post-verbal subjects incorrectly analysed as "obj"

[leoalenc]

Trazendo para cá, @arademaker, o problema detectado neste comentário:

https://github.com/LR-POR/PorGram/issues/72#issuecomment-1039003148

~/hpsg/por$ grep -EnA 5 "post-verbal" random-sample-*-entries.tdl`

random-sample-205-entries.tdl:293:; annotation error: post-verbal subject incorrectly annotated as obj
random-sample-205-entries.tdl-294-faltar_2 := trans-verb-lex & ; ERROR:nonrefl-int-verb-lex
random-sample-205-entries.tdl-295-  [ STEM < "faltar" >,
random-sample-205-entries.tdl-296-    SYNSEM.LKEYS.KEYREL.PRED "_faltar_v_2_rel" ]
random-sample-205-entries.tdl-297-  """
random-sample-205-entries.tdl-298-  Embora ainda faltem três dias@obj->nsubj@ de prova, a verdade é que o seu momento de forma é excelente, aliás como o comprovam os resultados alcançados na presente temporada.
--
random-sample-205-entries.tdl:868:; annotation error: post-verbal subject incorrectly annotated as obj
random-sample-205-entries.tdl-869-restar_1 := trans-verb-lex & ; ERROR:nonrefl-int-verb-lex
random-sample-205-entries.tdl-870-  [ STEM < "restar" >,
random-sample-205-entries.tdl-871-    SYNSEM.LKEYS.KEYREL.PRED "_restar_v_1_rel" ]
random-sample-205-entries.tdl-872-  """
random-sample-205-entries.tdl-873-  «Chegámos a acordo sobre a maior parte das questões, só restam algumas divergências@obj->nsubj@ menores», explicaram Yilmaz e Erbakan na conferência de imprensa conjunta que deram ontem.

Até agora, são dois casos, em 683 entradas analisadas manualmente, do total de 2059 geradas automaticamente, conforme as seguintes issues:

https://github.com/LR-POR/PorGram/issues/66 https://github.com/LR-POR/PorGram/issues/67 https://github.com/LR-POR/PorGram/issues/72 https://github.com/LR-POR/PorGram/issues/74

[arademaker]

Commit 03b29517c solves this issue.