Horton

Thread

Commits

  1. Fix missing installation/uninstallation rules for BackgroundPsql.pm

  2. Backport BackgroundPsql perl test module

  3. Add missing uninstallation rule for BackgroundPsql.pm

  4. Fix missing installation rules for BackgroundPsql.pm

  5. Skip \password TAP test on old IPC::Run versions

  6. Test SCRAM iteration changes with psql \password

  7. Refactor background psql TAP functions

  1. Making background psql nicer to use in tap tests

    Andres Freund <andres@anarazel.de> — 2023-01-30T19:43:50Z

    Hi,

Plenty tap tests require a background psql. But they're pretty annoying to
use.

I think the biggest improvement would be an easy way to run a single query and
get the result of that query. Manually having to pump_until() is awkward and
often leads to hangs/timeouts, instead of test failures, because one needs to
specify a match pattern to pump_until(), which on mismatch leads to trying to
keep pumping forever.

It's annoyingly hard to wait for the result of a query in a generic way with
background_psql(), and more generally for psql. background_psql() uses -XAtq,
which means that we'll not get "status" output (like "BEGIN" or "(1 row)"),
and that queries not returning anything are completely invisible.

A second annoyance is that issuing a query requires a trailing newline,
otherwise psql won't process it.


The best way I can see is to have a helper that issues the query, followed by
a trailing newline, an \echo with a recognizable separator, and then uses
pump_until() to wait for that separator.


Another area worthy of improvement is that background_psql() requires passing
in many variables externally - without a recognizable benefit afaict. What's
the point in 'stdin', 'stdout', 'timer' being passed in? stdin/stdout need to
point to empty strings, so we know what's needed - in fact we'll even reset
them if they're passed in.  The timer is always going to be
PostgreSQL::Test::Utils::timeout_default, so again, what's the point?

I think it'd be far more usable if we made background_psql() return a hash
with the relevant variables. The 031_recovery_conflict.pl test has:

my $psql_timeout = IPC::Run::timer($PostgreSQL::Test::Utils::timeout_default);
my %psql_standby = ('stdin' => '', 'stdout' => '');
$psql_standby{run} =
  $node_standby->background_psql($test_db, \$psql_standby{stdin},
        \$psql_standby{stdout},
        $psql_timeout);
$psql_standby{stdout} = '';

How about just returning a reference to a hash like that? Except that I'd also
make stderr available, which one can't currently access.


The $psql_standby{stdout} = ''; is needed because background_psql() leaves a
banner in the output, which it shouldn't, but we probably should just fix
that.


Brought to you by: Trying to write a test for vacuum_defer_cleanup_age.

- Andres

  2. Re: Making background psql nicer to use in tap tests

    Tom Lane <tgl@sss.pgh.pa.us> — 2023-01-30T20:06:46Z

    Andres Freund <andres@anarazel.de> writes:
> It's annoyingly hard to wait for the result of a query in a generic way with
> background_psql(), and more generally for psql. background_psql() uses -XAtq,
> which means that we'll not get "status" output (like "BEGIN" or "(1 row)"),
> and that queries not returning anything are completely invisible.

Yeah, the empty-query-result problem was giving me fits recently.
+1 for wrapping this into something more convenient to use.

			regards, tom lane

  3. Re: Making background psql nicer to use in tap tests

    Andres Freund <andres@anarazel.de> — 2023-01-31T00:00:47Z

    Hi,

On 2023-01-30 15:06:46 -0500, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > It's annoyingly hard to wait for the result of a query in a generic way with
> > background_psql(), and more generally for psql. background_psql() uses -XAtq,
> > which means that we'll not get "status" output (like "BEGIN" or "(1 row)"),
> > and that queries not returning anything are completely invisible.
>
> Yeah, the empty-query-result problem was giving me fits recently.
> +1 for wrapping this into something more convenient to use.

I've hacked some on this. I first tried to just introduce a few helper
functions in Cluster.pm, but that ended up being awkward. So I bit the bullet
and introduced a new class (in BackgroundPsql.pm), and made background_psql()
and interactive_psql() return an instance of it.

This is just a rough prototype. Several function names don't seem great, it
need POD documentation, etc.


The main convenience things it has over the old interface:
- $node->background_psql('dbname') is enough
- $psql->query(), which returns the query results as a string, is a lot easier
  to use than having to pump, identify query boundaries via regex etc.
- $psql->query_safe(), which dies if any query fails (detected via stderr)
- $psql->query_until() is a helper that makes it a bit easier to start queries
  that won't finish until a later point


I don't quite like the new interface yet:
- It's somewhat common to want to know if there was a failure, but also get
  the query result, not sure what the best function signature for that is in
  perl.
- query_until() sounds a bit too much like $node->poll_query_until(). Maybe
  query_wait_until() is better? OTOH, the other function has poll in the name,
  so maybe it's ok.
- right now there's a bit too much logic in background_psql() /
  interactive_psql() for my taste


Those points aside, I think it already makes the tests a good bit more
readable. My WIP vacuum_defer_cleanup_age patch shrunk by half with it.

I think with a bit more polish it's easy enough to use that we could avoid a
good number of those one-off psql's that we do all over.


I didn't really know what this, insrc/test/subscription/t/015_stream.pl, is
about:

$h->finish;    # errors make the next test fail, so ignore them here

There's no further test?

I'm somewhat surprised it doesn't cause problems in another ->finish later on,
where we then afterwards just use $h again. Apparently IPC::Run just
automagically restarts psql?


Greetings,

Andres Freund

  4. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-03-14T20:24:32Z

    > On 31 Jan 2023, at 01:00, Andres Freund <andres@anarazel.de> wrote:

> I've hacked some on this. I first tried to just introduce a few helper
> functions in Cluster.pm, but that ended up being awkward. So I bit the bullet
> and introduced a new class (in BackgroundPsql.pm), and made background_psql()
> and interactive_psql() return an instance of it.

Thanks for working on this!

> This is just a rough prototype. Several function names don't seem great, it
> need POD documentation, etc.

It might be rough around the edges but I don't think it's too far off a state
in which in can be committed, given that it's replacing something even rougher.
With documentation and some polish I think we can iterate on it in the tree.

I've played around a lot with it and it seems fairly robust.

> I don't quite like the new interface yet:
> - It's somewhat common to want to know if there was a failure, but also get
>  the query result, not sure what the best function signature for that is in
>  perl.

What if query() returns a list with the return value last?  The caller will get
the return value when assigning a single var as the return, and can get both in
those cases when it's interesting.  That would make for reasonably readable
code in most places?

   $ret_val = $h->query("SELECT 1;");
   ($query_result, $ret_val) = $h->query("SELECT 1;"); 

Returning a hash seems like a worse option since it will complicate callsites
which only want to know success/failure.

> - query_until() sounds a bit too much like $node->poll_query_until(). Maybe
>  query_wait_until() is better? OTOH, the other function has poll in the name,
>  so maybe it's ok.

query_until isn't great but query_wait_until is IMO worse since the function
may well be used for tests which aren't using longrunning waits.  It's also
very useful for things which aren't queries at all, like psql backslash
commands.  I don't have any better ideas though, so +1 for sticking with
query_until.

> - right now there's a bit too much logic in background_psql() /
>  interactive_psql() for my taste

Not sure what you mean, I don't think they're especially heavy on logic?

> Those points aside, I think it already makes the tests a good bit more
> readable. My WIP vacuum_defer_cleanup_age patch shrunk by half with it.

The test for \password in the SCRAM iteration count patch shrunk to 1/3 of the
previous coding.

> I think with a bit more polish it's easy enough to use that we could avoid a
> good number of those one-off psql's that we do all over.

Agreed, and ideally implement tests which were left unwritten due to the old
API being clunky.

+       # feed the query to psql's stdin, follwed by \n (so psql processes the

s/follwed/followed/

+A default timeout of $PostgreSQL::Test::Utils::timeout_default is set up,
+which can modified later.

This require a bit of knowledge about the internals which I think we should
hide in this new API.  How about providing a function for defining the timeout?

Re timeouts: one thing I've done repeatedly is to use short timeouts and reset
them per query, and that turns pretty ugly fast.  I hacked up your patch to
provide $h->reset_timer_before_query() which then injects a {timeout}->start
before running each query without the caller having to do it.  Not sure if I'm
alone in doing that but if not I think it makes sense to add.

--
Daniel Gustafsson

  5. Re: Making background psql nicer to use in tap tests

    Andres Freund <andres@anarazel.de> — 2023-03-15T01:03:21Z

    Hi,

On 2023-03-14 21:24:32 +0100, Daniel Gustafsson wrote:
> > On 31 Jan 2023, at 01:00, Andres Freund <andres@anarazel.de> wrote:
> 
> > I've hacked some on this. I first tried to just introduce a few helper
> > functions in Cluster.pm, but that ended up being awkward. So I bit the bullet
> > and introduced a new class (in BackgroundPsql.pm), and made background_psql()
> > and interactive_psql() return an instance of it.
> 
> Thanks for working on this!

Thanks for helping it move along :)


> > This is just a rough prototype. Several function names don't seem great, it
> > need POD documentation, etc.
> 
> It might be rough around the edges but I don't think it's too far off a state
> in which in can be committed, given that it's replacing something even rougher.
> With documentation and some polish I think we can iterate on it in the tree.

Cool.


> > I don't quite like the new interface yet:
> > - It's somewhat common to want to know if there was a failure, but also get
> >  the query result, not sure what the best function signature for that is in
> >  perl.
> 
> What if query() returns a list with the return value last?  The caller will get
> the return value when assigning a single var as the return, and can get both in
> those cases when it's interesting.  That would make for reasonably readable
> code in most places?

>    $ret_val = $h->query("SELECT 1;");
>    ($query_result, $ret_val) = $h->query("SELECT 1;");

I hate perl.


> Returning a hash seems like a worse option since it will complicate callsites
> which only want to know success/failure.

Yea. Perhaps it's worth having a separate function for this? ->query_rc() or such?



> > - right now there's a bit too much logic in background_psql() /
> >  interactive_psql() for my taste
> 
> Not sure what you mean, I don't think they're especially heavy on logic?

-EMISSINGWORD on my part. A bit too much duplicated logic.


> +A default timeout of $PostgreSQL::Test::Utils::timeout_default is set up,
> +which can modified later.
> 
> This require a bit of knowledge about the internals which I think we should
> hide in this new API.  How about providing a function for defining the timeout?

"definining" in the sense of accessing it? Or passing one in?


> Re timeouts: one thing I've done repeatedly is to use short timeouts and reset
> them per query, and that turns pretty ugly fast.  I hacked up your patch to
> provide $h->reset_timer_before_query() which then injects a {timeout}->start
> before running each query without the caller having to do it.  Not sure if I'm
> alone in doing that but if not I think it makes sense to add.

I don't quite understand the use case, but I don't mind it as a functionality.

Greetings,

Andres Freund

  6. Re: Making background psql nicer to use in tap tests

    Andrew Dunstan <andrew@dunslane.net> — 2023-03-15T14:10:20Z

    On 2023-01-30 Mo 19:00, Andres Freund wrote:
> Hi,
>
> On 2023-01-30 15:06:46 -0500, Tom Lane wrote:
>> Andres Freund<andres@anarazel.de>  writes:
>>> It's annoyingly hard to wait for the result of a query in a generic way with
>>> background_psql(), and more generally for psql. background_psql() uses -XAtq,
>>> which means that we'll not get "status" output (like "BEGIN" or "(1 row)"),
>>> and that queries not returning anything are completely invisible.
>> Yeah, the empty-query-result problem was giving me fits recently.
>> +1 for wrapping this into something more convenient to use.
> I've hacked some on this. I first tried to just introduce a few helper
> functions in Cluster.pm, but that ended up being awkward. So I bit the bullet
> and introduced a new class (in BackgroundPsql.pm), and made background_psql()
> and interactive_psql() return an instance of it.
>
> This is just a rough prototype. Several function names don't seem great, it
> need POD documentation, etc.


Since this class is only intended to have instances created from 
Cluster, I would be inclined just to put it at the end of Cluster.pm 
instead of creating a new file. That makes it clearer that the new 
package is not standalone. We already have instances of that.

The first param of the constructor is a bit opaque. If it were going to 
be called from elsewhere I'd want something a bit more obvious, but I 
guess we can live with it here. An alternative might be 
multiple_constructors (e.g. new_background, new_interactive) which use a 
common private routine.

Don't have comments yet on the other things, will continue looking.


cheers


andrew

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

  7. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-03-17T09:48:09Z

    > On 15 Mar 2023, at 02:03, Andres Freund <andres@anarazel.de> wrote:

>> Returning a hash seems like a worse option since it will complicate callsites
>> which only want to know success/failure.
> 
> Yea. Perhaps it's worth having a separate function for this? ->query_rc() or such?

If we are returning a hash then I agree it should be a separate function.
Maybe Andrew has input on which is the most Perl way of doing this.

>>> - right now there's a bit too much logic in background_psql() /
>>> interactive_psql() for my taste
>> 
>> Not sure what you mean, I don't think they're especially heavy on logic?
> 
> -EMISSINGWORD on my part. A bit too much duplicated logic.

That makes more sense, and I can kind of agree.  I don't think it's too bad but
I agree there is room for improvement.

>> +A default timeout of $PostgreSQL::Test::Utils::timeout_default is set up,
>> +which can modified later.
>> 
>> This require a bit of knowledge about the internals which I think we should
>> hide in this new API.  How about providing a function for defining the timeout?
> 
> "definining" in the sense of accessing it? Or passing one in?

I meant passing one in.

>> Re timeouts: one thing I've done repeatedly is to use short timeouts and reset
>> them per query, and that turns pretty ugly fast.  I hacked up your patch to
>> provide $h->reset_timer_before_query() which then injects a {timeout}->start
>> before running each query without the caller having to do it.  Not sure if I'm
>> alone in doing that but if not I think it makes sense to add.
> 
> I don't quite understand the use case, but I don't mind it as a functionality.

I've used it a lot when I want to run n command which each should finish
quickly or not at all.  So one time budget per command rather than having a
longer timeout for a set of commands that comprise a test.  It can be done
already today by calling ->start but it doesn't exactly make the code cleaner.

As mentioned off-list I did some small POD additions when reviewing, so I've
attached them here in a v2 in the hopes that it might be helpful.  I've also
included the above POC for restarting the timeout per query to show what I
meant.

--
Daniel Gustafsson

  8. Re: Making background psql nicer to use in tap tests

    Andrew Dunstan <andrew@dunslane.net> — 2023-03-17T13:48:31Z

    On 2023-03-17 Fr 05:48, Daniel Gustafsson wrote:
>> On 15 Mar 2023, at 02:03, Andres Freund<andres@anarazel.de>  wrote:
>>> Returning a hash seems like a worse option since it will complicate callsites
>>> which only want to know success/failure.
>> Yea. Perhaps it's worth having a separate function for this? ->query_rc() or such?
> If we are returning a hash then I agree it should be a separate function.
> Maybe Andrew has input on which is the most Perl way of doing this.


I think the perlish way is use the `wantarray` function. Perl knows if 
you're expecting a scalar return value or a list (which includes a hash).


    return wantarray ? $retval : (list or hash);


A few more issues:

A common perl idiom is to start private routine names with an 
underscore. so I'd rename wait_connect to _wait_connect;

Why is $restart_before_query a package/class level value instead of an 
instance value? And why can we only ever set it to 1 but not back again? 
Maybe we don't want to, but it looks odd.

If we are going to keep this as a separate package, then we should put 
some code in the constructor to prevent it being called from elsewhere 
than the Cluster package. e.g.

     # this constructor should only be called from PostgreSQL::Test::Cluster
     my ($package, $file, $line) = caller;

     die "Forbidden caller of constructor: package: $package, file: 
$file:$line"
       unless $package eq 'PostgreSQL::Test::Cluster';

This should refer to the full class name:

+=item $node->background_psql($dbname, %params) => BackgroundPsql instance


Still reviewing ...


cheers


andrew

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

  9. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-03-17T14:08:11Z

    > On 17 Mar 2023, at 14:48, Andrew Dunstan <andrew@dunslane.net> wrote:
> On 2023-03-17 Fr 05:48, Daniel Gustafsson wrote:
>>> On 15 Mar 2023, at 02:03, Andres Freund <andres@anarazel.de>
>>>  wrote:
>>> 
>>>> Returning a hash seems like a worse option since it will complicate callsites
>>>> which only want to know success/failure.
>>>> 
>>> Yea. Perhaps it's worth having a separate function for this? ->query_rc() or such?
>>> 
>> If we are returning a hash then I agree it should be a separate function.
>> Maybe Andrew has input on which is the most Perl way of doing this.
> 
> I think the perlish way is use the `wantarray` function. Perl knows if you're expecting a scalar return value or a list (which includes a hash).
> 
>    return wantarray ? $retval : (list or hash);

Aha, TIL. That seems like precisely what we want. 

> A common perl idiom is to start private routine names with an underscore. so I'd rename wait_connect to _wait_connect;

There are quite a few routines documented as internal in Cluster.pm which don't
start with an underscore.  Should we change them as well?  I'm happy to prepare
a separate patch to address that if we want that.

> Why is $restart_before_query a package/class level value instead of an instance value? And why can we only ever set it to 1 but not back again? Maybe we don't want to, but it looks odd.

It was mostly a POC to show what I meant with the functionality.  I think there
should be a way to turn it off (set it to zero) even though I doubt it will be
used much.

> If we are going to keep this as a separate package, then we should put some code in the constructor to prevent it being called from elsewhere than the Cluster package. e.g.
> 
>     # this constructor should only be called from PostgreSQL::Test::Cluster
>     my ($package, $file, $line) = caller;
>     
>     die "Forbidden caller of constructor: package: $package, file: $file:$line"
>       unless $package eq 'PostgreSQL::Test::Cluster';

I don't have strong feelings about where to place this, but Cluster.pm is
already quite long so I see a small upside to keeping it separate to not make
that worse.

--
Daniel Gustafsson

  10. Re: Making background psql nicer to use in tap tests

    Andrew Dunstan <andrew@dunslane.net> — 2023-03-17T16:25:14Z

    On 2023-03-17 Fr 10:08, Daniel Gustafsson wrote:
>
>> A common perl idiom is to start private routine names with an underscore. so I'd rename wait_connect to _wait_connect;
> There are quite a few routines documented as internal in Cluster.pm which don't
> start with an underscore.  Should we change them as well?  I'm happy to prepare
> a separate patch to address that if we want that.


Possibly. There are two concerns. First, make sure that they really are 
private. Last time I looked I think I noticed at least one thing that 
was alleged to be private but was called from a TAP script. Second, 
unless we backpatch it there will be some drift between branches, which 
can make backpatching things a bit harder. But by all means prep a patch 
so we can see the scope of the issue.


>
>> Why is $restart_before_query a package/class level value instead of an instance value? And why can we only ever set it to 1 but not back again? Maybe we don't want to, but it looks odd.
> It was mostly a POC to show what I meant with the functionality.  I think there
> should be a way to turn it off (set it to zero) even though I doubt it will be
> used much.


A common idiom is to have a composite getter/setter method for object 
properties something like this


    sub settingname

    {

       my ($self, $arg) = @_;

       $self->{settingname} = $arg if defined $arg;

       return $self->{settingname};

    }


>
>> If we are going to keep this as a separate package, then we should put some code in the constructor to prevent it being called from elsewhere than the Cluster package. e.g.
>>
>>      # this constructor should only be called from PostgreSQL::Test::Cluster
>>      my ($package, $file, $line) = caller;
>>      
>>      die "Forbidden caller of constructor: package: $package, file: $file:$line"
>>        unless $package eq 'PostgreSQL::Test::Cluster';
> I don't have strong feelings about where to place this, but Cluster.pm is
> already quite long so I see a small upside to keeping it separate to not make
> that worse.
>

Yeah, I can go along with that.


cheers


andrew

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

  11. Re: Making background psql nicer to use in tap tests

    Dagfinn Ilmari Mannsåker <ilmari@ilmari.org> — 2023-03-17T18:07:40Z

    Andrew Dunstan <andrew@dunslane.net> writes:

> On 2023-03-17 Fr 10:08, Daniel Gustafsson wrote:
>>> Why is $restart_before_query a package/class level value instead of
>>> an instance value? And why can we only ever set it to 1 but not back
>>> again? Maybe we don't want to, but it looks odd.
>> It was mostly a POC to show what I meant with the functionality.  I think there
>> should be a way to turn it off (set it to zero) even though I doubt it will be
>> used much.
>
>
> A common idiom is to have a composite getter/setter method for object
> properties something like this
>
>
>    sub settingname
>    {
>       my ($self, $arg) = @_;
>       $self->{settingname} = $arg if defined $arg;
>       return $self->{settingname};
>    }

Or, if undef is a valid value:


    sub settingname
    {
        my $self = shift;
        $self->[settingname} = shift if @_;
        return $self->{settingname};
    }

- ilmari

  12. Re: Making background psql nicer to use in tap tests

    Andrew Dunstan <andrew@dunslane.net> — 2023-03-17T22:12:58Z

    On 2023-03-17 Fr 14:07, Dagfinn Ilmari Mannsåker wrote:
> Andrew Dunstan<andrew@dunslane.net>  writes:
>
>> On 2023-03-17 Fr 10:08, Daniel Gustafsson wrote:
>>>> Why is $restart_before_query a package/class level value instead of
>>>> an instance value? And why can we only ever set it to 1 but not back
>>>> again? Maybe we don't want to, but it looks odd.
>>> It was mostly a POC to show what I meant with the functionality.  I think there
>>> should be a way to turn it off (set it to zero) even though I doubt it will be
>>> used much.
>>
>> A common idiom is to have a composite getter/setter method for object
>> properties something like this
>>
>>
>>     sub settingname
>>     {
>>        my ($self, $arg) = @_;
>>        $self->{settingname} = $arg if defined $arg;
>>        return $self->{settingname};
>>     }
> Or, if undef is a valid value:
>
>
>      sub settingname
>      {
>          my $self = shift;
>          $self->[settingname} = shift if @_;
>          return $self->{settingname};
>      }
>


Yes, I agree that's better (modulo the bracket typo)


cheers


andrew

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

  13. Re: Making background psql nicer to use in tap tests

    Andres Freund <andres@anarazel.de> — 2023-03-17T22:58:55Z

    Hi,

On 2023-03-17 12:25:14 -0400, Andrew Dunstan wrote:
> On 2023-03-17 Fr 10:08, Daniel Gustafsson wrote:
> > > If we are going to keep this as a separate package, then we should put some code in the constructor to prevent it being called from elsewhere than the Cluster package. e.g.
> > > 
> > >      # this constructor should only be called from PostgreSQL::Test::Cluster
> > >      my ($package, $file, $line) = caller;
> > >      die "Forbidden caller of constructor: package: $package, file: $file:$line"
> > >        unless $package eq 'PostgreSQL::Test::Cluster';
> > I don't have strong feelings about where to place this, but Cluster.pm is
> > already quite long so I see a small upside to keeping it separate to not make
> > that worse.
> > 
> 
> Yeah, I can go along with that.

Cool - I'd prefer a separate file. I do find Cluster.pm somewhat unwieldy at
this point, and I susect that we'll end up with additional helpers around
BackgroundPsql.

Greetings,

Andres Freund

  14. Re: Making background psql nicer to use in tap tests

    Andrew Dunstan <andrew@dunslane.net> — 2023-03-18T22:07:46Z

    On 2023-03-17 Fr 18:58, Andres Freund wrote:
> Hi,
>
> On 2023-03-17 12:25:14 -0400, Andrew Dunstan wrote:
>> On 2023-03-17 Fr 10:08, Daniel Gustafsson wrote:
>>>> If we are going to keep this as a separate package, then we should put some code in the constructor to prevent it being called from elsewhere than the Cluster package. e.g.
>>>>
>>>>       # this constructor should only be called from PostgreSQL::Test::Cluster
>>>>       my ($package, $file, $line) = caller;
>>>>       die "Forbidden caller of constructor: package: $package, file: $file:$line"
>>>>         unless $package eq 'PostgreSQL::Test::Cluster';
>>> I don't have strong feelings about where to place this, but Cluster.pm is
>>> already quite long so I see a small upside to keeping it separate to not make
>>> that worse.
>>>
>> Yeah, I can go along with that.
> Cool - I'd prefer a separate file. I do find Cluster.pm somewhat unwieldy at
> this point, and I susect that we'll end up with additional helpers around
> BackgroundPsql.
>

Yeah. BTW, a better test than the one above would be


    $package->isa("PostgreSQL::Test::Cluster")


cheers


andrew


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

  15. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-03-23T22:36:22Z

    > On 18 Mar 2023, at 23:07, Andrew Dunstan <andrew@dunslane.net> wrote:

> BTW, a better test than the one above would be
> 
>    $package->isa("PostgreSQL::Test::Cluster")

Attached is a quick updated v3 of the patch which, to the best of my Perl
abilities, tries to address the comments raised here.

--
Daniel Gustafsson

  16. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-03-31T20:33:23Z

    The attached v4 fixes some incorrect documentation (added by me in v3), and
fixes that background_psql() didn't honor on_error_stop and extraparams passed
by the user.  I've also added a commit which implements the \password test from
the SCRAM iteration count patchset as well as cleaned up a few IPC::Run
includes from test scripts.

--
Daniel Gustafsson

  17. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-04-02T20:24:16Z

    > On 31 Mar 2023, at 22:33, Daniel Gustafsson <daniel@yesql.se> wrote:
> 
> The attached v4 fixes some incorrect documentation (added by me in v3), and
> fixes that background_psql() didn't honor on_error_stop and extraparams passed
> by the user.  I've also added a commit which implements the \password test from
> the SCRAM iteration count patchset as well as cleaned up a few IPC::Run
> includes from test scripts.

And a v5 to fix a test failure in recovery tests.

--
Daniel Gustafsson

  18. Re: Making background psql nicer to use in tap tests

    Andres Freund <andres@anarazel.de> — 2023-04-02T21:37:58Z

    Hi,

On 2023-04-02 22:24:16 +0200, Daniel Gustafsson wrote:
> And a v5 to fix a test failure in recovery tests.

Thanks for workin gon this!


There's this XXX that I added:

> @@ -57,11 +51,10 @@ sub test_streaming
>  	COMMIT;
>  	});
>  
> -	$in .= q{
> -	COMMIT;
> -	\q
> -	};
> -	$h->finish;    # errors make the next test fail, so ignore them here
> +	$h->query_safe('COMMIT');
> +	$h->quit;
> +	# XXX: Not sure what this means
> +    # errors make the next test fail, so ignore them here
>  
>  	$node_publisher->wait_for_catchup($appname);

I still don't know what that comment is supposed to mean, unfortunately.

Greetings,

Andres Freund

  19. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-04-03T08:39:41Z

    > On 2 Apr 2023, at 23:37, Andres Freund <andres@anarazel.de> wrote:

> There's this XXX that I added:
> 
>> @@ -57,11 +51,10 @@ sub test_streaming
>> 	COMMIT;
>> 	});
>> 
>> -	$in .= q{
>> -	COMMIT;
>> -	\q
>> -	};
>> -	$h->finish;    # errors make the next test fail, so ignore them here
>> +	$h->query_safe('COMMIT');
>> +	$h->quit;
>> +	# XXX: Not sure what this means
>> +    # errors make the next test fail, so ignore them here
>> 
>> 	$node_publisher->wait_for_catchup($appname);
> 
> I still don't know what that comment is supposed to mean, unfortunately.

My reading of it is that it's ignoring any croak errors which IPC::Run might
throw if ->finish() isn't able to reap the psql process which had the \q.

I've added Amit who committed it in 216a784829c on cc: to see if he remembers
the comment in question and can shed some light.  Skimming the linked thread
yields no immediate clues.

--
Daniel Gustafsson

  20. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-04-05T21:44:31Z

    Unless there are objections I plan to get this in before the freeze, in order
to have better interactive tests starting with 16.  With a little bit of
documentation polish I think it's ready.

--
Daniel Gustafsson

  21. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-04-07T13:32:12Z

    > On 5 Apr 2023, at 23:44, Daniel Gustafsson <daniel@yesql.se> wrote:
> 
> Unless there are objections I plan to get this in before the freeze, in order
> to have better interactive tests starting with 16.  With a little bit of
> documentation polish I think it's ready.

When looking at the CFBot failure on Linux and Windows (not on macOS) I noticed
that it was down to the instance lacking IO::Pty.

[19:59:12.609](1.606s) ok 1 - scram_iterations in server side ROLE
Can't locate IO/Pty.pm in @INC (you may need to install the IO::Pty module) (@INC contains: /tmp/cirrus-ci-build/src/test/perl /tmp/cirrus-ci-build/src/test/authentication /etc/perl /usr/local/lib/i386-linux-gnu/perl/5.32.1 /usr/local/share/perl/5.32.1 /usr/lib/i386-linux-gnu/perl5/5.32 /usr/share/perl5 /usr/lib/i386-linux-gnu/perl/5.32 /usr/share/perl/5.32 /usr/local/lib/site_perl) at /usr/share/perl5/IPC/Run.pm line 1828.

Skimming the VM creation [0] it seems like it should be though?  On macOS the
module is installed inside Cirrus and the test runs fine.

I don't think we should go ahead with a patch that refactors interactive_psql
only to SKIP over it in CI (which is what the tab_completion test does now), so
let's wait until we have that sorted before going ahead.

--
Daniel Gustafsson

[0] https://github.com/anarazel/pg-vm-images/blob/main/scripts/linux_debian_install_deps.sh

  22. Re: Making background psql nicer to use in tap tests

    Andrew Dunstan <andrew@dunslane.net> — 2023-04-07T14:55:19Z

    On 2023-04-07 Fr 09:32, Daniel Gustafsson wrote:
>> On 5 Apr 2023, at 23:44, Daniel Gustafsson<daniel@yesql.se>  wrote:
>>
>> Unless there are objections I plan to get this in before the freeze, in order
>> to have better interactive tests starting with 16.  With a little bit of
>> documentation polish I think it's ready.
> When looking at the CFBot failure on Linux and Windows (not on macOS) I noticed
> that it was down to the instance lacking IO::Pty.
>
> [19:59:12.609](1.606s) ok 1 - scram_iterations in server side ROLE
> Can't locate IO/Pty.pm in @INC (you may need to install the IO::Pty module) (@INC contains: /tmp/cirrus-ci-build/src/test/perl /tmp/cirrus-ci-build/src/test/authentication /etc/perl /usr/local/lib/i386-linux-gnu/perl/5.32.1 /usr/local/share/perl/5.32.1 /usr/lib/i386-linux-gnu/perl5/5.32 /usr/share/perl5 /usr/lib/i386-linux-gnu/perl/5.32 /usr/share/perl/5.32 /usr/local/lib/site_perl) at /usr/share/perl5/IPC/Run.pm line 1828.
>
> Skimming the VM creation [0] it seems like it should be though?  On macOS the
> module is installed inside Cirrus and the test runs fine.
>
> I don't think we should go ahead with a patch that refactors interactive_psql
> only to SKIP over it in CI (which is what the tab_completion test does now), so
> let's wait until we have that sorted before going ahead.


It should probably be added to config/check_modules.pl if we're going to 
use it, but it seems to be missing for Strawberry Perl and msys/ucrt64 
perl and I'm not sure how easy it will be to add there. It would 
certainly add an installation burden for test instances at the very least.


cheers


andrew


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

  23. Re: Making background psql nicer to use in tap tests

    Andres Freund <andres@anarazel.de> — 2023-04-07T14:58:25Z

    Hi,

On 2023-04-07 15:32:12 +0200, Daniel Gustafsson wrote:
> > On 5 Apr 2023, at 23:44, Daniel Gustafsson <daniel@yesql.se> wrote:
> > 
> > Unless there are objections I plan to get this in before the freeze, in order
> > to have better interactive tests starting with 16.  With a little bit of
> > documentation polish I think it's ready.
> 
> When looking at the CFBot failure on Linux and Windows (not on macOS) I noticed
> that it was down to the instance lacking IO::Pty.
> 
> [19:59:12.609](1.606s) ok 1 - scram_iterations in server side ROLE
> Can't locate IO/Pty.pm in @INC (you may need to install the IO::Pty module) (@INC contains: /tmp/cirrus-ci-build/src/test/perl /tmp/cirrus-ci-build/src/test/authentication /etc/perl /usr/local/lib/i386-linux-gnu/perl/5.32.1 /usr/local/share/perl/5.32.1 /usr/lib/i386-linux-gnu/perl5/5.32 /usr/share/perl5 /usr/lib/i386-linux-gnu/perl/5.32 /usr/share/perl/5.32 /usr/local/lib/site_perl) at /usr/share/perl5/IPC/Run.pm line 1828.
> 
> Skimming the VM creation [0] it seems like it should be though?

Note it just fails on the 32build, not the 64bit build. Unfortunately I don't
think debian's multiarch in bullseye support installing enough of perl in
32bit and 64bit.

We can't have a hard dependency on non-default modules like IO::Pty anyway, so
the test needs to skip if it's not available.

On windows IO::Pty can't be installed, IIRC.


> I don't think we should go ahead with a patch that refactors interactive_psql
> only to SKIP over it in CI (which is what the tab_completion test does now), so
> let's wait until we have that sorted before going ahead.

Maybe I am a bit confused, but isn't that just an existing requirement? Why
would we expect this patchset to change what dependencies use of
interactive_psql() has?

Greetings,

Andres Freund

  24. Re: Making background psql nicer to use in tap tests

    Andres Freund <andres@anarazel.de> — 2023-04-07T15:04:13Z

    Hi,

On 2023-04-07 10:55:19 -0400, Andrew Dunstan wrote:
> It should probably be added to config/check_modules.pl if we're going to use
> it, but it seems to be missing for Strawberry Perl and msys/ucrt64 perl and
> I'm not sure how easy it will be to add there. It would certainly add an
> installation burden for test instances at the very least.

The last time I tried, it can't be installed on windows with cpan either, the
module simply doesn't have the necessary windows bits - likely because
traditionally windows didn't really have ptys. I think some stuff has been
added, but it probably would still require a bunch of portability work.

Note that we normally don't even build with readline support on windows - so
there's not really much point in using IO::Pty there. While I've gotten that
to work manually not too long ago, it's still manual and not documented etc.


Afaict the failures are purely about patch 2, not 1, right?

Greetings,

Andres Freund

  25. Re: Making background psql nicer to use in tap tests

    Tom Lane <tgl@sss.pgh.pa.us> — 2023-04-07T15:52:37Z

    Andres Freund <andres@anarazel.de> writes:
> On 2023-04-07 15:32:12 +0200, Daniel Gustafsson wrote:
>> I don't think we should go ahead with a patch that refactors interactive_psql
>> only to SKIP over it in CI (which is what the tab_completion test does now), so
>> let's wait until we have that sorted before going ahead.

> Maybe I am a bit confused, but isn't that just an existing requirement? Why
> would we expect this patchset to change what dependencies use of
> interactive_psql() has?

It is an existing requirement, but only for a test that's not too
critical.  If interactive_psql starts getting used for more interesting
things, we might be sad that the coverage is weak.

Having said that, weak coverage is better than no coverage.  I don't
think this point should be a show-stopper for committing.

			regards, tom lane

  26. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-04-07T15:55:08Z

    > On 7 Apr 2023, at 16:58, Andres Freund <andres@anarazel.de> wrote:

> Note it just fails on the 32build, not the 64bit build. Unfortunately I don't
> think debian's multiarch in bullseye support installing enough of perl in
> 32bit and 64bit.

I should probably avoid parsing logfiles with fever-induced brainfog, I
confused myself to think it was both =(

> Maybe I am a bit confused, but isn't that just an existing requirement? Why
> would we expect this patchset to change what dependencies use of
> interactive_psql() has?

Correct, there is no change from the current implementation.

--
Daniel Gustafsson

  27. Re: Making background psql nicer to use in tap tests

    Andres Freund <andres@anarazel.de> — 2023-04-07T15:58:23Z

    Hi,

On 2023-04-07 11:52:37 -0400, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > On 2023-04-07 15:32:12 +0200, Daniel Gustafsson wrote:
> >> I don't think we should go ahead with a patch that refactors interactive_psql
> >> only to SKIP over it in CI (which is what the tab_completion test does now), so
> >> let's wait until we have that sorted before going ahead.
> 
> > Maybe I am a bit confused, but isn't that just an existing requirement? Why
> > would we expect this patchset to change what dependencies use of
> > interactive_psql() has?
> 
> It is an existing requirement, but only for a test that's not too
> critical.  If interactive_psql starts getting used for more interesting
> things, we might be sad that the coverage is weak.

I don't really expect it to be used for non-critical things - after all,
interactive_psql() also depends on psql being built with readline support,
which we traditionally don't have on windows... For most tasks background_psql
should suffice...


> Having said that, weak coverage is better than no coverage.  I don't
> think this point should be a show-stopper for committing.

Yea.

One thing I wonder is whether we should have a central function for checking
if interactive_psql() is available, instead of copying 010_tab_completion.pl's
logic for it into multiple tests. But that could come later too.

Greetings,

Andres Freund

  28. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-04-07T16:14:37Z

    > On 7 Apr 2023, at 17:04, Andres Freund <andres@anarazel.de> wrote:

> Afaict the failures are purely about patch 2, not 1, right?

Correct.  The attached v6 wraps the interactive_psql test in a SKIP block with
a conditional on IO::Pty being available.

--
Daniel Gustafsson

  29. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-04-07T20:24:04Z

    > On 7 Apr 2023, at 18:14, Daniel Gustafsson <daniel@yesql.se> wrote:
>> On 7 Apr 2023, at 17:04, Andres Freund <andres@anarazel.de> wrote:

>> Afaict the failures are purely about patch 2, not 1, right?
> 
> Correct.  The attached v6 wraps the interactive_psql test in a SKIP block with
> a conditional on IO::Pty being available.

This version was green in the CFBot, so I ended up pushing it after some
documentation fixups and polish.

--
Daniel Gustafsson

  30. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-04-07T21:01:47Z

    > On 7 Apr 2023, at 22:24, Daniel Gustafsson <daniel@yesql.se> wrote:
> 
>> On 7 Apr 2023, at 18:14, Daniel Gustafsson <daniel@yesql.se> wrote:
>>> On 7 Apr 2023, at 17:04, Andres Freund <andres@anarazel.de> wrote:
> 
>>> Afaict the failures are purely about patch 2, not 1, right?
>> 
>> Correct.  The attached v6 wraps the interactive_psql test in a SKIP block with
>> a conditional on IO::Pty being available.
> 
> This version was green in the CFBot, so I ended up pushing it after some
> documentation fixups and polish.

Looks like morepork wasn't happy with the interactive \password test.

https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=morepork&dt=2023-04-07%2020%3A30%3A29

Looking into why the timer timed out in that test.

--
Daniel Gustafsson

  31. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-04-07T21:59:44Z

    > On 7 Apr 2023, at 23:01, Daniel Gustafsson <daniel@yesql.se> wrote:

> Looks like morepork wasn't happy with the interactive \password test.
> 
> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=morepork&dt=2023-04-07%2020%3A30%3A29
> 
> Looking into why the timer timed out in that test.

Staring at this I've been unable to figure out if there an underlying problem
here or a flaky testrun, since I can't reproduce it.  Maybe the animal owner
(on cc) have any insights?

The test has passed on several different platforms in the buildfarm, including
Linux, Solaris, macOS, NetBSD, FreeBSD and other OpenBSD animals.  It also
passed in an OpenBSD VM running with our Cirrus framework.

Unless there are objections raised I propose leaving it in for now, and I will
return to it tomorrow after some sleep, and install OpenBSD 6.9 to see if it's
reproducible.

--
Daniel Gustafsson

  32. Re: Making background psql nicer to use in tap tests

    Tom Lane <tgl@sss.pgh.pa.us> — 2023-04-07T22:35:06Z

    Daniel Gustafsson <daniel@yesql.se> writes:
>> On 7 Apr 2023, at 23:01, Daniel Gustafsson <daniel@yesql.se> wrote:
> Staring at this I've been unable to figure out if there an underlying problem
> here or a flaky testrun, since I can't reproduce it.  Maybe the animal owner
> (on cc) have any insights?

> The test has passed on several different platforms in the buildfarm, including
> Linux, Solaris, macOS, NetBSD, FreeBSD and other OpenBSD animals.  It also
> passed in an OpenBSD VM running with our Cirrus framework.

prion and mantid have now failed with the same symptom.  I don't
see a pattern, but it's not OpenBSD-only.  It will be interesting
to see if the failure is intermittent or not on those animals.

> Unless there are objections raised I propose leaving it in for now, and I will
> return to it tomorrow after some sleep, and install OpenBSD 6.9 to see if it's
> reproducible.

Agreed, we don't need a hasty revert here.  Better to gather data.

			regards, tom lane

  33. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-04-07T22:59:18Z

    > On 8 Apr 2023, at 00:35, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> 
> Daniel Gustafsson <daniel@yesql.se> writes:
>>> On 7 Apr 2023, at 23:01, Daniel Gustafsson <daniel@yesql.se> wrote:
>> Staring at this I've been unable to figure out if there an underlying problem
>> here or a flaky testrun, since I can't reproduce it.  Maybe the animal owner
>> (on cc) have any insights?
> 
>> The test has passed on several different platforms in the buildfarm, including
>> Linux, Solaris, macOS, NetBSD, FreeBSD and other OpenBSD animals.  It also
>> passed in an OpenBSD VM running with our Cirrus framework.
> 
> prion and mantid have now failed with the same symptom.  I don't
> see a pattern, but it's not OpenBSD-only.  It will be interesting
> to see if the failure is intermittent or not on those animals.

It would be interesting to know how far in the pumped input they get, if they
time out on the first one with nothing going through?  Will investigate further
tomorrow to see.

--
Daniel Gustafsson

  34. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-04-07T23:14:26Z

    > On 8 Apr 2023, at 00:59, Daniel Gustafsson <daniel@yesql.se> wrote:
> 
>> On 8 Apr 2023, at 00:35, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> 
>> Daniel Gustafsson <daniel@yesql.se> writes:
>>>> On 7 Apr 2023, at 23:01, Daniel Gustafsson <daniel@yesql.se> wrote:
>>> Staring at this I've been unable to figure out if there an underlying problem
>>> here or a flaky testrun, since I can't reproduce it.  Maybe the animal owner
>>> (on cc) have any insights?
>> 
>>> The test has passed on several different platforms in the buildfarm, including
>>> Linux, Solaris, macOS, NetBSD, FreeBSD and other OpenBSD animals.  It also
>>> passed in an OpenBSD VM running with our Cirrus framework.
>> 
>> prion and mantid have now failed with the same symptom.  I don't
>> see a pattern, but it's not OpenBSD-only.  It will be interesting
>> to see if the failure is intermittent or not on those animals.

morepork has failed again, which is good, since intermittent failures are
harder to track down.

> It would be interesting to know how far in the pumped input they get, if they
> time out on the first one with nothing going through?  Will investigate further
> tomorrow to see.

Actually, one quick datapoint.  prion and mantid report running IPC::Run
version 0.92, and morepork 0.96.  Animals that pass are running 20180523.0,
20200505.0, 20220807.0 or similar versions.  We don't print the IO::Pty version
during configure, but maybe this is related to older versions of the modules
and this test (not all of them apparently) need to SKIP if IO::Pty is missing
or too old?  Somewhere to start looking at the very least.

--
Daniel Gustafsson

  35. Re: Making background psql nicer to use in tap tests

    Andrew Dunstan <andrew@dunslane.net> — 2023-04-08T00:02:18Z

    On 2023-04-07 Fr 19:14, Daniel Gustafsson wrote:
>> On 8 Apr 2023, at 00:59, Daniel Gustafsson<daniel@yesql.se>  wrote:
>>
>>> On 8 Apr 2023, at 00:35, Tom Lane<tgl@sss.pgh.pa.us>  wrote:
>>>
>>> Daniel Gustafsson<daniel@yesql.se>  writes:
>>>>> On 7 Apr 2023, at 23:01, Daniel Gustafsson<daniel@yesql.se>  wrote:
>>>> Staring at this I've been unable to figure out if there an underlying problem
>>>> here or a flaky testrun, since I can't reproduce it.  Maybe the animal owner
>>>> (on cc) have any insights?
>>>> The test has passed on several different platforms in the buildfarm, including
>>>> Linux, Solaris, macOS, NetBSD, FreeBSD and other OpenBSD animals.  It also
>>>> passed in an OpenBSD VM running with our Cirrus framework.
>>> prion and mantid have now failed with the same symptom.  I don't
>>> see a pattern, but it's not OpenBSD-only.  It will be interesting
>>> to see if the failure is intermittent or not on those animals.
> morepork has failed again, which is good, since intermittent failures are
> harder to track down.
>
>> It would be interesting to know how far in the pumped input they get, if they
>> time out on the first one with nothing going through?  Will investigate further
>> tomorrow to see.
> Actually, one quick datapoint.  prion and mantid report running IPC::Run
> version 0.92, and morepork 0.96.  Animals that pass are running 20180523.0,
> 20200505.0, 20220807.0 or similar versions.  We don't print the IO::Pty version
> during configure, but maybe this is related to older versions of the modules
> and this test (not all of them apparently) need to SKIP if IO::Pty is missing
> or too old?  Somewhere to start looking at the very least.


Those aren't CPAN version numbers. See <https://metacpan.org/pod/IO::Pty>


prion was running 1.10 (dated to 2010). I have just updated it to 1.17 
(the CPAN latest). We'll see if that makes a difference.


cheers


andrew

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

  36. Re: Making background psql nicer to use in tap tests

    Tom Lane <tgl@sss.pgh.pa.us> — 2023-04-08T00:38:03Z

    Andrew Dunstan <andrew@dunslane.net> writes:
>> Actually, one quick datapoint.  prion and mantid report running IPC::Run
>> version 0.92, and morepork 0.96.  Animals that pass are running 20180523.0,
>> 20200505.0, 20220807.0 or similar versions.  We don't print the IO::Pty version
>> during configure, but maybe this is related to older versions of the modules
>> and this test (not all of them apparently) need to SKIP if IO::Pty is missing
>> or too old?  Somewhere to start looking at the very least.

> prion was running 1.10 (dated to 2010). I have just updated it to 1.17 
> (the CPAN latest). We'll see if that makes a difference.

I've been doing some checking with perlbrew locally.  It appears to not
be about IO::Pty so much as IPC::Run: it works with IPC::Run 0.99 but
not 0.79.  Still bisecting to identify exactly what's the minimum
okay version.

			regards, tom lane

  37. Re: Making background psql nicer to use in tap tests

    Tom Lane <tgl@sss.pgh.pa.us> — 2023-04-08T00:49:39Z

    I wrote:
> I've been doing some checking with perlbrew locally.  It appears to not
> be about IO::Pty so much as IPC::Run: it works with IPC::Run 0.99 but
> not 0.79.  Still bisecting to identify exactly what's the minimum
> okay version.

The answer is: it works with IPC::Run >= 0.98.  The version of IO::Pty
doesn't appear significant; it works at least back to 1.00 from early
2002.

IPC::Run 0.98 is relatively new (2018), so I don't think it'd fly
to make that our new minimum version across-the-board.  I recommend
just setting up this one test to SKIP if IPC::Run is too old.

			regards, tom lane

  38. Re: Making background psql nicer to use in tap tests

    Andres Freund <andres@anarazel.de> — 2023-04-08T00:57:33Z

    Hi,

On 2023-04-07 20:49:39 -0400, Tom Lane wrote:
> I wrote:
> > I've been doing some checking with perlbrew locally.  It appears to not
> > be about IO::Pty so much as IPC::Run: it works with IPC::Run 0.99 but
> > not 0.79.  Still bisecting to identify exactly what's the minimum
> > okay version.
> 
> The answer is: it works with IPC::Run >= 0.98.  The version of IO::Pty
> doesn't appear significant; it works at least back to 1.00 from early
> 2002.
> 
> IPC::Run 0.98 is relatively new (2018), so I don't think it'd fly
> to make that our new minimum version across-the-board.  I recommend
> just setting up this one test to SKIP if IPC::Run is too old.

Does the test actually take a while before it fails, or is it quick? It's
possible the failure is caused by 001_password.pl's use of
set_query_timer_restart(). I don't think other tests do something quite
comparable.

Greetings,

Andres Freund

  39. Re: Making background psql nicer to use in tap tests

    Tom Lane <tgl@sss.pgh.pa.us> — 2023-04-08T01:04:22Z

    Andres Freund <andres@anarazel.de> writes:
> On 2023-04-07 20:49:39 -0400, Tom Lane wrote:
>> IPC::Run 0.98 is relatively new (2018), so I don't think it'd fly
>> to make that our new minimum version across-the-board.  I recommend
>> just setting up this one test to SKIP if IPC::Run is too old.

> Does the test actually take a while before it fails, or is it quick?

It times out at whatever your PG_TEST_TIMEOUT_DEFAULT is.  I waited
3 minutes the first time, and then reduced that to 20sec for the
rest of the tries ...

			regards, tom lane

  40. Re: Making background psql nicer to use in tap tests

    Daniel Gustafsson <daniel@yesql.se> — 2023-04-08T07:53:58Z

    > On 8 Apr 2023, at 02:49, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> 
> I wrote:
>> I've been doing some checking with perlbrew locally.  It appears to not
>> be about IO::Pty so much as IPC::Run: it works with IPC::Run 0.99 but
>> not 0.79.  Still bisecting to identify exactly what's the minimum
>> okay version.
> 
> The answer is: it works with IPC::Run >= 0.98.  The version of IO::Pty
> doesn't appear significant; it works at least back to 1.00 from early
> 2002.

Thanks for investigating this!

> IPC::Run 0.98 is relatively new (2018), so I don't think it'd fly
> to make that our new minimum version across-the-board.  

Absolutely, that's not an option.

> I recommend
> just setting up this one test to SKIP if IPC::Run is too old.

Yes, will do that when I have a little more time at hand for monitoring the BF
later today.

--
Daniel Gustafsson

  41. Re: Making background psql nicer to use in tap tests

    Mikael Kjellström <mikael.kjellstrom@gmail.com> — 2023-04-08T08:00:50Z

    On 2023-04-08 09:53, Daniel Gustafsson wrote:

>> I recommend
>> just setting up this one test to SKIP if IPC::Run is too old.
> 
> Yes, will do that when I have a little more time at hand for monitoring the BF
> later today.

So what do you want me to do about grison and morepork?  I guess I could 
try to install a newer version of IPC::Run from CPAN or should I just 
leave it be?

/Mikael

  42. Re: Making background psql nicer to use in tap tests

    Tom Lane <tgl@sss.pgh.pa.us> — 2023-04-08T14:09:59Z

    =?UTF-8?Q?Mikael_Kjellstr=c3=b6m?= <mikael.kjellstrom@gmail.com> writes:
> So what do you want me to do about grison and morepork?  I guess I could 
> try to install a newer version of IPC::Run from CPAN or should I just 
> leave it be?

I think "leave it be" is fine.

			regards, tom lane

  43. Backporting BackgroundPsql

    Heikki Linnakangas <hlinnaka@iki.fi> — 2024-06-25T10:26:23Z

    While fixing a recent bug on visibility on a standby [1], I wrote a 
regression test that uses BackgroundPsql to run some queries in a 
long-running psql session. The problem is that that was refactored in 
v17, commit 664d757531. The test I wrote for v17 doesn't work as it is 
on backbranches. Options:

1. Write the new test differently on backbranches. Before 664d757531, 
the test needs to work a lot harder to use the background psql session, 
calling pump() etc. That's doable, but as noted in the discussion that 
led to 664d757531, it's laborious and error-prone.

2. Backport commit 664d757531. This might break out-of-tree perl tests 
that use the background_psql() function. I don't know if any such tests 
exist, and they would need to be changed for v17 anyway, so that seems 
acceptable. Anyone aware of any extensions using the perl test modules?

3. Backport commit 664d757531, but keep the existing background_psql() 
function unchanged. Add a different constructor to get the v17-style 
BackgroundPsql session, something like "$node->background_psql_new()".

I'm leaning towards 3. We might need to backport more perl tests that 
use background_psql() in the future, backporting the test module will 
make that easier.

Thoughts?

[1] 
https://www.postgresql.org/message-id/6b852e98-2d49-4ca1-9e95-db419a2696e0@iki.fi

-- 
Heikki Linnakangas
Neon (https://neon.tech)

  44. Re: Backporting BackgroundPsql

    Heikki Linnakangas <hlinnaka@iki.fi> — 2024-06-25T10:32:59Z

    On 25/06/2024 13:26, Heikki Linnakangas wrote:
> While fixing a recent bug on visibility on a standby [1], I wrote a 
> regression test that uses BackgroundPsql to run some queries in a 
> long-running psql session. The problem is that that was refactored in 
> v17, commit 664d757531. The test I wrote for v17 doesn't work as it is 
> on backbranches. Options:

Sorry, I meant v16. The BackgroundPsql refactorings went into v16. The 
backporting question remains for v15 and below.

-- 
Heikki Linnakangas
Neon (https://neon.tech)

  45. Re: Backporting BackgroundPsql

    Andres Freund <andres@anarazel.de> — 2024-06-25T11:40:38Z

    Hi,

On 2024-06-25 13:26:23 +0300, Heikki Linnakangas wrote:
> While fixing a recent bug on visibility on a standby [1], I wrote a
> regression test that uses BackgroundPsql to run some queries in a
> long-running psql session. The problem is that that was refactored in v17,
> commit 664d757531. The test I wrote for v17 doesn't work as it is on
> backbranches. Options:
> 
> 1. Write the new test differently on backbranches. Before 664d757531, the
> test needs to work a lot harder to use the background psql session, calling
> pump() etc. That's doable, but as noted in the discussion that led to
> 664d757531, it's laborious and error-prone.
> 
> 2. Backport commit 664d757531. This might break out-of-tree perl tests that
> use the background_psql() function. I don't know if any such tests exist,
> and they would need to be changed for v17 anyway, so that seems acceptable.
> Anyone aware of any extensions using the perl test modules?
> 
> 3. Backport commit 664d757531, but keep the existing background_psql()
> function unchanged. Add a different constructor to get the v17-style
> BackgroundPsql session, something like "$node->background_psql_new()".
> 
> I'm leaning towards 3. We might need to backport more perl tests that use
> background_psql() in the future, backporting the test module will make that
> easier.
> 
> Thoughts?

Yes, I've wished for this a couple times. I think 2 or 3 would be reasonable.
I think 1) often just leads to either tests not being written or being
fragile...

Greetings,

Andres Freund

  46. Re: Backporting BackgroundPsql

    Tom Lane <tgl@sss.pgh.pa.us> — 2024-06-25T14:26:00Z

    Andres Freund <andres@anarazel.de> writes:
> On 2024-06-25 13:26:23 +0300, Heikki Linnakangas wrote:
>> 1. Write the new test differently on backbranches. Before 664d757531, the
>> test needs to work a lot harder to use the background psql session, calling
>> pump() etc. That's doable, but as noted in the discussion that led to
>> 664d757531, it's laborious and error-prone.
>> 
>> 2. Backport commit 664d757531. This might break out-of-tree perl tests that
>> use the background_psql() function. I don't know if any such tests exist,
>> and they would need to be changed for v17 anyway, so that seems acceptable.
>> Anyone aware of any extensions using the perl test modules?
>> 
>> 3. Backport commit 664d757531, but keep the existing background_psql()
>> function unchanged. Add a different constructor to get the v17-style
>> BackgroundPsql session, something like "$node->background_psql_new()".

> Yes, I've wished for this a couple times. I think 2 or 3 would be reasonable.
> I think 1) often just leads to either tests not being written or being
> fragile...

I'd vote for (2).  (3) is just leaving a foot-gun for people to
hurt themselves with.

			regards, tom lane

  47. Re: Backporting BackgroundPsql

    Melanie Plageman <melanieplageman@gmail.com> — 2024-06-25T15:34:22Z

    On Tue, Jun 25, 2024 at 7:40 AM Andres Freund <andres@anarazel.de> wrote:
>
> Hi,
>
> On 2024-06-25 13:26:23 +0300, Heikki Linnakangas wrote:
> > While fixing a recent bug on visibility on a standby [1], I wrote a
> > regression test that uses BackgroundPsql to run some queries in a
> > long-running psql session. The problem is that that was refactored in v17,
> > commit 664d757531. The test I wrote for v17 doesn't work as it is on
> > backbranches. Options:
> >
> > 1. Write the new test differently on backbranches. Before 664d757531, the
> > test needs to work a lot harder to use the background psql session, calling
> > pump() etc. That's doable, but as noted in the discussion that led to
> > 664d757531, it's laborious and error-prone.
> >
> > 2. Backport commit 664d757531. This might break out-of-tree perl tests that
> > use the background_psql() function. I don't know if any such tests exist,
> > and they would need to be changed for v17 anyway, so that seems acceptable.
> > Anyone aware of any extensions using the perl test modules?
> >
> > 3. Backport commit 664d757531, but keep the existing background_psql()
> > function unchanged. Add a different constructor to get the v17-style
> > BackgroundPsql session, something like "$node->background_psql_new()".
> >
> > I'm leaning towards 3. We might need to backport more perl tests that use
> > background_psql() in the future, backporting the test module will make that
> > easier.
> >
> > Thoughts?
>
> Yes, I've wished for this a couple times. I think 2 or 3 would be reasonable.
> I think 1) often just leads to either tests not being written or being
> fragile...

+1 to backporting background psql!

I'm also okay with 2 or 3. But note that for 2, there are several
tests with skip_all and at least one of them uses background_psql
(031_recovery_conflict.pl), so we'll just have to remember to update
those. I assume that is easy enough to do if you grep for
background_psql -- but, just in case you were going to be 100%
test-driven :)

- Melanie

  48. Re: Backporting BackgroundPsql

    Andrew Dunstan <andrew@dunslane.net> — 2024-06-25T15:59:01Z

    On 2024-06-25 Tu 10:26 AM, Tom Lane wrote:
> Andres Freund<andres@anarazel.de>  writes:
>> On 2024-06-25 13:26:23 +0300, Heikki Linnakangas wrote:
>>> 1. Write the new test differently on backbranches. Before 664d757531, the
>>> test needs to work a lot harder to use the background psql session, calling
>>> pump() etc. That's doable, but as noted in the discussion that led to
>>> 664d757531, it's laborious and error-prone.
>>>
>>> 2. Backport commit 664d757531. This might break out-of-tree perl tests that
>>> use the background_psql() function. I don't know if any such tests exist,
>>> and they would need to be changed for v17 anyway, so that seems acceptable.
>>> Anyone aware of any extensions using the perl test modules?
>>>
>>> 3. Backport commit 664d757531, but keep the existing background_psql()
>>> function unchanged. Add a different constructor to get the v17-style
>>> BackgroundPsql session, something like "$node->background_psql_new()".
>> Yes, I've wished for this a couple times. I think 2 or 3 would be reasonable.
>> I think 1) often just leads to either tests not being written or being
>> fragile...
> I'd vote for (2).  (3) is just leaving a foot-gun for people to
> hurt themselves with.
>
> 			


+1


I'd like to get rid of it in its current form at least. Just about all 
the uses I'm aware of could be transformed to use the Session object 
I've been working on, based either on FFI or a small XS wrapper for some 
of libpq.

cheers

andrew

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

  49. Re: Backporting BackgroundPsql

    Daniel Gustafsson <daniel@yesql.se> — 2024-06-25T20:47:11Z

    > On 25 Jun 2024, at 16:26, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> 
> Andres Freund <andres@anarazel.de> writes:
>> On 2024-06-25 13:26:23 +0300, Heikki Linnakangas wrote:
>>> 1. Write the new test differently on backbranches. Before 664d757531, the
>>> test needs to work a lot harder to use the background psql session, calling
>>> pump() etc. That's doable, but as noted in the discussion that led to
>>> 664d757531, it's laborious and error-prone.
>>> 
>>> 2. Backport commit 664d757531. This might break out-of-tree perl tests that
>>> use the background_psql() function. I don't know if any such tests exist,
>>> and they would need to be changed for v17 anyway, so that seems acceptable.
>>> Anyone aware of any extensions using the perl test modules?
>>> 
>>> 3. Backport commit 664d757531, but keep the existing background_psql()
>>> function unchanged. Add a different constructor to get the v17-style
>>> BackgroundPsql session, something like "$node->background_psql_new()".
> 
>> Yes, I've wished for this a couple times. I think 2 or 3 would be reasonable.
>> I think 1) often just leads to either tests not being written or being
>> fragile...
> 
> I'd vote for (2).  (3) is just leaving a foot-gun for people to
> hurt themselves with.

I agree with this, if we're backporting we should opt for 2.  The only out of
tree user of background_psql() that I could find was check_pgactivity but they
seem to have vendored an old copy of our lib rather than use anything in a our
tree so we should be fine there.

Before pulling any triggers I think https://commitfest.postgresql.org/48/4959/
should be considered, since Tom found some flaws in the current code around how
timers and timeouts are used.

However, since Andrew is actively aiming to replace all of this shortly, should
we wait a see where that lands to avoid having to backport another library
change?

--
Daniel Gustafsson

  50. Re: Backporting BackgroundPsql

    Tom Lane <tgl@sss.pgh.pa.us> — 2024-06-25T20:57:17Z

    Daniel Gustafsson <daniel@yesql.se> writes:
> Before pulling any triggers I think https://commitfest.postgresql.org/48/4959/
> should be considered, since Tom found some flaws in the current code around how
> timers and timeouts are used.

That's certainly another issue to consider, but is it really a blocker
for this one?

> However, since Andrew is actively aiming to replace all of this shortly, should
> we wait a see where that lands to avoid having to backport another library
> change?

I would like to see what he comes up with ... but is it likely to
be something we'd risk back-patching?

			regards, tom lane

  51. Re: Backporting BackgroundPsql

    Daniel Gustafsson <daniel@yesql.se> — 2024-06-25T21:10:14Z

    > On 25 Jun 2024, at 22:57, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> 
> Daniel Gustafsson <daniel@yesql.se> writes:
>> Before pulling any triggers I think https://commitfest.postgresql.org/48/4959/
>> should be considered, since Tom found some flaws in the current code around how
>> timers and timeouts are used.
> 
> That's certainly another issue to consider, but is it really a blocker
> for this one?

It's not a blocker, but when poking at the code it seems useful to consider the
open items around it.

>> However, since Andrew is actively aiming to replace all of this shortly, should
>> we wait a see where that lands to avoid having to backport another library
>> change?
> 
> I would like to see what he comes up with ... but is it likely to
> be something we'd risk back-patching?

Maybe it'll be a stretch given that it's likely to introduce new dependencies.

--
Daniel Gustafsson

  52. Re: Backporting BackgroundPsql

    Alvaro Herrera <alvherre@alvh.no-ip.org> — 2024-06-26T00:12:42Z

    On 2024-Jun-25, Tom Lane wrote:

> Daniel Gustafsson <daniel@yesql.se> writes:

> > However, since Andrew is actively aiming to replace all of this shortly, should
> > we wait a see where that lands to avoid having to backport another library
> > change?
> 
> I would like to see what he comes up with ... but is it likely to
> be something we'd risk back-patching?

FWIW I successfully used the preliminary PqFFI stuff Andrew posted to
write a test program for bug #18377, which I think ended up being better
than with BackgroundPsql, so I think it's a good way forward.  As for
back-patching it, I suspect we're going to end up backpatching the
framework anyway just because we'll want to have it available for
backpatching future tests, even if we keep a backpatch minimal by doing
only the framework and not existing tests.

I also backpatched the PqFFI and PostgreSQL::Session modules to older PG
branches, to run my test program there.  This required only removing
some lines from PqFFI.pm that were about importing libpq functions that
older libpq didn't have.

Of course, the PostgreSQL::Session stuff is not ready yet, so if we want
this test in the tree soon, I don't think we should wait.


I'll note, though, that Test::More doesn't work terribly nicely with
perl threads, but that only relates to my test program and not to PqFFI.

-- 
Álvaro Herrera        Breisgau, Deutschland  —  https://www.EnterpriseDB.com/
"La gente vulgar sólo piensa en pasar el tiempo;
el que tiene talento, en aprovecharlo"

  53. Re: Backporting BackgroundPsql

    Michael Paquier <michael@paquier.xyz> — 2024-06-26T00:25:13Z

    On Wed, Jun 26, 2024 at 02:12:42AM +0200, Alvaro Herrera wrote:
> FWIW I successfully used the preliminary PqFFI stuff Andrew posted to
> write a test program for bug #18377, which I think ended up being better
> than with BackgroundPsql, so I think it's a good way forward.  As for
> back-patching it, I suspect we're going to end up backpatching the
> framework anyway just because we'll want to have it available for
> backpatching future tests, even if we keep a backpatch minimal by doing
> only the framework and not existing tests.
> 
> I also backpatched the PqFFI and PostgreSQL::Session modules to older PG
> branches, to run my test program there.  This required only removing
> some lines from PqFFI.pm that were about importing libpq functions that
> older libpq didn't have.

Nice!  I definitely +1 the backpatching of the testing bits.  This
stuff can make validating bugs so much easier, particularly when there
are conflicting parts in the backend after a cherry-pick.
--
Michael

  54. Re: Backporting BackgroundPsql

    Heikki Linnakangas <hlinnaka@iki.fi> — 2024-06-26T07:34:31Z

    On 26/06/2024 03:25, Michael Paquier wrote:
> On Wed, Jun 26, 2024 at 02:12:42AM +0200, Alvaro Herrera wrote:
>> FWIW I successfully used the preliminary PqFFI stuff Andrew posted to
>> write a test program for bug #18377, which I think ended up being better
>> than with BackgroundPsql, so I think it's a good way forward.  As for
>> back-patching it, I suspect we're going to end up backpatching the
>> framework anyway just because we'll want to have it available for
>> backpatching future tests, even if we keep a backpatch minimal by doing
>> only the framework and not existing tests.
>>
>> I also backpatched the PqFFI and PostgreSQL::Session modules to older PG
>> branches, to run my test program there.  This required only removing
>> some lines from PqFFI.pm that were about importing libpq functions that
>> older libpq didn't have.
> 
> Nice!  I definitely +1 the backpatching of the testing bits.  This
> stuff can make validating bugs so much easier, particularly when there
> are conflicting parts in the backend after a cherry-pick.

I haven't looked closely at the new PgFFI stuff but +1 on that in 
general, and it makes sense to backport that once it lands on master. In 
the meanwhile, I think we should backport BackgroundPsql as it is, to 
make it possible to backport tests using it right now, even if it is 
short-lived.

-- 
Heikki Linnakangas
Neon (https://neon.tech)

  55. Re: Backporting BackgroundPsql

    Robert Haas <robertmhaas@gmail.com> — 2024-06-26T11:54:42Z

    On Wed, Jun 26, 2024 at 3:34 AM Heikki Linnakangas <hlinnaka@iki.fi> wrote:
> I haven't looked closely at the new PgFFI stuff but +1 on that in
> general, and it makes sense to backport that once it lands on master. In
> the meanwhile, I think we should backport BackgroundPsql as it is, to
> make it possible to backport tests using it right now, even if it is
> short-lived.

+1. The fact that PgFFI may be coming isn't a reason to not back-patch
this. The risk of back-patching testing infrastructure is also very
low as compared with code; in fact, there's a lot of risk from NOT
back-patching popular testing infrastructure.

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

  56. Re: Backporting BackgroundPsql

    Heikki Linnakangas <hlinnaka@iki.fi> — 2024-06-27T16:35:05Z

    On 26/06/2024 14:54, Robert Haas wrote:
> On Wed, Jun 26, 2024 at 3:34 AM Heikki Linnakangas <hlinnaka@iki.fi> wrote:
>> I haven't looked closely at the new PgFFI stuff but +1 on that in
>> general, and it makes sense to backport that once it lands on master. In
>> the meanwhile, I think we should backport BackgroundPsql as it is, to
>> make it possible to backport tests using it right now, even if it is
>> short-lived.
> 
> +1. The fact that PgFFI may be coming isn't a reason to not back-patch
> this. The risk of back-patching testing infrastructure is also very
> low as compared with code; in fact, there's a lot of risk from NOT
> back-patching popular testing infrastructure.

Ok, I pushed commits to backport BackgroundPsql down to v12. I used 
"option 2", i.e. I changed background_psql() to return the new 
BackgroundPsql object.

-- 
Heikki Linnakangas
Neon (https://neon.tech)

  57. Re: Backporting BackgroundPsql

    Pavan Deolasee <pavan.deolasee@gmail.com> — 2024-06-29T04:38:19Z

    On Thu, Jun 27, 2024 at 10:05 PM Heikki Linnakangas <hlinnaka@iki.fi> wrote:

>
> Ok, I pushed commits to backport BackgroundPsql down to v12. I used
> "option 2", i.e. I changed background_psql() to return the new
> BackgroundPsql object.
>
>
Don't we need to add install and uninstall rules for the new module, like
we did in
https://git.postgresql.org/pg/commitdiff/a4c17c86176cfa712f541b81b2a026ae054b275e
and
https://git.postgresql.org/pg/commitdiff/7039c7cff6736780c3bbb41a90a6dfea0f581ad2
?

Thanks,
Pavan

  58. Re: Backporting BackgroundPsql

    Daniel Gustafsson <daniel@yesql.se> — 2024-07-01T07:39:25Z

    > On 29 Jun 2024, at 06:38, Pavan Deolasee <pavan.deolasee@gmail.com> wrote:

> Don't we need to add install and uninstall rules for the new module, like we did in https://git.postgresql.org/pg/commitdiff/a4c17c86176cfa712f541b81b2a026ae054b275e and  https://git.postgresql.org/pg/commitdiff/7039c7cff6736780c3bbb41a90a6dfea0f581ad2 ?

Thats correct, we should backport those as well.

--
Daniel Gustafsson

  59. Re: Backporting BackgroundPsql

    Pavan Deolasee <pavan.deolasee@gmail.com> — 2024-07-01T14:11:35Z

    H i Daniel,

On Mon, Jul 1, 2024 at 1:09 PM Daniel Gustafsson <daniel@yesql.se> wrote:

> > On 29 Jun 2024, at 06:38, Pavan Deolasee <pavan.deolasee@gmail.com>
> wrote:
>
> > Don't we need to add install and uninstall rules for the new module,
> like we did in
> https://git.postgresql.org/pg/commitdiff/a4c17c86176cfa712f541b81b2a026ae054b275e
> and
> https://git.postgresql.org/pg/commitdiff/7039c7cff6736780c3bbb41a90a6dfea0f581ad2
> ?
>
> Thats correct, we should backport those as well.
>

Thanks for confirming. Attaching patches for PG15 and PG14, but this will
need backporting all the way up to PG12.

Thanks,
Pavan

  60. Re: Backporting BackgroundPsql

    Heikki Linnakangas <hlinnaka@iki.fi> — 2024-07-01T16:33:47Z

    On 01/07/2024 17:11, Pavan Deolasee wrote:
> H i Daniel,
> 
> On Mon, Jul 1, 2024 at 1:09 PM Daniel Gustafsson <daniel@yesql.se 
> <mailto:daniel@yesql.se>> wrote:
> 
>      > On 29 Jun 2024, at 06:38, Pavan Deolasee
>     <pavan.deolasee@gmail.com <mailto:pavan.deolasee@gmail.com>> wrote:
> 
>      > Don't we need to add install and uninstall rules for the new
>     module, like we did in
>     https://git.postgresql.org/pg/commitdiff/a4c17c86176cfa712f541b81b2a026ae054b275e <https://git.postgresql.org/pg/commitdiff/a4c17c86176cfa712f541b81b2a026ae054b275e> and https://git.postgresql.org/pg/commitdiff/7039c7cff6736780c3bbb41a90a6dfea0f581ad2 <https://git.postgresql.org/pg/commitdiff/7039c7cff6736780c3bbb41a90a6dfea0f581ad2> ?
> 
>     Thats correct, we should backport those as well.
> 
> Thanks for confirming. Attaching patches for PG15 and PG14, but this 
> will need backporting all the way up to PG12.

Thanks! Pushed to v12-v15.

-- 
Heikki Linnakangas
Neon (https://neon.tech)