Hello.
Dne 10. dubna 2012 22:29 Martin Sucha <such...@uniba.sk> napsal(a):
> Hi Tobias,
>
> This seems to be useful functionality, but I have a few comments:
>
> On Tuesday 10 April 2012 19:09:01 Tobias Börtitz wrote:
>> Over the above I added the letter_check(wchar_t) function to the libc. It
>> checks whether the assigned character contains a printable character
>> (a-zA-Z) or not.
> First of all, HelenOS uses unicode strings so I think checking for (a-zA-Z) is
> not correct. A word may start with different characters as well, e.g.
> "čučoriedková" is a word that would not match the check.
>
> Secondly, I'm not familar with unicode, but I think there is no definition of
> what a unicode "printable character" is - even some characters that are not
> printable by themselves may be printed when used in a surrogate pair (correct
> me if I'm wrong). What about checking whether a character is not a word
> delimiter instead?
>
> As for what is a word delimiter, I'd suggest using more characters as word
> delimiters (e.g. punctuation characters such as comma, em dash, etc. or
> various types of spaces that are present in unicode).
I agree. If you would replace in pt_is_word_beginning the last switch
and the call to letter check with something like:
if (is_word_delimiter(end_str[0])) {
ret = false;
goto exit;
}
ret = is_word_delimiter(*prev_char);
where is_word_delimiter() would check for space, tab, comma etc., your
patch would correctly jump over words. At least, it would jump
correctly over examples in /textdemo, regardless of the language.
One more thing, Tobias - please, do put space between "if" and the condition.
Cheers,
- Vojta
>
> Regards,
> Martin Sucha
>
>
> _______________________________________________
> HelenOS-devel mailing list
> HelenOS-devel@lists.modry.cz
> http://lists.modry.cz/cgi-bin/listinfo/helenos-devel
_______________________________________________
HelenOS-devel mailing list
HelenOS-devel@lists.modry.cz
http://lists.modry.cz/cgi-bin/listinfo/helenos-devel
