Re: longfin and tamandua aren't too happy but I'm not sure why

On Wed, Sep 28, 2022 at 1:15 AM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Dilip Kumar <dilipbalaut(at)gmail(dot)com> writes:
> > wrasse is also failing with a bus error,
>
> Yeah. At this point I think it's time to call for this patch
> to get reverted. It should get tested *off line* on some
> non-Intel, non-64-bit, alignment-picky architectures before
> the rest of us have to deal with it any more.

I don't really understand how you expect me or Dilip to do this. Is
every PostgreSQL hacker supposed to have a bunch of antiquated servers
in their basement so that they can test this stuff? I don't think I
have had easy access to non-Intel, non-64-bit, alignment-picky
hardware in probably 25 years, unless my old Raspberry Pi counts.

I admit that I should have checked the CI results before pushing this
commit, but as you say yourself, that missed a bunch of stuff, and I'd
say it was the important stuff. Unless and until CI is able to check
all the same configurations that the buildfarm can check, it's not
going to be possible to get test results on some of these platforms
except by checking the code in and seeing what happens. If I revert
this, I'm just going to be sitting here not knowing where any of the
problems are and having no way to find them.

Maybe I'm missing something here. Apart from visual inspection of the
code and missing fewer mistakes than I did, how would you have avoided
these problems in one of your commits?

--
Robert Haas
EDB: http://www.enterprisedb.com