 |
|
|
|
|
|
JPython Source Access Via CVS
In the spirit of Python's
open source
code, we are making JPython's source code available in a
similar manner. This will aid both the JPython developer community
and the JPython maintainers by allowing those with the particular
interest to view the very latest version of the source.
Important: In order to get access to the JPython source
code, you will need to read and agree to JPython's
licensing terms. You may want to stop
now and read these terms.
This page briefly describes the JPython public CVS tree, what is
available, and how to do checkouts of the code. You can also
get notified of new checkins via email. If you
plan on
generating patches
to JPython, we highly encourage you to grab
the CVS snapshot and generate patches against it. It is possible
that your problem has already been fixed, and if not, it makes our job
of integrating your changes much easier if your patches apply to the
current source snapshot.
You should probably also read the section of the documentation on
Compiling JPython from source.
What is available?
Most of the JPython distribution is available, however some essential
pieces that come with the installer distribution are missing. Most
notable are the OROMatcher regular expression library, and the
installer generated files. We are prohibited from distributing the
OROMatcher classes separately, although you can download them for use
with JPython from ORO, Inc.'s own
Website. The JPython installer also generates some files when it
runs, and these are not kept under CVS control. See the instructions below on running JPython from the CVS
tree.
All of our changes to the internal source tree are immediately
reflected in the read-only CVS mirror. This means that if you want to
stay on the bleeding edge of JPython development, you can easily do so
by checking out the source tree, and updating as changes are checked
in. If you are more inclined to wait until stable versions
are officially released, then wait for announcements on the
appropriate mailing lists, newsgroups, and Web sites.
|
What is CVS?
`CVS' stands for the Concurrent Version System, a free version control
system that I use to manage changes in the development tree of the
JPython source code. CVS clients are freely available for Unix,
Windows, and MacOS so any developer who wants to have access to the
JPython source tree should be able to do so. For more information on
obtaining a CVS client for your system, see the
Cyclic Web pages. Note that you
will want to use CVS 1.9 or better.
The first thing you need to do is log into the remote CVS
server. You do this by executing the following command (assuming
Unix1):
CVSROOT=":pserver:jpythoncvs@cvs.python.org:/projects/cvsroot"
export CVSROOT
% cvs login
This will prompt you for a password; read
these instructions on getting the password.
Once you've logged in, you can do your initial checkout:
% cvs -z6 co -d jpython jpython/dist
Once you've done the intial checkout, you can stay current by
cd'ing into the jpython subdirectory and typing:
% cvs -z6 -q up -P -d
where -q says to be quiet, -z6 says to use
level 6 compression (a good tradeoff between improved network
performance and CPU usage), -P says to prune empty
directories (good for ignoring obsolete directories), and
-d says to check out newly added directories.
Note that after the initial checkout, you do not need to specify
the CVS repository to use or to set the CVSROOT environment
variable. CVS remembers this automatically.
When Barry or any other member of the development team, makes changes
to the internal source tree, two things happen immediately. First,
the change is copied out to the public read-only mirror on
cvs.python.org. This means that you always have access to exactly the
same tree that we have in front of us. Second, the log message for the change
is sent to a mailing list that you can
subscribe to. Note that it is a read-only list; you cannot post
messages, so use normal channels for communicating back to Barry and
the development team.
Also note that this mailing list is not archived, since all the
information is available already through the CVS log commands.
Other distribution mechanisms
Currently anonymous CVS is the only access method available, however
many people (especially those outside the USA) have indicated that
remote CVS is pretty inefficient. As time allows, we'll investigate
making other options for accessing the source tree available. Watch
here for further instructions.
Once you've done the checkout as described above, you need to do a
little extra work in order to get JPython running. Essentially you
need to manually do what the installer ordinarily generates a script
to do. Here's a sample Bourne shell script
which I use to start up
JPython on my Solaris box, using the Sun production 1.2 JVM. You will
probably need to change the projroot variable and
oro variables, and possibly others. Sorry I can't help much
for running JPython from CVS under Windows. If you have a sample
script you'd like to donate for these pages,
let me know.
Note that I have set JAVA_COMPILER=NONE in the script. In
my experience, Sun's JIT-enabled JVMs have too many bugs to make them
work reliably for JPython, so I just turn 'em off unconditionally.
1The environment variable command shown here is valid for
the Bourne or Bash shells (sh or bash). For csh derivatives use:
setenv CVSROOT ":pserver:jpythoncvs@cvs.python.org:/projects/cvsroot"
