Troubles for building a correct chess pgn grammar.

[laurent bernabe]

Hello everyone,

I am trying to build my own chess PGN grammar : compilation is fine, but when parsing a PGN file, the flow is broken since the first caracter !!!

This is the error I get :

______________________________________________________________

line 1:0 mismatched input '[Event "?"]' expecting {'[', MOVE_NUMBER}

______________________________________________________________

What's wrong with my grammar ? (I would like, for each game, to have the collected tags in a group called tag, and in the notation group, to collect each single/pair of halfmoves individually. And that can explain some of the grammar choices).

I've joined the pgn file I've used for the test.

Regards

[Terence Parr]

sounds like a UTF-8 or other encoding issue.  

Ter

--
You received this message because you are subscribed to the Google Groups "antlr-discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to antlr-discussi...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
<Scandinave.pgn>

[laurent bernabe]

Hello,

thank you for your answer,

I did not though about this problem.

Meanwhile, I notice that I forgot to give an updated version of the grammar I am working on (now it's done).

Furthermore, as I added modifications, I noticed that errors and result tree was quite differents. That make me think I am still on the learning way of the grammars building :).

This time I get these errors :

____________________________

line 1:7 token recognition error at: '"?'

line 1:9 token recognition error at: '"]'

line 2:6 token recognition error at: '"?'

line 2:8 token recognition error at: '"]'

line 3:6 token recognition error at: '"?'

line 3:8 token recognition error at: '?'

line 3:9 token recognition error at: '?'

line 3:10 token recognition error at: '?'

line 3:12 token recognition error at: '?'

line 3:13 token recognition error at: '?'

line 3:15 token recognition error at: '?'

line 3:16 token recognition error at: '?'

line 3:17 token recognition error at: '"]'

line 4:7 token recognition error at: '"?'

line 4:9 token recognition error at: '"]'

line 5:7 token recognition error at: '"o'

line 5:27 token recognition error at: '"]'

line 6:7 token recognition error at: '"o'

line 6:27 token recognition error at: '"]'

line 7:8 token recognition error at: '"*'

line 7:10 token recognition error at: '"]'

line 8:5 token recognition error at: '"C'

line 8:9 token recognition error at: '"]'

line 2:0 mismatched input '[' expecting TAG_VALUE

line 3:0 mismatched input '[' expecting TAG_VALUE

line 3:11 mismatched input '.' expecting TAG_VALUE

line 5:0 mismatched input '[' expecting TAG_VALUE

line 5:9 mismatched input 'peningTrainerWhite' expecting TAG_VALUE

line 6:9 mismatched input 'peningTrainerBlack' expecting TAG_VALUE

line 8:0 mismatched input '[' expecting TAG_VALUE

line 8:7 mismatched input '01' expecting TAG_VALUE

________________________________________________

Regards

[Terence Parr]

i'm afraid you'll have to read the manual/book and see examples rather than just pasting in rexpr stuff like this: ;)

TAG_VALUE:  '"'[:print:]*'"';

Ter

--

Dictation in use. Please excuse homophones, malapropisms, and nonsense. 

[laurent bernabe]

Thank you very much :)

I used this kind of notation([:print:]) in Regex, but it can't apply here.

My pgn grammar is nearly complete : I've just problems with the NEWLINE lexer : I can't figure why it can't resolve '\n' into the lexer NEWLINE.

Errors are :

_____________________________

line 1:11 mismatched input '\n' expecting NEWLINE

line 2:10 mismatched input '\n' expecting NEWLINE

line 3:19 mismatched input '\n' expecting NEWLINE

line 4:11 mismatched input '\n' expecting NEWLINE

line 5:29 mismatched input '\n' expecting NEWLINE

line 6:29 mismatched input '\n' expecting NEWLINE

line 7:12 mismatched input '\n' expecting NEWLINE

line 8:11 mismatched input '\n\n1.' expecting NEWLINE

_________________________________

Regards

Le samedi 28 décembre 2013 23:55:25 UTC+1, the_antlr_guy a écrit :

i'm afraid you'll have to read the manual/book and see examples rather than just pasting in rexpr stuff like this: ;)

TAG_VALUE:  '"'[:print:]*'"';

Ter

[laurent bernabe]

I've gone one step further :

The tags are well parsed, but I still have problem for discarding newlines.

Problematic token is in red :

________________________________________________________________

[@0,0:0='[',<5>,1:0]

[@1,1:5='Event',<10>,1:1]

[@2,7:9='"?"',<11>,1:7]

[@3,10:10=']',<6>,1:10]

[@4,12:12='[',<5>,2:0]

[@5,13:16='Site',<10>,2:1]

[@6,18:20='"?"',<11>,2:6]

[@7,21:21=']',<6>,2:9]

[@8,23:23='[',<5>,3:0]

[@9,24:27='Date',<10>,3:1]

[@10,29:40='"????.??.??"',<11>,3:6]

[@11,41:41=']',<6>,3:18]

[@12,43:43='[',<5>,4:0]

[@13,44:48='Round',<10>,4:1]

[@14,50:52='"?"',<11>,4:7]

[@15,53:53=']',<6>,4:10]

[@16,55:55='[',<5>,5:0]

[@17,56:60='White',<10>,5:1]

[@18,62:82='"openingTrainerWhite"',<11>,5:7]

[@19,83:83=']',<6>,5:28]

[@20,85:85='[',<5>,6:0]

[@21,86:90='Black',<10>,6:1]

[@22,92:112='"openingTrainerBlack"',<11>,6:7]

[@23,113:113=']',<6>,6:28]

[@24,115:115='[',<5>,7:0]

[@25,116:121='Result',<10>,7:1]

[@26,123:125='"*"',<11>,7:8]

[@27,126:126=']',<6>,7:11]

[@28,128:128='[',<5>,8:0]

[@29,129:131='ECO',<10>,8:1]

[@30,133:137='"C01"',<11>,8:5]

[@31,138:138=']',<6>,8:10]

[@32,139:142='\n\n1.',<13>,8:11]

[@33,144:145='e4',<13>,10:3]

[@34,147:148='d5',<13>,10:6]

[@35,150:151='2.',<13>,10:9]

[@36,153:154='d4',<13>,10:12]

[@37,156:156='(',<3>,10:15]

[@38,158:159='2.',<13>,10:17]

[@39,161:164='exd5',<13>,10:20]

[@40,166:169='Qxd5',<13>,10:25]

[@41,171:171='(',<3>,10:30]

[@42,173:176='2...',<13>,10:32]

[@43,178:180='Nf6',<13>,10:37]

[@44,182:182=')',<4>,10:41]

[@45,184:185='3.',<13>,10:43]

[@46,187:189='Nc3',<13>,10:46]

[@47,191:193='Qa5',<13>,10:50]

[@48,195:195=')',<4>,10:54]

[@49,197:197='(',<3>,10:56]

[@50,199:200='2.',<13>,10:58]

[@51,202:203='f3',<13>,10:61]

[@52,205:208='dxe4',<13>,10:64]

[@53,210:211='3.',<13>,10:69]

[@54,213:216='fxe4',<13>,10:72]

[@55,218:218=')',<4>,10:77]

[@56,220:223='2...',<13>,10:79]

[@57,225:226='e6',<13>,10:84]

[@58,228:229='3.',<13>,10:87]

[@59,231:234='exd5',<13>,10:90]

[@60,236:239='exd5',<13>,10:95]

[@61,242:242='*',<1>,10:101]

[@62,243:242='<EOF>',<-1>,10:102]

line 8:11 extraneous input '\n\n1.' expecting {'[', MOVE_NUMBER}

(file (game (tag [ Event "?" ]) (tag [ Site "?" ]) (tag [ Date "????.??.??" ]) (tag [ Round "?" ]) (tag [ White "openingTrainerWhite" ]) (tag [ Black "openingTrainerBlack" ]) (tag [ Result "*" ]) (tag [ ECO "C01" ]) \n\n1. e4 d5 2. d4 ( 2. exd5 Qxd5 ( 2... Nf6 ) 3. Nc3 Qa5 ) ( 2. f3 dxe4 3. fxe4 ) 2... e6 3. exd5 exd5 * notation))

_________________________________________________________________________________________________________________________________________

The new grammar and the Scandinave.pgn test file are joined.

Regards

[Terence Parr]

On Dec 29, 2013, at 7:49 AM, laurent bernabe wrote:

I've gone one step further :

The tags are well parsed, but I still have problem for discarding newlines.

doesn't look that way to me

T

Regards

--
You received this message because you are subscribed to the Google Groups "antlr-discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to antlr-discussi...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

<Scandinave.pgn><Pgn.g4>

[laurent bernabe]

It is strange, because as when debugging in AntlrWorks2,

it has some troubles with some newlines.

Maybe should I use the antlr4 command line tools, and not rely on AntlrWork2.

[Terence Parr]

AW2 and the command-line might be using diff versions of antlr lib?

T

[laurent bernabe]

No, I've said nonsense.

It can't be AW2 version nor the Antlr bin I've setup in my PATH, as they are the latest.

But I've always a problem for ignoring all '\n' inputs, as I got a token whose value is '\n\n1.',

instead of just having the token '1' and the token '.' .

Regards

[John B. Brodie]

Greetings!

Should the NO_SPECIAL_GRAPH negation also include the \t \r and \n chars?

Hope this helps....
   -jbb

[laurent bernabe]

Hello,

Thank you for your suggestion.

Now all the newlines are removed :).

Now I've got to improve the tokens for the notation rule.

Regards

[laurent bernabe]

Far better now : all tokens are well grouped.

I am skipping the points in move numbers.

There's just a little problem with some parenthesis which seem to have been excluded from the parser stream.

Regards.

[laurent bernabe]

Finally got it !!!

Thank you for all your answers :)

My final grammar is joined to this message.