On Thu, Mar 01, 2012 at 12:04:39AM +0100, Martin Nowak wrote:
[...]
> Mmh, I've retested and you're right dmd's lexer is about 2x faster.
> The main overhead stems from using ranges and enforce.
>
> Quick profiling shows that 25% is spent in popFront and
> std.utf.stride. Last time I worked on this I rewrote std.utf.decode
> to be much faster. But utf characters are still "decoded" twice, once
> for front and then again for popFront. Also stride uses table lookup
> and can't be inlined.
[...]
One way to not decode characters twice is by using a single-dchar buffer
in your range. Something like:
struct MyRange {
private File src;
char buf[];
dchar readahead;
this(File _src) {
// ... fill up buf from src here
popFront();
}
@property pure dchar front() {
return readahead;
}
void popFront() {
int stride;
readahead = decode(buf, stride);
buf = buf[stride..$];
// ... fill up buf more if needed
}
}
T
--
"A man's wife has more power over him than the state has." -- Ralph Emerson
