguson dated of 2006-01-05, i.e., it's
been stalled for 1 year). My last effort was to get a SOC contributor,
but our proposal didn't get a ranking high enough to get the funding.
Fortunately, though, the student, Petar, haven't give up and continue to
work in the project; I've been in talk with him privately in the last
days, but haven't got the time to apply his work yet.
I hope Filipe can give some feedback on possible actions and if he can find
some time to guide this?
It's Felipe, not Filipe :-).
Anyway, I think the most important think to do is to finish the Cargo
integration and release Cactus 1.8 (hopefully using the M2 build). Doing
so will renew the project energies, opening the discussions regarding
the project future (for instance, which Java EE versions to support).
For Filipe : another possibility could that we get a discussion started on
general to see if there
are any Jakarta people willing to help out ?
Yes, that might be a good idea. There were many people interested in the
testing.apache.org effort last year, but once it got the first negative
feedback from the boarding meeting, the enthusiasm faded away (I'm not
blaming everybody, much the opposite - I am the one who should keep the
TLP effort going, but for personal reasons could not make it).
Anyway, even if get the attention of these people, they would most
likely work in the project as volunteers. So, another think that might
help, is to find a company to sponsor the project, i.e., to hire people
to work full-time in Cactus (and other testing projects in general;
that's one direction I would like to take the testing.apache.org if it
got created).
I know that's harder, but maybe it's a question of approaching the right
people in the right companies (for instance, it might be interesting
for Sun to have Cactus providing Java EE 5 testing for Glassfish
integrated into NetBeans). Notice that most of the successfully active
projects (Tomcat, Maven, Derby, etc...) are backed by paid developers.
I, for one, would love to be such a developer, but unfortunately do not
have the 'political influence' to pursue it (i.e., do not know the right
people in the right places :-(.
The JIRA does show much commit during the last year.
That's true. Most - if not all - committed work was the changes
regarding the bundled LGPLed JBoss jar.
J2EE 1.4 is not already supported and the 1.5 spec is out.
The problem in supporting J2EE 1.4 and Java EE 5 is that it would
require some heavy refactoring in the way the build is structured. We
might need to make some sacrifices for that refactoring, like stop
supporting J2EE 1.2, for instance (I don't that would be a real
sacrifice, though). Anyway, porting the build to Maven 2 might help this
effort.
Globally it leaves the impression that not much is happening.
Unfortunately, that's mostly true. I mean, it's just not completely true
due to Petar's work, which has been in 'stealth mode' so far
If so, what is the reason, replace by a more popular solution for similar
tests.
I don't think there is an alternative for Cactus; the main alternative
would be out-of-the-container testing (or no testing at all, as it
happens most of the time :-). The main problem, as Martin already
pointed out, is the lack active developers.
-- Felipe
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
