Re: [icu-bug] incoming/3746

> Then what I stated before, is now my proposed fix for this issue.

Assuming that everyone is agreeable that ulocdata_getExemplarSet should 
return a case-insensitive set, then I'm OK with that solution.

Deborah Goldsmith
Internationalization, Unicode liaison
Apple Computer, Inc.
[email protected]

On May 11, 2004, at 3:10 PM, George Rhoten wrote:

> I see now.  You are talking about the other internal function, which is
> called closeOver. :-(
>
> Then what I stated before, is now my proposed fix for this issue.
>
> George Rhoten
> IBM Globalization Center of Competency/ICU  San José, CA, USA
> ICU main website: http://oss.software.ibm.com/icu/index.html
>
>
>
>
> Deborah Goldsmith <[email protected]>
> 05/11/2004 02:50 PM
>
>         To:     George Rhoten/San Jose/IBM@IBMUS
>         cc:     Mark Davis <[email protected]>,
> [email protected]
>         Subject:        Re: [icu-bug] incoming/3746
>
>
> It's a public API if you have a pattern. It doesn't help if you've
> already got a USet: give an existing USet, the only way to apply the
> case closure to it is to convert it back into a pattern, and then apply
> that pattern to a new USet with the appropriate options. That seems
> rather circuitous.
>
> What Mark (I think) and I are saying is that it would be useful to be
> able to apply the appropriate closures to an existing *set*. I know
> it's already possible if you have the *pattern*.
>
> Deborah Goldsmith
> Internationalization, Unicode liaison
> Apple Computer, Inc.
> [email protected]
>
> On May 11, 2004, at 2:38 PM, George Rhoten wrote:
>
>> This is already public API.  You can read the uset.h API 
>> documentation,
>> and it will say the same thing as I'm saying.  Mark is remembering the
>> _old_ state of the C++ API that this API mirrors.  The
>> uset_applyPattern
>> API was needed by ulocdata, and by icuio.  This is not internal.
>>
>> I remember this public API proposal quite well:
>> http://oss.software.ibm.com/pipermail/icu/2003-October/005775.html
>>
>> I'm not proposing the fix, I'm just stating what would fix this issue,
>> which I agree is an issue.  I know, it's a minor difference in what 
>> I'm
>> saying ;-)
>>
>> George Rhoten
>> IBM Globalization Center of Competency/ICU  San José, CA, USA
>> ICU main website: http://oss.software.ibm.com/icu/index.html
>>
>>
>>
>>
>> Deborah Goldsmith <[email protected]>
>> 05/11/2004 02:20 PM
>>
>>         To:     George Rhoten/San Jose/IBM@IBMUS
>>         cc:     Mark Davis <[email protected]>,
>> [email protected]
>>         Subject:        Re: [icu-bug] incoming/3746
>>
>>
>> Is this a fix that's proposed for ulocdata_getExemplarSet?
>>
>> I will echo Mark in that I think having public API for doing the
>> appropriate closures on an *existing* USet would be useful. That way
>> clients could choose whether they wanted the case closure or not.
>>
>> Deborah Goldsmith
>> Internationalization, Unicode liaison
>> Apple Computer, Inc.
>> [email protected]
>>
>> On May 11, 2004, at 10:44 AM, George Rhoten wrote:
>>
>>> Here is the 2 line fix for this issue.
>>>
>>>         uset_applyPattern(fillIn, exemplarChars, len,
>>>                           USET_IGNORE_SPACE | USET_CASE_INSENSITIVE,
>>> status);
>>>
>>> This API is draft 2.8.  It is not internal.
>>>
>>> George Rhoten
>>> IBM Globalization Center of Competency/ICU  San José, CA, USA
>>> ICU main website: http://oss.software.ibm.com/icu/index.html
>>>
>>>
>>>
>>>
>>> "Mark Davis" <[email protected]>
>>> Sent by: [email protected]
>>> 05/11/2004 10:21 AM
>>>
>>>         To:     <[email protected]>
>>>         cc:
>>>         Subject:        Re: [icu-bug] incoming/3746
>>>
>>>
>>> Hmmm. We do have internal API that computes the case-closure of a 
>>> set,
>>> and
>>> that
>>> computes the canonical-equivalence closure. Perhaps we should think
>>> about
>>> making
>>> those public APIs in the future.
>>>
>>> Mark
>>> __________________________________
>>> http://www.macchiato.com
>>> ► शिष्यादिच्छेत्पराजयम् ◄
>>>
>>> ----- Original Message -----
>>> From: <[email protected]>
>>> To: <[email protected]>;
>>> <[email protected]>
>>> Sent: Tue, 2004 May 11 08:53
>>> Subject: [icu-bug] incoming/3746
>>>
>>>
>>>> new message incoming/3746
>>>> URL: http://www.jtcsv.com/cgi-bin/icu-bugs?findid=3746
>>>>
>>>> ====> ORIGINAL MESSAGE FOLLOWS <====
>>>>
>>>> From: [email protected]
>>>> Date: Tue May 11 11:53:18 2004
>>>> Subject: Should exemplar character set be case-insensitive?
>>>>
>>>> Full_Name: Deborah Goldsmith
>>>> Version: 2.8
>>>> OS: all
>>>> PROJECT: ICU4C
>>>> JAVA:
>>>> Submission from: (NULL) (24.6.137.96)
>>>>
>>>>
>>>> Right now the API that returns the exemplar character set in
>>>> ulocdata.h
>>> just
>>>> creates a set directly from the (lowercase-only) locale data.
>>>> Shouldn't
>>> it be
>>>> specifying the case-insensitive option when creating the set from 
>>>> the
>>> pattern
>>> in
>>>> the locale data? One problem is that once the set is returned, the
>>>> only
>>> way to
>>>> get the case-insensitive version is to turn it back into a pattern,
>>>> then
>>>> recreate the set with the case-insensitive option. There is no 
>>>> public
>>> API to
>>>> compute the case-closure of an existing set (that would also be an
>>> acceptable
>>>> solution).
>>>>
>>>> _______________________________________________
>>>> icu-bugrfe mailing list
>>>> [email protected]
>>>>
>>> http://oss.software.ibm.com/developerworks/oss/mailman/listinfo/icu-
>>> bugrfe
>>>>
>>>
>>> _______________________________________________
>>> icu mailing list
>>> [email protected]
>>> http://oss.software.ibm.com/developerworks/oss/mailman/listinfo/icu
>>>
>>> _______________________________________________
>>> icu mailing list
>>> [email protected]
>>> http://oss.software.ibm.com/developerworks/oss/mailman/listinfo/icu
>>
>>
>
>