[Issue 11254] std.string.strip is not nothrow

[d-bugmail]

http://d.puremagic.com/issues/show_bug.cgi?id=11254


monarchdo...@gmail.com changed:

   What|Removed |Added

 Status|NEW |RESOLVED
 CC||monarchdo...@gmail.com
 Resolution||INVALID


--- Comment #1 from monarchdo...@gmail.com 2013-10-14 05:55:11 PDT ---
(In reply to comment #0)
> import std.string: strip;
> void main() nothrow {
> " hello ".strip;
> }
> 
> 
> 
> dmd 2.064beta gives:
> 
> test.d(3): Error: 'std.string.strip!(immutable(char)).strip' is not nothrow
> test.d(2): Error: function 'D main' is nothrow yet may throw
> 
> 
> I don't know if this can be done. Often string functions need to decode UTF,
> and this could raise exceptions. In most cases, or for ASCII strings, a strip
> can't throw exceptions.
> 
> If this can't be done then please close down this issue.

strip is a unicode aware function, that can remove unicode whites, so it *must*
decode. So even if "most of the time", it won't throw, in the generic case, it
can.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
--- You are receiving this mail because: ---

[Issue 11254] std.string.strip is not nothrow

[d-bugmail]

http://d.puremagic.com/issues/show_bug.cgi?id=11254



--- Comment #2 from bearophile_h...@eml.cc 2013-10-14 09:46:51 PDT ---
(In reply to comment #1)

> strip is a unicode aware function, that can remove unicode whites, so it 
> *must*
> decode. So even if "most of the time", it won't throw, in the generic case, it
> can.

Some possible alternative solutions:
- A strip-like function that works on ubyte[] (the return type of
std.string.representation if you give it a string);
- A compile-time switch for std.string.strip that compiles out the
unicode-aware parts.
- A std.ascii.astrip nothrow function designed to work only on ASCII
strings/char[].

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
--- You are receiving this mail because: ---

[Issue 11254] std.string.strip is not nothrow

[d-bugmail]

http://d.puremagic.com/issues/show_bug.cgi?id=11254



--- Comment #3 from monarchdo...@gmail.com 2013-10-14 11:50:50 PDT ---
(In reply to comment #2)
> (In reply to comment #1)
> 
> > strip is a unicode aware function, that can remove unicode whites, so it 
> > *must*
> > decode. So even if "most of the time", it won't throw, in the generic case, 
> > it
> > can.
> 
> Some possible alternative solutions:
> - A strip-like function that works on ubyte[] (the return type of
> std.string.representation if you give it a string);
> - A compile-time switch for std.string.strip that compiles out the
> unicode-aware parts.
> - A std.ascii.astrip nothrow function designed to work only on ASCII
> strings/char[].

You should try the new generic std.algorithm.strip:

//
import std.string, std.ascii, std.algorithm;

void main(string[] args) nothrow pure
{
string s = " hello! ";
s = cast(string)s.representation.strip!isWhite();
}
//

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
--- You are receiving this mail because: ---

[Issue 11254] std.string.strip is not nothrow

[d-bugmail]

http://d.puremagic.com/issues/show_bug.cgi?id=11254


Jonathan M Davis  changed:

   What|Removed |Added

 CC||jmdavisp...@gmx.com


--- Comment #4 from Jonathan M Davis  2013-10-14 19:41:32 
PDT ---
I think that we should probably move towards overloading string functions with
ubyte[] so that they can have ASCII-specific versions, and more of those would
be able to be nothrow.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
--- You are receiving this mail because: ---