Re: ispell/hunspell imprecision in error message

Поиск
Список
Период
Сортировка
От Tom Lane
Тема Re: ispell/hunspell imprecision in error message
Дата
Msg-id 6356.1469542023@sss.pgh.pa.us
обсуждение исходный текст
Ответ на Re: ispell/hunspell imprecision in error message  (Artur Zakirov <a.zakirov@postgrespro.ru>)
Список pgsql-hackers
Artur Zakirov <a.zakirov@postgrespro.ru> writes:
> On 25.07.2016 19:53, Peter Eisentraut wrote:
>> But I found that this actually talks about Hunspell format dictionaries.
>> (So the man page is hunspell(5) as opposed to ispell(5).)  While as far
>> as the tsearch interfaces are concerned, those two are lumped together,
>> should we not change the error message to refer to Hunspell?

> As I understand, this error message refers to the Ispell dictionary 
> template. This template can use various dictionary formats. Which is 
> confusing. Maybe would be better to change dictionary template name. But 
> it can break backward compatibility...

I think the error message has to refer to the ispell dictionary type by
its assigned name, otherwise you'll just create even more confusion.
(In other words, I read the message as talking about PG's ispell
dictionary code, not about somebody else's ispell man file.)

The right place to explain that ispell accepts hunspell-formatted files,
or to link to specifications of what that means, is in the documentation.
Section 12.6.5 already says that, but maybe it could be expanded/rewritten
a little.
        regards, tom lane



В списке pgsql-hackers по дате отправления:

Предыдущее
От: Chapman Flack
Дата:
Сообщение: Re: AdvanceXLInsertBuffer vs. WAL segment compressibility
Следующее
От: Tom Lane
Дата:
Сообщение: Re: Proposal: revert behavior of IS NULL on row types