pgxn-tester update

[Tomas Vondra]

Hi there,

I've updated all the pgxn-tester machines with the same changes I did ~3
weeks ago on 'hactar' while getting the multicorn work. I've also
updated all the branches to the last minor release, and added the
9.5alpha1 that was released recently.

I had to discard all the old results, because some of those were likely
broken just like multicorn, and identifying individual packages would be
a major PITA. All the results should be available already.

I've noticed a few new failures, especially in the 'check' phase, i.e.
when running regression tests. This looks a bit annoying:

\set ECHO 0
+ unrecognized value "0" for "ECHO"; assuming "none"

Apparently the 'ECHO' no longer recognizes "0" and many regression tests
fail because of that.

Let me know if you run into anything else, especially if you believe
it's a bug in the pgxn-tester tooling.

regards
Tomas

[Ronan Dunklau]

Thank you tomas.

It seems that not every server is configured the same with regards to python
installation:

http://pgxn-tester.org/distributions/multicorn/1.2.3

On hal-9000:
Found Python 2.6, but 2.7 is required.

On hactar:
ImportError: No module named setuptools

Is this on purpose ? Is there a page summarizing the server configuration /
installation ?

Thank you !


>
> regards
> Tomas

[Tomas Vondra]

On 07/20/2015 10:50 AM, Ronan Dunklau wrote:
>
> It seems that not every server is configured the same with regards to python
> installation:
>
> http://pgxn-tester.org/distributions/multicorn/1.2.3
>
> On hal-9000:
> Found Python 2.6, but 2.7 is required.
>
> On hactar:
> ImportError: No module named setuptools
>
> Is this on purpose ? Is there a page summarizing the server
configuration / installation ?

No, that's not intentional - I'll fix that.

The only "user" documentation is at http://pgxn-tester.org/about, but it
does not contain the information you're asking about. I'll look into
that too.

regards
Tomas

[Jim Nasby]

It would be nice if we had variation on the server config, but that
would presumably require some way to specify external dependencies.
--
Jim Nasby, Data Architect, Blue Treble Consulting, Austin TX
Data in Trouble? Get it in Treble! http://BlueTreble.com

[Tomas Vondra]

On 07/20/2015 06:21 PM, Tomas Vondra wrote:
>
>
> On 07/20/2015 10:50 AM, Ronan Dunklau wrote:
>>
>> It seems that not every server is configured the same with regards to
>> python
>> installation:
>>
>> http://pgxn-tester.org/distributions/multicorn/1.2.3
>>
>> On hal-9000:
>> Found Python 2.6, but 2.7 is required.

Fixed. clearly I forgot to set PATH properly when initializing the
virtualenv, so it was using the default Python version.

>>
>> On hactar:
>> ImportError: No module named setuptools

This is a bit strange, though. The virtualenv on hactar is definitely
using the right python version, but I've found failures like this:

http://pgxn-tester.org/results/d7e0d1bc-75fa-4732-9da4-229409d2a16b

INFO: best version: multicorn 1.0.0
INFO: saving /var/pgxn-tester/hactar/tmp/tmpEcOwQY/multicorn-1.0.0.zip
INFO: unpacking: /var/pgxn-tester/hactar/tmp/tmpEcOwQY/multicorn-1.0.0.zip
INFO: building extension
Python version is 2.6
./preflight-check.sh
cp sql/multicorn.sql sql/multicorn--1.0.0.sql
...
INFO: installing extension
Python version is 2.6
./preflight-check.sh
...
python2.6 ./setup--1.0.0.py install
Traceback (most recent call last):
File "./setup--1.0.0.py", line 2, in
from setuptools import setup, find_packages, Extension

ImportError: No module named setuptools

gmake: *** [python_code] Error 1


Clearly it still refers to Python 2.6 for some reason, so it does not
find the setuptools installed in the virtualenv. I'm not sure where the
$(PYTHON) variable in Makefile comes from, but I assume it comes from
the PostgreSQL build - I'll try to rebuild it with the correct PATH (so
that it finds Python 2.7). Hope that'll fix it.

regards
Tomas