[developers] ^M after features in MRS in result file

Stephan Oepen oe at ifi.uio.no
Fri Nov 11 23:04:10 CET 2011

many thanks for the diagnostic help, woodley,

i must agree, the VPM specification is incomplete,
seeing that it fails to determine whether or not ^M
(aka CR) is to be considered whitespace or not (on
platforms where it does not form part of a newline,
when followed by LF).  i accept the desire to have
the TDL and VPM readers be comparable at least
in this respect, to better support grammarians who
work cross-platform (DOS and Un*x conventions).

i have checked in a patch to the LOGON sources,
which will be part of the LOGON run-time binaries
next time they are re-generated, and which i will
propagate into the LKB trunk at some point in the
not too distant future.

all best, oe

On Thu, Nov 10, 2011 at 19:02, Woodley Packard <sweaglesw at sweaglesw.org> wrote:
> Hi Emily,
> The problem is that the "semi.vpm" file has DOS line endings. I would reckon
> you could call that an LKB bug, seeing as LKB seems happy to accept the DOS
> line endings in the TDL files.  If you reformat the semi.vpm file to have
> regular UNIX line endings, the problem goes away (at least as tested on my
> end).
> Cheers,
> Woodley
> On 11/10/2011 09:33 AM, Emily M. Bender wrote:
>> Dear all,
>> We are experiencing strange behavior with the fine system
>> (using the latest and greatest logon from http://svn.emmtee.net/trunk)
>> using one particular grammar (attached).
>> In particular, we're seeing ^M after each variable property
>> name in the MRS (see the attached file).  This works okay
>> for me (Ubuntu 10.04), but for Varya (presumably a different
>> linux version) it's causing the fine system to not be able to
>> read the results files, as the ^M characters are apparently
>> being understood as newlines.  (An example can be found
>> in the attached grammar directory under tsdb/home/bad-newlines.)
>> I can confirm that this is not happening for at least one other
>> grammar using the same logon on the same machine, so
>> I'm guessing it's somehow a property of the attached grammar.
>> A quick look through the grammar and skeleton files
>> doesn't show anything suspicious.  Has anyone seen similar
>> behavior?  Any suggestions on where to look?
>> Thanks,
>> Emily and Varya
>> --
>> Emily M. Bender
>> Associate Professor
>> Department of Linguistics
>> Check out CLMS on facebook! http://www.facebook.com/uwclma

+++ 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