Re: proposal: fix corner use case of variadic fuctions usage

On Sat, Dec 1, 2012 at 1:14 PM, Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>wrote:

> Hello
>
> >
> > Hi Pavel.
> >
> > I am trying to review this patch and on my work computer everything
> compiles
> > and tests perfectly. However, on my laptop, the regression tests don't
> pass
> > with "cache lookup failed for type XYZ" where XYZ is some number that
> does
> > not appear to be any type oid.
> >
> > I don't really know where to go from here. I am asking that other people
> try
> > this patch to see if they get errors as well.
> >
>
> yes, I checked it on .x86_64 and I had a same problems
>
> probably there was more than one issue - I had to fix a creating a
> unpacked params and I had a issue with gcc optimalization when I used
> a stack variable for fcinfo.
>
> Now I fixed these issues and I hope so it will work on all platforms
>

It appears to work a lot better, yes. I played around with it a little bit
and wasn't able to break it, so I'm marking it as ready for committer.
Some wordsmithing will need to be done on the code comments.