Peter Kümmel wrote:
> And am I right that the current plan to complete the Unicode support is:
>
> 1. find a 32bit per character string class: docstring
We have that already.
> 2. find docstring streams
Yes (almost done, we only need to decide how to resolve the character
widening problem).
Abdelrazak Younes wrote:
> Georg Baum wrote:
>> Abdelrazak Younes wrote:
>>
>>> Exactly and that was my argument but Lars insisted that the vector
>>> solution is better. He even said that docstring would probably better
>>> use vector instead.
>>
>> I must have overread that.
>
> Just for referen
Georg Baum wrote:
Abdelrazak Younes wrote:
Exactly and that was my argument but Lars insisted that the vector
solution is better. He even said that docstring would probably better
use vector instead.
I must have overread that.
Just for reference, here are his exact words:
Abdelrazak Younes
Abdelrazak Younes wrote:
> Exactly and that was my argument but Lars insisted that the vector
> solution is better. He even said that docstring would probably better
> use vector instead.
I must have overread that.
Georg
Georg Baum wrote:
Abdelrazak Younes wrote:
But wait... that's docstring we are talking about here ;-)
IMHO we must merge this docstring and the vector.
I don't agree.
Well I think you agree with me below ;-)
I just meant that we don't need two types.
We need
only one type! Lars seems to
Georg Baum wrote:
Abdelrazak Younes wrote:
Georg Baum wrote:
And he is right. Therefore the comment that it should go eventually. The
problem is that we currently have _a lot_ of utf8 encoded data in
std::string. All these strings need to be converted to docstring. It is
not practical to do th
Abdelrazak Younes wrote:
> But wait... that's docstring we are talking about here ;-)
>
> IMHO we must merge this docstring and the vector.
I don't agree.
> We need
> only one type! Lars seems to hate the basic_string based one (docstring)
> so let's just use the vector based one (aka ucs4stri
Abdelrazak Younes wrote:
> Georg Baum wrote:
>> And he is right. Therefore the comment that it should go eventually. The
>> problem is that we currently have _a lot_ of utf8 encoded data in
>> std::string. All these strings need to be converted to docstring. It is
>> not practical to do that in on
Abdelrazak Younes <[EMAIL PROTECTED]> writes:
| Georg Baum wrote:
| > Abdelrazak Younes wrote:
| >
| >> Georg Baum wrote:
| >
| >>> std::string const to_utf8(docstring const & ucs4)
| >>> {
| >>> std::vector const utf8 =
| >>> ucs4_to_utf8(std::vector(ucs4.begin(),
| >>>
Georg Baum wrote:
Abdelrazak Younes wrote:
Georg Baum wrote:
std::string const to_utf8(docstring const & ucs4)
{
std::vector const utf8 =
ucs4_to_utf8(std::vector(ucs4.begin(),
ucs4.end()));
return std::string(utf8.begin(), utf8.end());
}
Hum.
Angus Leeming wrote:
Abdelrazak Younes wrote:
a similar >> operator for reading will be needed in place of your
from_utf8() function. The idea is that we don't need to think about
utf8 at all internally. I am not sure the conversion is done correctly
in the above code but you get the idea.
Abdelrazak Younes wrote:
> Georg Baum wrote:
>> std::string const to_utf8(docstring const & ucs4)
>> {
>> std::vector const utf8 =
>> ucs4_to_utf8(std::vector(ucs4.begin(),
>> ucs4.end()));
>> return std::string(utf8.begin(), utf8.end());
>> }
>
>
Abdelrazak Younes wrote:
I am more and more convinced that we should just implement some stream
operators for ucs4 strings that will deal with utf8 internally and
directly with the output... something like that (not tested and only
with little endian system:
std::ostream& operator<<(
std:
Georg Baum wrote:
We should always be explicit about character set conversions. In my patch
one can already see some problematic cases. This would not be possible if
docstring had a constructor from std::string. If performance is an issue we
could inline these methods, but having from_utf8 and f
Abdelrazak Younes wrote:
> Edwin Leuven wrote:
>> when i load the userguide and keep the pgdn button pressed i get the
>> following crash:
>>
>> terminate called after throwing an instance of 'std::length_error'
>> what(): basic_string::_S_create
>
> Confirmed. Here is my (perhaps stupid) und
Edwin Leuven wrote:
when i load the userguide and keep the pgdn button pressed i get the
following crash:
terminate called after throwing an instance of 'std::length_error'
what(): basic_string::_S_create
Confirmed. Here is my (perhaps stupid) understanding:
The problem is that docstring
16 matches
Mail list logo
