[developers] tmr inclusion in logon distributions
glenn at thai-language.com
Thu Apr 11 10:00:48 CEST 2013
Sorry, this may be a false alarm: apparently I have a few different LOGON installations on my linux machine, and the one in question
was not updated to /trunk.
Nevertheless, it does still appear that recovering the file set used to build a .grm file would be nice.
From: Glenn Slayden [mailto:glenn at thai-language.com]
Sent: Wednesday, April 10, 2013 8:50 PM
To: developers at delph-in.net
Subject: tmr inclusion in logon distributions
While tracking an inconsistency between parsers, I discovered, by dumping out the 'english.grm' in the distribution, that it does
not contain all of the token mapping rules implied by the list of 'tmr' files specified in 'english.tdl'. Further, I was not able to
find out where information about which '.set' file was flopped--for each respective distribution set (i.e. Redwoods)--is stored.
The [incr tsdb()] 'run' file does contain the number of AVM instances (and lexical entries) in the grammar image that was used, but
it isn't convenient to work it out this way. And I could discern no method whereby PET activates (or disables) particular token
mapping rules from the loaded grammar.
Insights would be appreciated. I'm on ERG trunk. Many thanks,
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the developers