Re: Postgres perl module namespace - Mailing list pgsql-hackers

On 2022-04-18 Mo 13:43, Tom Lane wrote:
> Andrew Dunstan <andrew@dunslane.net> writes:
>> On 2022-04-18 Mo 11:52, Noah Misch wrote:
>>> On Mon, Apr 18, 2022 at 07:15:30AM -0700, Andres Freund wrote:
>>>> I just, again, tried to backport a test as part of a bugfix. The
>>>> renaming between 14 and 15 makes that task almost comically harder. The
>>>> only way I see of dealing with that for the next 5 years is to just
>>>> never backpatch tests to < 15. Which seems like a bad outcome.
>> I'm not sure how often we do things like that. But I don't agree it's
>> impossibly hard, although I can see it might be a bit annoying.
> I think we back-patch test cases *all the time*.  So I think Andres
> is quite right to be concerned about making that harder, although I'm
> not sure that his estimate of the conversion difficulty is accurate.
> I plan to keep a copy of Noah's script and see if applying that to
> the patch files alleviates the pain.  In a few months we should have
> a better idea of whether that's sufficient, or we want to go to the
> work of back-patching the renaming.
>
> I doubt that just plopping the new Cluster.pm in alongside the old
> file could work --- wouldn't the two modules need to share state
> somehow?


No, I think we could probably just port the whole of src/test/PostreSQL
back if required, and have it live alongside the old modules. Each TAP
test is a separate miracle - see comments elsewhere about port
assignment in parallel TAP tests.


But that would mean we have some tests in the old flavor and some in the
new flavor in the back branches, which might get confusing.


>
> Another thing that ought to be on the table is back-patching
> 549ec201d (Replace Test::More plans with done_testing).  Those
> test counts are an even huger pain for back-patching than the
> renaming, because the count is often different in each branch.
>
>             


+1 for doing that


cheers


andrew

--
Andrew Dunstan
EDB: https://www.enterprisedb.com