Re: Row security violation error is misleading

Dean,

* Dean Rasheed (dean(dot)a(dot)rasheed(at)gmail(dot)com) wrote:
> On 7 April 2015 at 16:21, Stephen Frost <sfrost(at)snowman(dot)net> wrote:
> > Agreed and we actually have a patch from Dean already to address this,
> > it's just been waiting on me (with a couple of other ones). It'd
> > certainly be great if you have time to take a look at those, though,
> > generally speaking, I feel prety happy about where those are and believe
> > they really just need to be reviewed/tested and maybe a bit of word-
> > smithing around the docs.
>
> The first of those patches [1] has bit-rotted somewhat, due to the
> recent changes to the way rowmarks are handled, so here's an updated
> version. It wasn't a trivial merge, because commit
> cb1ca4d800621dcae67ca6c799006de99fa4f0a5 made a change to
> ExecBuildAuxRowMark() without a matching change to
> preprocess_targetlist(), and one of the new RLS-with-inheritance tests
> fell over that.

Thanks a lot for this. Please take a look at the attached. It still
includes the preprocess_targetlist() changes, so the regression tests
don't fail, but that'll be committed independently (hopefully soon).

I've taken an initial look at the second patch (actually, a few times)
and plan to do a thorough review soon. I'd definitely like to get these
both committed and done with very shortly. Again, apologies about the
delays; this past weekend was quite a bit busier than I originally
anticipated.

Thanks!

Stephen