[OpenIndiana-discuss] Bash brace expansion in non-C locales??

Hans J. Albertsson hans.j.albertsson at branneriet.se
Wed Mar 13 08:04:43 UTC 2013


not such an easy thing then...

however. I suppose given a complete dictionary, a system could try 
splitting a word between c and h and see if the resulting subwords are 
in the dictionary. Hmm
On 2013-03-13 08:57, Marcel Telka wrote:
> On Wed, Mar 13, 2013 at 08:01:16AM +0100, Hans J. Albertsson wrote:
>> Can that confusion happen at the start of a word or only inside
>> words?? And how many rulebreaking words are there, can they be
>> enumerated??
> It can happen only in the middle. Only in a case when the word is constructed
> from two initially (semi-)separate words. In my example the word "viachlasný"
> is "viac" + "hlasný" (it is something like multi+voice). It will happen always
> when you combine a word ending by "c" with another word starting with "h". Such
> combinations are not very often used in the language, but because they are
> constructed as a combination of two words, you can construct a lot of such
> words. The situation is even worse because those two (semi-)separate words used
> for the construction might not exist as separate words in such form as it was
> used for the construction.
>
> BTW, the problem is caused by a fact that our "ch" (the single letter) is
> equivalent to Russian "X", with the exception that we do not have a single
> character for it (we probably should have). AFAIK, Czech language is similar
> with "ch" (it is a single letter too), but I am not sure whether there is an
> example of a real word where "ch" are two letters.
>
>> On 2013-03-12 22:50, Sašo Kiselkov wrote:
>>> On 03/12/2013 10:10 PM, Marcel Telka wrote:
>>>> On Tue, Mar 12, 2013 at 10:02:27PM +0100, Sašo Kiselkov wrote:
>>>>> I'm pretty sure nobody in bash development actually considers
>>>>> locale-specific letter ordering rules. Language-specific idiosyncrasies
>>>>> are a never ending stream of hurt and implementation problems (e.g. in
>>>>> my language "ch" is supposed to be treated as a single letter for
>>>>> sorting purposes).
>>>> Interestingly, "ch" is not always a single letter. It depends on a word:
>>>> "viachlasný" is an example of a word where "ch" are two letters...
>>>>
>>>> Yes, our language is Slovak.
>>> Correct, that's another twisty-twist I forgot to mention. Slovak
>>> sucks... (for computing)
>>>
>>> --
>>> Saso




More information about the OpenIndiana-discuss mailing list