[developers] [lkb] Compiling PET with (R)MRS support

Stephan Oepen oe at ifi.uio.no
Sun Apr 18 19:14:07 CEST 2010

hi katya,

> cat test.fsc | cheap -tok=fsc -limit=1000 -default-les=all -cm \
>   -mrs=xml -verbose=10 -results=1 grammar.grm

which grammar is this?  have you confirmed that MRS extraction works in
the LKB?  if so, does PET load the right set of MRS globals?  could you
try your PET binary on the ERG, e.g. something like

  cheap -yy -packing -verbose=4 -mrs -cm \
    -default-les=all english.grm < pet/sample.yy

> Does someone have suggestions about how to fix the problem?

it is hard to say whether this is a grammar-related problem (e.g. some
missing setting, or something), or related to you using an ECL version
that is rumoured to be problematic (in the PET `README').  in general,
it might be worth trying pre-compiled binaries, which are known to work
with several of the larger grammars.  the so-called LOGON tree includes
PET binaries (and quite a number of grammars and other resources); see:


once you find out more about the original problem, please let us know!

                                                          best  -  oe

+++ Universitetet i Oslo (IFI); Boks 1080 Blindern; 0316 Oslo; (+47) 2284 0125
+++    --- oe at ifi.uio.no; stephan at oepen.net; http://www.emmtee.net/oe/ ---

More information about the developers mailing list