| Author |
[split $longstring {}]
|
|
| Andreas Leitgeb 2006-01-27, 3:59 am |
| Kaitzschu <kaitzschu@kaitzschu.cjb.net.nospam.plz.invalid> wrote:
> split {abcdefghijklmnopqrstuvwxyz} {}
I wonder, if for (very) long strings e.g. > 100000 and more
it would be worth the trouble to reuse the Tcl_Objects for
the single characters, or whether this is already happening.
Just like with [binary scan $longdata c* var].
| |
| suchenwi 2006-01-27, 3:59 am |
| It is happening, for several years now already :^)
| |
| Donal K. Fellows 2006-01-27, 3:59 am |
| Actually, I made that optimization work for [split ... {}] first. The
[binary scan] one came later (and was much harder to write too).
Donal.
| |
| Andreas Leitgeb 2006-01-27, 7:59 am |
| Donal K. Fellows <donal.k.fellows@man.ac.uk> wrote:
> Actually, I made that optimization work for [split ... {}] first. The
> [binary scan] one came later (and was much harder to write too).
> Donal.
Sorry, I mixed it up with a situation discussed last year
about split. (but that was about split with non-empty
split-chars, and is unrelated to this case.)
| |
| Donal K. Fellows 2006-01-27, 7:59 am |
| Andreas Leitgeb wrote:
> Sorry, I mixed it up with a situation discussed last year
> about split. (but that was about split with non-empty
> split-chars, and is unrelated to this case.)
That's the case which isn't optimized on the grounds that it is probably
not worth doing on the normal lengths of data and frequencies of words
that people use.
Donal.
|
|
|
|