Re: [icu-support] Collation question

[Mark D. <mar...@jt...>]

With the first case, I believe what is happening is that one or more of t=
he
characters are whitespace, so what the Collator sees is < < which is an
error. You probably have to double-slash them.

For the second, we have a long-standing design (since 2002) for changing =
the
order of scripts ( see

http://dev.icu-project.org/cgi-bin/viewcvs.cgi/*checkout*/icuhtml/design/=
collation/ICU_collation_design.htm

and search for "scriptOrder" and "Script Order"). But it has never been a
priority among our tasks, so you have to do it the hard way, by copying t=
he
UCA rules for the script + whatever script tailorings there are.

This can be done programmatically, I think. Eg to put tamil in front of
latin, do something like the following.

Let X be the UCA rules
Let A be the position of the first character matching
[[:script=3Dtamil:][:letter:]]
Let B be the position of the first character after A matching
[:^script=3Dtamil:]
Let X1 be the substring of X from A upto but not including B
Let Y be the Tamil rules.
Let Z be "& [before 1] a < " + X1 + Y
Create new collator from Z.

For multiple scripts, just add more pairs to Z.

This will probably need some tweeks beyond this, but you could start from
here.

=E2=80=8EMark

----- Original Message -----=20
From: "Rajeev J Sebastian" <raj...@di...>
To: <icu...@li...>
Sent: Friday, May 27, 2005 10:09
Subject: [icu-support] Collation question


> Hi,
>
> I have a question regarding collator implementation for ml_IN locale in
ICU
> 2.6
>
> I need to reassign the positions of u200d and u200c as shown below:
>
>
>
>
>     CollationElements {
>         Version { "1.0" }
>         Sequence { "[normalization on]&[top]"
>           "<\u200d<\u200c"
>  ....
>
>
> However, this always results in the following error while building:
>
>
>
> ICU_DATA=3D../data/out/build
>
LD_LIBRARY_PATH=3D../common:../i18n:../tools/toolutil:../layout:../layout=
ex:..
/extra/ustdio:../tools/ctestfw:../data/out:../data:../stubdata/:
> $LD_LIBRARY_PATH  ../tools/genrb/genrb -k -q -p icudt26l -s
../data/locales
> -d ../data/out/build ml_IN.txt
> ../data/locales/ml_IN.txt:108: parse error. Stopped parsing with
> U_INVALID_FORMAT_ERROR
> couldn't parse the file ml_IN.txt. Error:U_INVALID_FORMAT_ERROR
> make[1]: *** [../data/out/build/icudt26l_ml_IN.res] Error 3
> make[1]: Leaving directory `/home/kane/font_maker/icu/icu/source/data'
> make: *** [all-recursive] Error 2
>
>
>
>
> Trying the rule in Locale Explorer works. Perhaps it is because 200c/20=
0d
are
> already assigned as equal to [last tertiary ignorable] ? Is there any w=
ay
to
> reassign weights of these two codepoints in my locale ?
>
> Also, how do I move sections around ? for e.g., I would like to include=
 in
the
> ml_IN locale, the following order:
>
> malayalam, tamil, latin, devanagari ....
>
> So, how can I move say the Malayalam section to just before Tamil and s=
o
on ?
> Or do I have to copy the collation rules from the appropriate files int=
o
my
> ml_IN locale ?
>
>
> Rajeev J Sebastian
>
>
>
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by Yahoo.
> Introducing Yahoo! Search Developer Network - Create apps using Yahoo!
> Search APIs Find out how you can build Yahoo! directly into your own
> Applications - visit http://developer.yahoo.net/?fr=3Doffad-ysdn-ostg-q=
22005
> _______________________________________________
> icu-support mailing list - icu...@li...
> To Un/Subscribe: https://lists.sourceforge.net/lists/listinfo/icu-suppo=
rt
>
>