parser problem?

[Ben Abbott]

On Mar 13, 2009, at 3:36 PM, John W. Eaton wrote:

> On 13-Mar-2009, Ben Abbott wrote:
>
> | On Mar 13, 2009, at 10:56 AM, Ben Abbott wrote:
> |
> | > Do I infer correctly that this is a parser problem?
> | >
> | > octave:1> warning error
> | > octave:2> help cellstr
> | > error: potential auto-insertion of `,' near line 47 of file / 
> Users/
> | > bpabbott/Development/mercurial/octave-3.1.54/scripts/help/which.m
> | > error: called from:
> | > error:   /Users/bpabbott/Development/mercurial/octave-3.1.54/ 
> scripts/
> | > help/help.m at line 81, column 5
> | > octave:2> help test
> | > error: potential auto-insertion of `,' near line 89 of file / 
> Users/
> | > bpabbott/Development/mercurial/octave-3.1.54/scripts/strings/ 
> strcat.m
> | > error: called from:
> | > error:   /Users/bpabbott/Development/mercurial/octave-3.1.54/ 
> scripts/
> | > help/__makeinfo__.m at line 154, column 12
> | > error:   /Users/bpabbott/Development/mercurial/octave-3.1.54/ 
> scripts/
> | > help/__makeinfo__.m at line 111, column 16
> | > error:   /Users/bpabbott/Development/mercurial/octave-3.1.54/ 
> scripts/
> | > help/help.m at line 62, column 25
> | > octave:2>
> | >
> | > Line 47 of which.m is
> | >
> | > 	varargout = {m.file};
> | >
> | > Line 89 of strchr.m is
> | >
> | >        varargin{nv}(:) = {str};
> | >
> | > From the Octave prompt
> | >
> | > octave:6> warning error
> | > octave:7> a = "a";
> | > octave:8> b = {a}
> | > error: potential auto-insertion of `,' near line 8
> | > octave:8> b = {(a)}
> | > b =
> | >
> | > {
> | >  [1,1] = a
> | > }
> | >
> | > octave:9> b = {a(:)}
> | > b =
> | >
> | > {
> | >  [1,1] = a
> | > }
> | >
> | > Ben
> | >
> | > p.s. my last build was about 2 days ago.
> |
> | another example
> |
> | octave:16> a = 1;
> | octave:17> b = {a, a}
> | error: potential auto-insertion of `,' near line 17
> | octave:17> b = {1, 1}
> | b =
> |
> | {
> |    [1,1] =  1
> |    [1,2] =  1
> | }
> |
> | octave:18> b = {a, a,} ## notice the 2nd comma
> | b =
> |
> | {
> |    [1,1] =  1
> |    [1,2] =  1
> | }
>
> The following change will avoid this problem.  But I don't think it is
> a complete solution.  I'd guess that there are some other cases that
> will incorrectly trigger this warning message.
>
> jwe
>
> diff --git a/src/lex.l b/src/lex.l
> --- a/src/lex.l
> +++ b/src/lex.l
> @@ -2812,6 +2812,11 @@
>       if (index_op && ! spc_gobbled)
> 	return;
>
> +      // There is no need to insert a comma if the next token ends a
> +      // matrix or cell array list.
> +      if (c1 == ']' || c1 == '}')
> +	return;
> +
>       maybe_warn_separator_insert (',');
>
>       xunput (',', yytext);

Works for me!

Thanks
Ben