Re: trouble caused by change in 7.3 handling of '' in
От | Doug McNaught |
---|---|
Тема | Re: trouble caused by change in 7.3 handling of '' in |
Дата | |
Msg-id | m3el8eqblq.fsf@varsoon.wireboard.com обсуждение исходный текст |
Ответ на | Re: trouble caused by change in 7.3 handling of '' in (Bruce Momjian <pgman@candle.pha.pa.us>) |
Список | pgsql-general |
Vivek Khera <khera@kcilink.com> writes: > It is not hard. What is hard is when you change behavior abruptly > between releases without a transition period in which you warn about > an upcoming change when that deprecated feature is used. Ideally, it > would have logged this error for the 7.3 release, and disallowed it > for 7.4, something akin to the LIMIT transition. > > I guess nobody realized how some popular applications used this > non-compliant feature before the change was implemented. Perhaps we > learn from this experience and don't do this type of change again. > Perhaps not. Yeah, I can personally see a case for deprecating nonstandard features and spitting out a warning, rather than dropping them abruptly. The deprecation would last for one major release cycle. On the other hand, if the nonstandard wart/feature is standing in the way of implementing a high-quality and desired new feature (not saying this did or didn't happen in this case; I don't know), is it worth putting off adding the new feature rather than just excising the wart and putting up with the pain? -Doug
В списке pgsql-general по дате отправления: