Re: Proposal: add uloc_setKeywordValue and uloc_removeKeyword

OK Updated proposal:
  * Keywords will be sorted as they are added, the function will not 
re-sort unrelated keywords
  * I retract the proposal for the removeKeyword function.
  * 0-length OR null keyword values will result in removal


On 25-Awi-2004, at 2:43 AM, George Rhoten wrote:

> I like the single uloc_setKeywordValue() function idea.  It seems like
> less duplication of effort for removing keywords.
>
> Regarding the question in the proposal, I presume that the keywords 
> will
> be sorted to any rules that uloc_getName uses for keywords.
>
> I presume that a Locale class API is not in this proposal, since we 
> don't
> provide similar set functions for languages, scripts, countries or
> variants.
>
> 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]>
> Sent by: [email protected]
> 08/24/2004 05:00 PM
>
> To
> "Steven R. Loomis" <[email protected]>
> cc
> Icu <[email protected]>
> Subject
> Re: Proposal:  add uloc_setKeywordValue and uloc_removeKeyword
>
>
>
>
>
>
> I don't know of a need for zero-length keyword values, I was just
> saying they could be accommodated if necessary. Yes, I was suggesting
> removing the removeKeyword function.
>
> Deborah
>
> On Aug 24, 2004, at 4:36 PM, Steven R. Loomis wrote:
>
>> That's already in the proposal, but i think you mean a modification to
>> (1) allow 0-length values to be set and (2) remove the removeKeyword
>> function
>>
>> actually i am implementing remove as just calling set with a null
>>
>> I believe getKeywordValue returns 0 (length 0) for both a missing and
>> a 0-length value.   I'd rather not allow 0-length values in that case.
>>
>>>  * @param keywordValue value of the keyword to be set. If 0-length or
>>>  *  NULL, will result in the keyword being removed.
>>
>>
>> On 24-Awi-2004, at 9:44 PM, Deborah Goldsmith wrote:
>>
>>> A possible alternative: use a NULL keywordValue passed to
>>> uloc_setKeywordValue to mean "remove". I don't know if empty keyword
>>> values are allowed, but even if they are that can be distinguished
>>> from the case of the pointer being NULL.
>>>
>>> This makes the API a little smaller, if that's important. This is the
>>> way I did it in my own implementation.
>>>
>>> Deborah
>>>
>>> On Aug 24, 2004, at 10:26 AM, Steven R. Loomis wrote:
>>>
>>>>
>>>> Proposal:  add uloc_setKeywordValue and uloc_deleteKeyword
>>>> Date:    2004 Aug 24
>>>> Bug: 3651 http://www.jtcsv.com/cgibin/icu-bugs?findid=3651
>>>>
>>>>  Quoting from the bug: "Since the UCalendarType argument to 
>>>> ucal_open
>>>> is deprecated, there isn't a good way to set an explicit calendar to
>>>> override the locale. By contrast, for a numeric format you can set
>>>> the
>>>> currency programmatically.
>>>>   "Provide functions in uloc.h for setting keywords as well as
>>>> getting
>>>> them. Right now, you have to iterate through all the keywords
>>>> yourself
>>>> to see if that keyword is already present (you don't want to set two
>>>> copies of the same keyword). In addition to uloc_getKeywordValue
>>>> there would be uloc_setKeywordValue and uloc_deleteKeyword."
>>>>
>>>> Q: should the keywords be sorted?
>>>>
>>>> /**
>>>>  * Set the value of the specified keyword.
>>>>  * NOTA BENE: Unlike almost every other ICU function which takes a
>>>>  * buffer, this function will NOT truncate the output text. If a
>>>>  * BUFFER_OVERFLOW_ERROR is received, it means that the original
>>>>  * buffer is untouched. This is done to prevent incorrect or 
>>>> possibly
>>>>  * even malformed locales from being generated and used.
>>>>  * Missing keywords on delete won't be marked as an error.
>>>>  *
>>>>  * @param keywordName name of the keyword to be set. Case
>>>> insensitive.
>>>>  * @param keywordValue value of the keyword to be set. If 0-length 
>>>> or
>>>>  *  NULL, will result in the keyword being removed.
>>>>  * @param buffer input buffer containing locale to be modified.
>>>>  * @param bufferCapacity capacity of receiving buffer
>>>>  * @param status containing error code - buffer not big enough.
>>>>  * @return the length needed for the buffer
>>>>  * @see uloc_removeKeyword
>>>>  * @draft ICU 3.2
>>>>  */
>>>> U_DRAFT int32_t U_EXPORT2
>>>> uloc_setKeywordValue(const char* keywordName,
>>>>                      const char* keywordValue,
>>>>                      char* buffer, int32_t bufferCapacity,
>>>>                      UErrorCode* status);
>>>>
>>>>
>>>> /**
>>>>  * Remove a specified keyword.
>>>>  * Missing keywords on delete won't be marked as an error.
>>>>  * @param keywordName name of the keyword to be removed. Case
>>>> insensitive.
>>>>  * @param buffer input buffer containing locale to be modified.
>>>>  * @param bufferCapacity capacity of receiving buffer
>>>>  * @param status containing error code - buffer not big enough.
>>>>  * @return the length needed for the buffer
>>>>  * @see uloc_removeKeyword
>>>>  * @draft ICU 3.2
>>>>  */
>>>> U_DRAFT int32_t U_EXPORT2
>>>> uloc_removeKeyword(const char* keywordName,
>>>>                    char* buffer, int32_t bufferCapacity,
>>>>                    UErrorCode* status);
>>>>
>>>> _______________________________________________
>>>> 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
>>
>
> _______________________________________________
> 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