[Biojava-l] Re: biocorba idl

[Matthew Pocock]

Dear all,

I have just checked in a server & client under biojava-live/demos that
works!!!! Ref counting is not properly implemented yet. But - the corba
objects get spat out as biojava objects, so biojava code can run over remove
objects with no change to the API. Realy cool stuff.

I guess next week we move into a QC phaze, and iron out any last glitches in
the API.

If you want to play with it here are some simple instructions:

put java into your path - at the sanger centre log into one of the ics
machines & type:
> setenv PATH /usr.opt./java122/bin:${PATH}

check out biojava from cvs & cd to biojava-live
Copy xml.jar into biojava-live and build the project - may take some time (we
need to split it into modules)
> sh build/buildall.sh

Then build the client/server
> cd demos
> javac -classpath ../xml.jar:../biojava.jar:. *.java

To run the server type:
> java -classpath ../xml.jar:../biojava.jar:. CorbaDBServer ChrI.prom.fasta
DNA
This will spit out an IOR for the database object. Copy this as you will need
to paste it onto the command line for the client.

To run the client, you will need to bg the server, or move to a different
terminal/machine & type
> java -classpath ../xml.jar:../biojava.jar:. CorbaDBClient
the_very_long_org_string

This will then connect to the server, and retrieve each sequence in the
database and print out its name. Then it will exit. Hack the sript - get it to
print out the sequence, or only the sequences that you chose, or whatever. You
will notice that the names are screwed up, but they should be correct if you
do an explicit fetch-by-name on the sequence. (Yes I need to fix this).

Anyway, it is there to break.

Matthew