"Jonah H. Harris" <[EMAIL PROTECTED]> writes:
> On 6/22/06, Tom Lane <[EMAIL PROTECTED]> wrote:
>> Yes it would. The most obvious point is that memory management and
>> error handling conventions inside the backend are quite different from
>> what you'd expect to employ in a standalone program.
> No, this wouldn't really be that hard, especially if he created a few
> macros to handle the differences.
Jonah, I've been working with this system for years, and it's not that
easy to "handle the differences with a few macros". I've seen people
try, repeatedly, and seen their code break repeatedly. The relatively
small number of files that we use in both frontend and backend scenarios
are all extremely limited-in-function and tend to break easily. I would
never try to build something nontrivial that would work both ways
... especially not on a very tight time budget. Diogo will have enough
challenges delivering something useful that works in one environment.
> The goal
> was to start with xlogdump and enhance it, so starting with
> command-line is probably the best anyway.
Diogo, are you working from my old xlogdump hack? If so what version?
I can send you the latest off-list. I add stuff to it periodically when
I need it, and I don't think I've published it lately.
regards, tom lane
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
