fixed a bug causing too much gluing of Ita,Spa nominal forms with their complements; should only apply to clitic pronouns. And an even worse bug in Fre, due to the use of course-of-value tables in DiffFre (instead of full explicit tables). After the present participle was added to Romance verbs, these tables lacked one form, but this was not reported by the type checker. The result was garbled linearizations and, in some cases, run-time errors in the PGF interpreter. Something to fix in the concrete syntax type checker of course.

lib/src/romance/ResRomance.gf

@@ -276,18 +276,14 @@ oper
 --- have a "-" with possibly a special verb form with "t":
 --- "comment fera-t-il" vs. "comment fera Pierre"
 
-  infVP : VP -> Agr -> Str = \vp ->
+  infVP : VP -> Agr -> Str = nominalVP VInfin ;
+
+  gerVP : VP -> Agr -> Str = nominalVP (\_ -> VGer) ;
+
+  nominalVP : (Bool -> VF) -> VP -> Agr -> Str = \vf,vp,agr ->
       let
         iform = orB vp.clit3.hasClit (isVRefl vp.s.vtyp) ;
-        vf    = VInfin iform ;    
-      in
-      nominalVP vf iform vp ;
-
-  gerVP : VP -> Agr -> Str = nominalVP VGer True ;
-
-  nominalVP : VF -> Bool -> VP -> Agr -> Str = \vf,iform,vp,agr ->
-      let
-        inf   = vp.s.s ! vf ;
+        inf   = vp.s.s ! vf iform ;
         neg   = vp.neg ! RPos ;             --- Neg not in API
         obj   = vp.s.p ++ vp.comp ! agr ++ vp.ext ! RPos ; ---- pol
         refl  = case isVRefl vp.s.vtyp of {