Re: is this legal ntriples? / raptor ntriple parser bug with long bnode labels?

>>>Dan Brickley said:
> danbri@fireball:~/servers/cvs/w3.org/WWW/2001/sw/Europe/200301/x2r/ht/a1$ mak
    e checkrdf
> rdfdump --ntriples file:po-rdf.nt
> rdfdump: Parsing URI file:po-rdf.nt
> rdfdump: Error - URI file:po-rdf.nt:1 column 8 - Saw '7', expected <URIref> o
    r _:bnodeID
> rdfdump: Error - URI file:po-rdf.nt:2 column 8 - Saw '7', expected <URIref> o
    r _:bnodeID
> rdfdump: Parsing returned 0 statements
> make: *** [checkrdf] Error 1
> danbri@fireball:~/servers/cvs/w3.org/WWW/2001/sw/Europe/200301/x2r/ht/a1$ cat
     po-rdf.nt
> _:id2590758 <http://example.com/nomappednsuri/#productName>  "Lawnmower" .
> _:id2590758 <http://example.com/nomappednsuri/#quantity>  "1" .
> 
> 
> ...is it that the bnode labels are too long? That appears to be the case:

No it was a bug in the digit checking code, '0's were mistakenly
matched as illegal.  Fixed in CVS or apply the patch below.

Dave

Index: ntriples_parse.c
===================================================================
***************
*** 269,275 ****
  
  /* These are for 7-bit ASCII and not locale-specific */
  #define IS_ASCII_ALPHA(c) (((c)>0x40 && (c)<0x5B) || ((c)>0x60 && (c)<0x7B))
! #define IS_ASCII_DIGIT(c) ((c)>0x30 && (c)<0x3A)
  #define IS_ASCII_PRINT(c) ((c)>0x1F && (c)<0x7F)
  
  /**
--- 269,275 ----
  
  /* These are for 7-bit ASCII and not locale-specific */
  #define IS_ASCII_ALPHA(c) (((c)>0x40 && (c)<0x5B) || ((c)>0x60 && (c)<0x7B))
! #define IS_ASCII_DIGIT(c) ((c)>0x2F && (c)<0x3A)
  #define IS_ASCII_PRINT(c) ((c)>0x1F && (c)<0x7F)
  
  /**

Received on Tuesday, 21 January 2003 12:27:01 UTC