On Thu, Mar 05, 2020 at 12:53:54PM +0900, Michael Paquier wrote:
> On Wed, Mar 04, 2020 at 09:21:45AM +0100, Julien Rouhaud wrote:
>
> > Should we add some regression
> > tests for that? I guess most of it could be borrowed from the patch
> > to fix the toast index issue I sent last week.
>
> I have doubts when it comes to use a strategy based on
> pg_cancel_backend() and a match of application_name (see for example
> 5ad72ce but I cannot find the associated thread). I think that we
> could design something more robust here and usable by all tests, with
> two things coming into my mind:
> - A new meta-command for isolation tests to be able to cancel a
> session with PQcancel().
> - Fault injection in the backend.
> For the case of this thread, the cancellation command would be a better
> match.
Here's a patch to add an optional "timeout val" clause to isolationtester's
step definition. When used, isolationtester will actively wait on the query
rather than continuing with the permutation next step, and will issue a cancel
once the defined timeout is reached. I also added as a POC the previous
regression tests for invalid TOAST indexes, updated to use this new
infrastructure (which won't pass as long as the original bug for invalid TOAST
indexes isn't fixed).
I'll park that in the next commitfest, with a v14 target version.
