Re: SQLState
От | Neil Conway |
---|---|
Тема | Re: SQLState |
Дата | |
Msg-id | 1084406840.19208.61.camel@tokyo обсуждение исходный текст |
Ответ на | Re: SQLState ("Mahesh Swamy" <maheshs@fast.fujitsu.com.au>) |
Список | pgsql-patches |
On Wed, 2004-05-12 at 00:55, Mahesh Swamy wrote: > I think that I have missed explaining a point. We are developing a > proprietary extension to PostgreSQL. This means that some errors that are > thrown in our extension will never be thrown by the OSS community code. > Thus, for these cases we should use a code that will not be used by the OSS > community (hopefully) so we will not have to keep changing it in future > updates. I hope i am clear? Thus, we invented using the letter 'x', as it > seems a bit out of the way... Ah, right -- makes sense. I don't think a (lower-case) 'x' is the best choice, though; my copy of SQL2003 Section 23.1 sayeth: Class values that begin with one of the <digit>s 0 , 1 , 2 , 3 , or 4 or one of the <simple Latin upper case letter>s A , B , C , D , E , F , G , or H are returned only for conditions defined in ISO/IEC 9075 or in any other International Standard. [ ... ] Subclass values associated with such classes that also begin with one of those 13 characters are returned only for conditions defined in ISO/IEC 9075 or some other International Standard. [... ] Subclass values associated with such classes that begin with one of the <digit>s 5 , 6 , 7 , 8 , or 9 or one of the <simple Latin upper case letter>s I , J , K , L , M , N , O , P , Q , R , S , T , U , V , W , X , Y , or Z are reserved for implementation-specified conditions and are called implementation-defined subclasses. So an upper-case 'X' or other suitable letter (for those errors thrown _only_ by the proprietary extension) seems best, as the standard explicitly says that such a subclass value is legal for implementation-defined conditions. -Neil
В списке pgsql-patches по дате отправления: