Peter Eisentraut <peter_e(at)gmx(dot)net> writes:
> This is an interesting idea, although you may want to give the user the
> option to customize his prompt.
Seems cool. I am a bit worried about whether the transaction-block
detection mechanism is reliable, though. We might need to add something
to the FE/BE protocol to make this work correctly.
>> Secondly, it adds a "begin transaction" option that, when
>> enabled, ensures that you are always inside a transaction
>> while in psql, so you can always rollback.
> This should be done in the backend.
Agreed. If I recall recent discussions correctly, the spec says that
certain SQL commands should open a transaction and others should not.
It's not reasonable to have that logic in psql rather than the backend.
regards, tom lane
In response to
pgsql-hackers by date
|Next:||From: Hannu Krosing||Date: 2002-02-07 18:40:53|
|Subject: Re: Threaded PosgreSQL server|
|Previous:||From: Tom Lane||Date: 2002-02-07 17:58:50|
|Subject: Re: Suggestions for 7.3 date handling |
pgsql-patches by date
|Next:||From: Tom Lane||Date: 2002-02-07 19:05:35|
|Subject: Re: contrib/pg_filedump - PostgreSQL File Dump Utility |
|Previous:||From: Patrick Macdonald||Date: 2002-02-07 14:59:17|
|Subject: Re: contrib/pg_filedump - PostgreSQL File Dump Utility|