> Yeah, I think you could do that, I agree it feels better that way.
> You'll still need new copytup and comparetup functions, though, to deal
> with HeapTupleHeaders instead of MinimalTuples, or modify the existing
> ones to handle both.
You meant HeapTuple, not HeapTupleHeaders, right?
Mmh, didn't think of those two functions; I might as well start with Gregory
Stark's patch (that is: using HeapTuple)
> And some way to indicate that you want to preserve
> the visibility information when you create the tuplesort, maybe a new
> parameter to tuplesort_begin_heap().
I guess that using Gregory Stark's patch there's no need for it, since it uses
A patch that:
1) uses always the old CLUSTER method for non-btree indexes and for
2) add a whole set of new functions to tuplesort (as in Gregory Stark's patch)
would be rejected "for sure"? Or can be thought as a "better than nothing,
works in 90% cases" patch?
In response to
pgsql-hackers by date
|Next:||From: Greg Stark||Date: 2010-01-15 16:01:25|
|Subject: Re: New XLOG record indicating WAL-skipping|
|Previous:||From: Kevin Grittner||Date: 2010-01-15 15:47:28|
|Subject: Re: Testing with concurrent sessions|