Re: [sqlite] toupper/tolower

[Zbigniew Baniewski]

I'm sorry: totally missed `The "function" method' :(  So, it seems, that
the problem's solved. :)
-- 
pozdrawiam / regards

Zbigniew Baniewski
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users

Re: [sqlite] toupper/tolower

[Zbigniew Baniewski]

On Sun, Jun 21, 2009 at 08:34:23PM -0400, Igor Tandetnik wrote:

> http://www.sqlite.org/cvstrac/fileview?f=sqlite/ext/icu/README.txt

Yes, I've found libraries made by someone, and ready to use. Unfortunately,
they are of several MB size... :(

No, I would to keep the application "compact", and because of this, the
better way - if possible - would be suggested completion of TCL API. Even
that I'm aware, that such external method will be slower.
-- 
pozdrawiam / regards

Zbigniew Baniewski
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users

Re: [sqlite] toupper/tolower

[Igor Tandetnik]

Zbigniew Baniewski wrote:
> And yes - alas - upper/lower are unable to do it properly for
> characters "outside ASCII". But I've found some more info:

http://www.sqlite.org/cvstrac/fileview?f=sqlite/ext/icu/README.txt

Igor Tandetnik



___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users

Re: [sqlite] toupper/tolower

[Zbigniew Baniewski]

On Sun, Jun 21, 2009 at 05:12:18PM -0700, Cory Nelson wrote:

> Unfortunately, no.  A correct toupper/tolower would require locale
> support and passing in a string pointer, not a char.  Even a wide
> character can't store all the information needed to make something
> upper/lower.

OK, maybe different way then:

There is - in SQLite's TCL API - a "collate" method:

#v+
  This method registers new text collating sequences. There are two
  arguments: the name of the collating sequence and the name of a TCL
  procedure that implements a comparison function for the collating sequence.

  For example, the following code implements a collating sequence called
  "NOCASE" that sorts in text order without regard to case:

proc nocase_compare {a b} {
  return [string compare [string tolower $a] [string tolower $b]]
}
db collate NOCASE nocase_compare
#v-

Perhaps could be possible to add to that API something like external
"tolower/toupper" methods, that could be registered from within TCL?
-- 
pozdrawiam / regards

Zbigniew Baniewski
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users

Re: [sqlite] toupper/tolower

[Cory Nelson]

On Sun, Jun 21, 2009 at 4:34 PM, Zbigniew Baniewski wrote:
> In the docs there is a remark:
>
> #v+
>  upper(X)  Return a copy of input string X converted to all upper-case
>  letters. The implementation of this function uses the C library routine
>  toupper() which means it may not work correctly on non-ASCII UTF-8
>  strings.
> #v-
>
> And yes - alas - upper/lower are unable to do it properly for characters
> "outside ASCII". But I've found some more info:
>
> http://www.warpspeed.com.au/cgi-bin/inf2html.cmd?..%5Chtml%5Cbook%5CToolkt40%5CXPG4REF.INF+297
>
> #v+
>  Note: toupper and tolower can only be used for single-byte characters.
>  towupper and towlower should be used for case conversion of wide characters
>  that are equivalent to both single-byte and double-byte characters.
> #v-
>
> It looks, like the problem could be easily solved just by replacing
> "toupper/tolower" occurences with "towupper/towlower". Am I right?
> If so - perhaps could be such change introduced in new version?

Unfortunately, no.  A correct toupper/tolower would require locale
support and passing in a string pointer, not a char.  Even a wide
character can't store all the information needed to make something
upper/lower.

-- 
Cory Nelson
http://int64.org
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users