ADDCOHORT hangs

[flammie]

this was discussed in irc before christmas by @lynnda-hill:

$ cd github/giellalt/lang-smj && make
$ echo Dan maŋŋela la Sámedigge giellastivrra ja fylkka bálkáv vaddám. | hfst-tokenise -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/tokeniser-gramcheck-gt-desc.pmhfst'  | divvun-blanktag '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/analyser-gt-whitespace.hfst'  | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/../../src/cg3/valency.cg3'  | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/../../tools/tokenisers/mwe-dis.cg3'  | cg-mwesplit  | divvun-cgspell -n 10 -b 15.000000 -w 5000.000000 -u 0.400000 -l '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/acceptor.default.hfst' -m '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/errmodel.default.hfst'  | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/valency-postspell.cg3'  | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/./grc-disambiguator.cg3'  | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/./spellchecker.cg3'  | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/after-speller-disambiguator.cg3' 
"<Dan>"
	"dat" Pron <smj> <smj> Dem Sg Gen <W:0.0> SELECT:2630:GenPo SUBSTITUTE:4194 SUBSTITUTE:4188
;	"Da" N Prop Sem/Org Ess <W:0.0> SELECT:2630:GenPo
;	"Da" N Prop Sem/Org Sg Ine <W:0.0> SELECT:2630:GenPo
;	"Dan" N Prop Sem/Mal Attr <W:0.0> REMOVE:1816
;	"Dan" N Prop Sem/Mal Sg Nom <W:0.0> REMOVE:1816
;	"dat" Pron Dem Sg Err/Infl Ela Attr <W:0.0> SELECT:2630:GenPo
;	"dat" Pron Dem Sg Gen Attr <W:0.0> SELECT:2630:GenPo REMOVE:2168
;	"dat" Pron Dem Sg Ill Attr <W:0.0> SELECT:2630:GenPo
;	"dat" Pron Dem Sg Ine Attr <W:0.0> SELECT:2630:GenPo
: 
"<maŋŋela>"
	"maŋŋela" Po <smj> <smj> <W:0.0> SELECT:2842 SUBSTITUTE:4192 SUBSTITUTE:4186
;	"maŋŋel" N Sem/Dummytag Sg Gen <W:0.0> SELECT:2842
;	"maŋŋela" Adv <W:0.0> SELECT:2842
;	"maŋŋela" Pr <W:0.0> SELECT:2842
: 
"<la>"
	"liehket" V <smj> <smj> IV Ind Prs Sg2 <W:0.0> MAP:3350 SUBSTITUTE:4189 @+FAUXV SUBSTITUTE:4183
	"liehket" V <smj> <smj> IV Ind Prs Sg3 <W:0.0> MAP:3350 SUBSTITUTE:4189 @+FAUXV SUBSTITUTE:4183
;	"liehket" V IV Ind Prs ConNeg <W:0.0> REMOVE:3098
: 
"<Sámedigge>"
	"Sámedigge" N <smj> <smj> Prop Sem/Org Sg Nom <W:0.0> SELECT:3258 SUBSTITUTE:4187 SUBSTITUTE:4181
;	"sámedigge" N Sem/Build_Org Sg Nom <W:0.0> SELECT:3258
;	"digge" N Sem/Org Sg Nom <W:0.0>
;		"sábme" N Sem/Hum_Lang Cmp/SgGen Cmp <W:0.0> REMOVE:2110:longest-match
;	"digge" N Sem/Org Sg Nom <W:0.0>
;		"sáme" A Cmp/Attr Cmp <W:0.0> REMOVE:2110:longest-match
;	"digge" N Sem/Org Sg Nom <W:0.0>
;		"sáme" N Sem/Dummytag Cmp/Unass Cmp <W:0.0> REMOVE:2110:longest-match
: 
"<giellastivrra>"
	"stivrra" N <smj> <smj> Sem/Org Sg Nom <W:0.0> SUBSTITUTE:4187 SUBSTITUTE:4181
		"giella" N Sem/Lang_Tool-catch Cmp/SgNom Cmp <W:0.0>
: 
"<ja>"
	"ja" CC <smj> <smj> <W:0.0> MAP:2012 SELECT:2068 SUBSTITUTE:4195 @CNP SUBSTITUTE:4189
;	"ja" CC <W:0.0> @CVP MAP:2012 SELECT:2068
: 
"<fylkka>"
	"fylkka" v1 N <smj> <smj> Sem/Org Sg Nom <W:0.0> SUBSTITUTE:4187 SUBSTITUTE:4181
: 
"<bálkáv>"
	"bálkká" N <smj> <smj> Sem/Money Sg Acc <W:0.0> SUBSTITUTE:4187 SUBSTITUTE:4181
: 
"<vaddám>"
	"vaddet" V <smj> <smj> <RE-Ill-Ani><TH-Acc-*Ani> <RE-Ill-Ani> TV PrfPrc <W:0.0> SUBSTITUTE:1054 SUBSTITUTE:1061 MAP:3381 SELECT:4072 SUBSTITUTE:4189 @-FMAINV SUBSTITUTE:4183
;	"vadde" N Sem/Hum Gram/NomAg Sg Acc PxSg1 <W:0.0> REMOVE:3980
;	"vadde" N Sem/Hum Gram/NomAg Sg Gen PxSg1 <W:0.0> REMOVE:3538
;	"vadde" N Sem/Hum Gram/NomAg Sg Nom PxSg1 <W:0.0> SELECT:4072
"<.>"
	"." CLB <W:0.0>
:\n

$ echo Dan maŋŋela la Sámedigge giellastivrra ja fylkka bálkáv vaddám. | hfst-tokenise -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/tokeniser-gramcheck-gt-desc.pmhfst'  | divvun-blanktag '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/analyser-gt-whitespace.hfst'  | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/../../src/cg3/valency.cg3'  | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/../../tools/tokenisers/mwe-dis.cg3'  | cg-mwesplit  | divvun-cgspell -n 10 -b 15.000000 -w 5000.000000 -u 0.400000 -l '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/acceptor.default.hfst' -m '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/errmodel.default.hfst'  | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/valency-postspell.cg3'  | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/./grc-disambiguator.cg3'  | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/./spellchecker.cg3'  | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/after-speller-disambiguator.cg3' | vislcg3 --trace -g '/home/flammie/github/giellalt/lang-smj/tools/grammarcheckers/./grammarchecker.cg3' 

[hangs up]

The offending rule is commented out in git:

diff --git a/tools/grammarcheckers/grammarchecker.cg3 b/tools/grammarcheckers/grammarchecker.cg3
index 7c3cc6f01..8c4f51e6d 100644
--- a/tools/grammarcheckers/grammarchecker.cg3
+++ b/tools/grammarcheckers/grammarchecker.cg3
@@ -2158,7 +2158,7 @@ COPY:syn-nom_nom-gen_nom (Gen &SUGGEST) EXCEPT (Nom &syn-nom_nom-gen_nom) TARGET
 
 ADD:xsyn-nom-nom-comma (&syn-nom-nom-comma) TARGET (Gram/3syll Actio Nom) OR (N Sg Nom) IF (1 (N Sg Nom) OR (Actio Nom))(2 (CC) LINK 1 (Nom)) (NEGATE 0 (Gen)) (NEGATE 1 (V Ind) OR Adv OR Po) (NEGATE 0 EVEN-VERBS + Der/NomAct) (NEGATE 1 (Gram/TAbbr)) (NEGATE 0 (A Sg Nom)) (NEGATE 1 ("jurra") LINK *1 (V Ind) BARRIER NOT-CC - Adv) (NEGATE -1 (N Sg Nom) LINK 1 (Actio Nom)) (NEGATE 1 (Prop Sg Nom))  ;
    
-#ADDCOHORT  ("<,>" "," CLB &ADDED &syn-nom-nom-comma) AFTER &syn-nom-nom-comma;
+ADDCOHORT  ("<,>" "," CLB &ADDED &syn-nom-nom-comma) AFTER &syn-nom-nom-comma;
 ADDRELATION (RIGHT) (&syn-nom-nom-comma) TO (1 (&ADDED)) ;
    
     ## Dan maŋŋela la Sámedigge giellastivrra ja fylkka bálkáv vaddám.

the pipeline used is tools/grammarcheckers/modes/trace-gramchecker-dev.mode which can be regenerated with make dev in the grammarchers directory.

[TinoDidriksen]

Without having looked at any code, isn't that simply an endless loop? It's adding a new &syn-nom-nom-comma after every just-added &syn-nom-nom-comma.

Require that it can only hit existing readings: ADDCOHORT ("<,>" "," CLB &ADDED &syn-nom-nom-comma) AFTER &syn-nom-nom-comma - &ADDED; ?