On 24/01/2011, at 6:51 AM, john skaller wrote:
>
> On 24/01/2011, at 6:41 AM, john skaller wrote:
>
>>
>> Be warned, to implement the match thing I have changed the merge rule to keep
>> all possible interpretations of expressions. If you do something ambiguous it
>> may propagate into statements. These don't allow ambiguity.
>
>
> Oops.. now I get a weird bug in tools/flx2html.flx saying I have duplicate
> Flx2Html,
> which is module .. no idea why my changes caused this! May be to backtrack
> on my last promise .. :)
Ok, I have fixed this problem: the original cause was that I was using the
"aging" rule
for merges to support
x.(0) --> tuple component 0
but this is an override as a special case of a more general operation. This is
fixed now.
Instead the special case is handled in the compiler instead of the parser.
There is also a change to the lexer: you can no longer write .5 for 1/2. Instead
you must write 0.5. This now allows:
x.0
to mean the 0'th projection of x (no parens needed). Much cleaner.
To Dypgen list: the symptom of this problem was a duplicate MODULE name FFS!
I have no idea how "keep_all" merge rule for expressions could possibly
propagate
like this! Modules are statements!
It is not as if there was a duplicate parse: Felix bugs out if there is a
duplicate
parse of the whole file.
Anyhow after removing this duplicate production ... the standard library parse
time dropped from 200 seconds to 125 seconds!!!!!!!! I have no idea why,
because previously the duplication was removed by a merge rule.
Now there's no duplication, but there merge rule keeps all parses.
--
john skaller
[email protected]
------------------------------------------------------------------------------
Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)!
Finally, a world-class log management solution at an even better price-free!
Download using promo code Free_Logger_4_Dev2Dev. Offer expires
February 28th, so secure your free ArcSight Logger TODAY!
http://p.sf.net/sfu/arcsight-sfd2d
_______________________________________________
Felix-language mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/felix-language