[m-rev.] for review: improve string__replace and string__replace_all
Ralph Becket
rafe at cs.mu.OZ.AU
Tue Dec 7 11:44:24 AEDT 2004
Peter Ross, Monday, 6 December 2004:
> On Mon, Dec 06, 2004 at 11:40:20AM +1100, Ralph Becket wrote:
> > Peter Ross, Sunday, 5 December 2004:
> > > Index: library/string.m
> > > ===================================================================
> > > RCS file: /home/mercury1/repository/mercury/library/string.m,v
> > > retrieving revision 1.221
> > > diff -u -r1.221 string.m
> > > --- library/string.m 10 Nov 2004 07:12:38 -0000 1.221
> > > +++ library/string.m 5 Dec 2004 16:26:38 -0000
> > > @@ -609,7 +609,9 @@
> > >
> > > string__replace_all(Str, Pat, Subst, Result) :-
> > > ( Pat = "" ->
> > > - copy(Str, Result)
> > > + F = (func(C, L) = [char_to_string(C) ++ Subst | L]),
> > > + Foldl = string__foldl(F, Str, []),
> > > + Result = append_list([Subst | list__reverse(Foldl)])
> > > ;
> > > PatLength = string__length(Pat),
> > > ReversedChunks = replace_all(Str, Pat, Subst, PatLength, 0, []),
> >
> > Just as a minor efficiency hack, I'd move the char_to_string(C) computation
> > out of the lambda - loop invariant hoisting isn't turned on until -O5 or
> > 6 IIRC and even then it might not hoist it.
> >
> I would love to move it out, but then the program wouldn't be correct.
> As the value for C supplied is for each character in the string, so it
> is always changing.
Ah, yes. I am a fool.
Can I suggest the following implementation for the then-goal?
F = ( func(C, L) = [Subst, char_to_string(C) | L] ),
Result = string__append_list(foldr(F, Str, [Subst]))
This avoids the call to reverse and some excess string copying.
-- Ralph
--------------------------------------------------------------------------
mercury-reviews mailing list
post: mercury-reviews at cs.mu.oz.au
administrative address: owner-mercury-reviews at cs.mu.oz.au
unsubscribe: Address: mercury-reviews-request at cs.mu.oz.au Message: unsubscribe
subscribe: Address: mercury-reviews-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------
More information about the reviews
mailing list