Comment #0 by bearophile_hugs — 2012-03-11T14:07:42Z
Is this difference/inconsistency between split() and splitter() desired and good?
import std.string, std.array, std.algorithm, std.range;
void main() {
char[] s = cast(char[])[131, 64, 32, 251, 22];
assert(std.string.split(s).length == 2); // no error
assert(walkLength(std.array.splitter(s)) == 2); // Invalid UTF-8 sequence
assert(walkLength(std.algorithm.splitter(s)) == 2); // Invalid UTF-8 sequence
}
Output, DMD 2.059head:
std.utf.UTFException@std\utf.d(645): Invalid UTF-8 sequence (at index 1)
----------------
...\dmd2\src\phobos\std\array.d(469): dchar std.array.front!(char[]).front(char[])
...\dmd2\src\phobos\std\algorithm.d(2110): D3std9algorithm47__T8splitterS28...
...\dmd2\src\phobos\std\range.d(971): D3std5range97__...
----------------
Comment #1 by monarchdodra — 2012-10-22T23:06:22Z
(In reply to comment #0)
> Is this difference/inconsistency between split() and splitter() desired and
> good?
>
>
> import std.string, std.array, std.algorithm, std.range;
> void main() {
> char[] s = cast(char[])[131, 64, 32, 251, 22];
> assert(std.string.split(s).length == 2); // no error
> assert(walkLength(std.array.splitter(s)) == 2); // Invalid UTF-8 sequence
> assert(walkLength(std.algorithm.splitter(s)) == 2); // Invalid UTF-8
> sequence
> }
>
>
> Output, DMD 2.059head:
>
> std.utf.UTFException@std\utf.d(645): Invalid UTF-8 sequence (at index 1)
> ----------------
> ...\dmd2\src\phobos\std\array.d(469): dchar
> std.array.front!(char[]).front(char[])
> ...\dmd2\src\phobos\std\algorithm.d(2110): D3std9algorithm47__T8splitterS28...
> ...\dmd2\src\phobos\std\range.d(971): D3std5range97__...
> ----------------
This is a bug in string.split (which is actually a public import of array.split).
Currently array.split only supports ascii white, and is oblivious to longer utf whites (but it does work on unicode).
Comment #2 by bearophile_hugs — 2013-11-18T02:36:29Z