Thread: [HACKERS] Latest Data::Dumper breaks hstore_plperl regression test

[HACKERS] Latest Data::Dumper breaks hstore_plperl regression test

From
Tom Lane
Date:
Or at least, that's what I surmise from the fact that buildfarm critter
caiman has been failing that test for the last day or so, with symptoms
indicating whitespace changes in Data::Dumper output.  Some poking into
the Fedora repo shows that rawhide updated perl-Data-Dumper from 2.161
to 2.167 on May 11, so that fits ...
        regards, tom lane



Re: [HACKERS] Latest Data::Dumper breaks hstore_plperl regression test

From
Robert Haas
Date:
On Sat, May 13, 2017 at 6:22 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Or at least, that's what I surmise from the fact that buildfarm critter
> caiman has been failing that test for the last day or so, with symptoms
> indicating whitespace changes in Data::Dumper output.  Some poking into
> the Fedora repo shows that rawhide updated perl-Data-Dumper from 2.161
> to 2.167 on May 11, so that fits ...

Depending on the precise details of how Data::Dumper prints things
doesn't seem like a particularly good idea.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company



Re: [HACKERS] Latest Data::Dumper breaks hstore_plperl regression test

From
Craig Ringer
Date:


On 14 May 2017 11:33, "Robert Haas" <robertmhaas@gmail.com> wrote:
On Sat, May 13, 2017 at 6:22 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Or at least, that's what I surmise from the fact that buildfarm critter
> caiman has been failing that test for the last day or so, with symptoms
> indicating whitespace changes in Data::Dumper output.  Some poking into
> the Fedora repo shows that rawhide updated perl-Data-Dumper from 2.161
> to 2.167 on May 11, so that fits ...

Depending on the precise details of how Data::Dumper prints things
doesn't seem like a particularly good idea.

Surely Test::More::is_deeply is the way to go here.

Re: [HACKERS] Latest Data::Dumper breaks hstore_plperl regressiontest

From
Andrew Dunstan
Date:

On 05/13/2017 11:43 PM, Craig Ringer wrote:
>
>
> On 14 May 2017 11:33, "Robert Haas" <robertmhaas@gmail.com
> <mailto:robertmhaas@gmail.com>> wrote:
>
>     On Sat, May 13, 2017 at 6:22 PM, Tom Lane <tgl@sss.pgh.pa.us
>     <mailto:tgl@sss.pgh.pa.us>> wrote:
>     > Or at least, that's what I surmise from the fact that buildfarm
>     critter
>     > caiman has been failing that test for the last day or so, with
>     symptoms
>     > indicating whitespace changes in Data::Dumper output.  Some
>     poking into
>     > the Fedora repo shows that rawhide updated perl-Data-Dumper from
>     2.161
>     > to 2.167 on May 11, so that fits ...
>
>     Depending on the precise details of how Data::Dumper prints things
>     doesn't seem like a particularly good idea.
>
>
> Surely Test::More::is_deeply is the way to go here.

This isn't a TAP test. It's a standard pg_regress test.

cheers

andrew

-- 
Andrew Dunstan                https://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services




Re: [HACKERS] Latest Data::Dumper breaks hstore_plperl regressiontest

From
Andrew Dunstan
Date:

On 05/13/2017 11:32 PM, Robert Haas wrote:
> On Sat, May 13, 2017 at 6:22 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Or at least, that's what I surmise from the fact that buildfarm critter
>> caiman has been failing that test for the last day or so, with symptoms
>> indicating whitespace changes in Data::Dumper output.  Some poking into
>> the Fedora repo shows that rawhide updated perl-Data-Dumper from 2.161
>> to 2.167 on May 11, so that fits ...
> Depending on the precise details of how Data::Dumper prints things
> doesn't seem like a particularly good idea.
>


I'd be inclined to set $Data::Dumper::Indent to 0 which would suppress
all indentation, and adjusting the test results accordingly. We already
set $Data::Dumper::Sortkeys to 1, so there's precedent for controlling
how it presents data back to us.

cheers

andrew



-- 
Andrew Dunstan                https://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services




Andrew Dunstan <andrew.dunstan@2ndquadrant.com> writes:
> On 05/13/2017 11:32 PM, Robert Haas wrote:
>> Depending on the precise details of how Data::Dumper prints things
>> doesn't seem like a particularly good idea.

> I'd be inclined to set $Data::Dumper::Indent to 0 which would suppress
> all indentation, and adjusting the test results accordingly. We already
> set $Data::Dumper::Sortkeys to 1, so there's precedent for controlling
> how it presents data back to us.

Sounds plausible to me.
        regards, tom lane



Re: [HACKERS] Latest Data::Dumper breaks hstore_plperl regressiontest

From
Andrew Dunstan
Date:

On 05/14/2017 12:04 AM, Andrew Dunstan wrote:
>
> On 05/13/2017 11:32 PM, Robert Haas wrote:
>> On Sat, May 13, 2017 at 6:22 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>>> Or at least, that's what I surmise from the fact that buildfarm critter
>>> caiman has been failing that test for the last day or so, with symptoms
>>> indicating whitespace changes in Data::Dumper output.  Some poking into
>>> the Fedora repo shows that rawhide updated perl-Data-Dumper from 2.161
>>> to 2.167 on May 11, so that fits ...
>> Depending on the precise details of how Data::Dumper prints things
>> doesn't seem like a particularly good idea.
>>
>
> I'd be inclined to set $Data::Dumper::Indent to 0 which would suppress
> all indentation, and adjusting the test results accordingly. We already
> set $Data::Dumper::Sortkeys to 1, so there's precedent for controlling
> how it presents data back to us.
>


Here's a patch along those lines.

cheers

andrew


-- 
Andrew Dunstan                https://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Attachment
Andrew Dunstan <andrew.dunstan@2ndquadrant.com> writes:
> Here's a patch along those lines.

Looks sane in a quick once-over.  Note this needs to be back-patched.
        regards, tom lane