[developers] extension to [incr tsdb()] client API

Stephan Oepen oe at ifi.uio.no
Sun Feb 22 18:08:45 CET 2009

dear all,

after many years of stability, i just revised the [incr tsdb()] client
API, i.e. the protocol used in communication (via PVM) with processing
clients like PET or (TR)ALE.  this revision is not dramatic (one extra
argument to process_item()), and it may even be possible for old client
binaries to work with new [incr tsdb()] code (but not vice versa).  but
everyone compiling their own PET binaries needs to make sure to update
both PET and [incr tsdb()] at the same time.

the new API has the following signature:

  extern int 
  capi_register(int (*)(const char *, int, const char *,
                        int, int, const char *),
                int (*)(int, const char *, int, int, 
                        int, int, int, const char *),
                int (*)(const char *),
                int (*)(int, const char *));

where the final `const char *' argument to process_item() is new (and 
provides a way of passing `custom' information to the client).

i have updated the [incr tsdb()] code both in the LinGO CVS repository
(where a new `test' build is now available) and the LOGON tree; i have
also applied the corresponding changes to `tsdb++.cpp' in PET, in both
the `trunk' and `cm' branch.  when compiling new client binaries, make
sure to link against an up-to-date `libitsdb.a' or `libitsdb.so'.  for
environments where `libitsdb.so' is copied `outside' the [incr tsdb()]
tree (e.g. the Heart of Gold, i believe), please also make sure to get
the latest version of the shared library.

i hope these changes will not cause any issues.  feel free to email me
in case you experience surprising [incr tsdb()] client behavior!

                                                  best wishes  -  oe

+++ Universitetet i Oslo (IFI); Boks 1080 Blindern; 0316 Oslo; (+47) 2284 0125
+++     CSLI Stanford; Ventura Hall; Stanford, CA 94305; (+1 650) 723 0515
+++       --- oe at ifi.uio.no; oe at csli.stanford.edu; stephan at oepen.net ---

More information about the developers mailing list