Estonian verb subject gets dropped when Sno is imported about gf-midwife

[kristiank]

Relevant code and pointers:

The Estonian linearization puts in the "default" subject.

GF-midwife/MidwifeEst.gf

Line 48 in e7a0dbc

FeelsFetalMovements level = mkS (mkCl (mkNP (mkN "patsient")) (mkVP (mkVP (mkV2 (mkV "tundma" "tunda" "tunneb")) (mkNP (mkN "loote" (mkN "liigutusi")))) level)) ;

Whereas the Snomed linearization only puts the level-record

GF-midwife/MidwifeSno.gf

Line 41 in e7a0dbc

FeelsFetalMovements level = level ;

which is simply

GF-midwife/MidwifeSno.gf

Line 42 in e7a0dbc

Strong = {s = "276371006 |Strong fetal movements (finding)|"} ;

[inariksit]

All still works, but I can't understand why the parser fails. Before restarting the gf environment and re-importing the grammars for writing this issue, then the subject (patsient) simply vanished and I could only parse "tunneb hästi looteliigutusi".

Not sure what you mean. If we look at the output (annotations mine):

> p "patsient tunneb hästi looteliigutusi"
FeelsFetalMovements Strong   -- This is the correct parse from MidwifeEst

The parser failed at token 1: "patsient" -- This is the output from trying to parse the sentence in MidwifeSno

If you have all languages imported, and only want to parse in one, you can add the -lang=Est flag in the command, like this:

Midwife> p -lang=Est "patsient tunneb hästi looteliigutusi"
FeelsFetalMovements Strong

[kristiank]

If you have all languages imported, and only want to parse in one, you can add the -lang= flag in the command

Okay, this explains it. I guess it would have been more self-evident for me, if the fault message in the output would somehow be marked by language.

Thank you very much!