Thanks to all. it is now working smoothly.
I dug through some old documentation and found this jem:
"In general it is best to use only letters and numbers in any name. If you
wish to use special characters you do so at your own risk; they may not be
supported by all of the tools.
In general, the following characters are NOT allowed to be in an Attribute
name:
( ) , . ' " (and no white space - blank, \t, \n etc)"
Thus my L0 rule for attribute names became
<id> ~ [^\s\.(),'"]+
and the rule for literal comparison words became
<literal_word> ~ [^\s(),'"]+
The single and double quoted string contents became
<string_without_single_quote_or_vertical_space> ~ [^']+
<string_without_double_quote_or_vertical_space> ~ [^"]+
The current level of the project is here https://ibm.biz/BdRnrC - although
frankly it has zero interest outside of the customers running this product.
I have produced advisory messages for 20 common failure cases.
One funny thing - I used the eval {} to get control after a parse
exception. On my largest test case of 6000 situation formula, there were 6
parse failures. One was a null string. One was a series of quoted strings
inside parenthesis - except that a quote was missing!! One was a single
colon character... and so on.
Anyway, I expect this will make the product more efficient and reliable. It
will probably get some folks irritated because some of the problems are a
style choice.
Thanks again.
John Alvord
On Fri, Mar 28, 2014 at 9:57 PM, Jeffrey Kegler <
[email protected]> wrote:
> Cleaning off my desktop, I also have this note -- in your gist this line
>
> <idx> ~ [^\s\.]+
>
> also looks like it might be dangerously broad. It's useful, even when
> using LATM, to ask for every lexeme, "What stops it?" Lexemes which easily
> go on forever make sense in some places, but more usually it's a potential
> problem. Ron Savage's website keeps a lot of SQL
> BNF<http://savage.net.au/SQL/index.html>,
> which might be useful for determining for safe (or at least
> standard-conformant) ways to restrict the syntax.
>
> Btw, there is also an IRC channel for Marpa: #marpa on irc.freenode.net.
> Please feel free to use whichever forum you are most comfortable with.
>
>
> -- jeffrey
>
> On 03/27/2014 12:26 PM, John Alvord wrote:
>
> Here is the gist
>
> https://gist.github.com/jalvo2014/9816011
>
> I will get started on that debugging documentation, thanks
>
> John Alvord
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "marpa parser" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> For more options, visit https://groups.google.com/d/optout.
>
--
You received this message because you are subscribed to the Google Groups
"marpa parser" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
