On 2019-05-17 15:26:38 -0400, Tom Lane wrote:
> Robert Haas <robertmhaas@gmail.com> writes:
> > So I think this basically just doesn't work right now. I am
> > sympathetic to Andres's position that we shouldn't go whacking the
> > code around too much at this late date, and he's probably right that
> > we're going to find lots of other problems with tableam as well and
> > you have to draw the line someplace, but on the other hand given your
> > experience and mine, it's probably pretty likely that anybody who
> > tries to use tableam for anything is going to run into this problem,
> > so maybe it's not crazy to think about a few last-minute changes.
>
> It seems to me that the entire tableam project is still very much WIP,
Agreed on that front.
> and if anybody is able to do anything actually useful with a different
> AM right at the moment, that's just mighty good fortune for them.
I think this is too negative. Yes, there's a warts, but you can write
something like zheap without tableam related code modifications (undo
however...). You can write something like zedstore, and it will works,
with a few warts. Yes, a bit of code duplication, and a few efficiency
losses are to be expected. But that's different from it being impossible
to write an AM.
> "I can't do X in an external AM" is not a bug, not for v12 anyway.
Indeed.
Greetings,
Andres Freund