Hi!
On Mar 1, 2009, at 6:52 PM, Jim Starkey wrote:
Before people get seriously silly about pluggable parsers, may I
suggest that a little thought be given to the big picture? Parsing
itself is not a big deal. The execution structure, on the other
hand, is, as it determines the performance potential of the entire
system. Some serious architectural thought at this point, ladies
and gentlemen, could have significant payback over the next three or
four years.
I am fine with seeing what we can do today... but long term?
Personally I find the lack of an API around the item generation to be
a bad idea. That system has just been growing over time.
For DDL we are moving toward a tree defined by proto for instance. It
has an API, it is versioned, and can be easily serialized (heck... the
serialized objects can even be recreated for multiple languages). I'd
really like to see something like this happen long term for the parse
tree.
Cheers,
-Brian
--
_______________________________________________________
Brian "Krow" Aker, brian at tangent.org
Seattle, Washington
http://krow.net/ <-- Me
http://tangent.org/ <-- Software
_______________________________________________________
You can't grep a dead tree.
_______________________________________________
Mailing list: https://launchpad.net/~drizzle-discuss
Post to : [email protected]
Unsubscribe : https://launchpad.net/~drizzle-discuss
More help : https://help.launchpad.net/ListHelp
