Andrew Borley wrote:
[snip]
A couple of questions:
What do you see going into --enable-most-extensions?
My own view would be CPP, Axis2 and one of the scripting languages to
give a
good view of what SCACPP can do. Alternatively, it could check for
certain
environment variables on the system (e.g. AXIS2C_HOME, PYTHON_HOME)
and only
compile those extensions that have the required variable defined.
Should we follow the Java side of things and release binaries containing
different groups of extensions?
As Simon says, the runtime only loads and uses dlls it finds and needs to
use, so releasing a single binary with all the libraries in it should
be OK.
Currently we have no libraries that clash (like Java does with the
Axis and
Celtix binding.ws implementations), but when we do I guess we'll
either have
to release different binaries, or have some installation process where
the
user chooses which particular extension is to be used.
Cheers
Andy
Andy,
I was not sure what people would say about the --enable-most-extensions
:) Here's what I propose (not a definite list, but an approach to help
define the list):
- For our next Milestone release, the stable/complete enough extensions,
we vote and decide what goes in
- In the future, our users should help us decide what they want in the
"most" group
I'd prefer to have a fixed list without any magic based on your
environment variables, just so that you know what you get / you get what
you asked for when you invoked configure. However I could be easily
convinced otherwise if several of you think that it's nicer to have the
build check the build machine's envionment.
I agree with Simon's approach to release just one binary. Even with
everything in it our distribution is small so let's take this
opportunity to keep things simple :) We just need a good set of
--enable- / --disable configure options, this will allow people with
specific packaging requirements to easily build the distribution they need.
--
Jean-Sebastien
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
