Re^5: How to do automated tests of a database centric CPAN module

[erix]

in reply to Re^4: How to do automated tests of a database centric CPAN module
in thread How to do automated tests of a database centric CPAN module

That's true but in that it's not different from Test::postgresql, as far as I can see.

From Test::postgresql docs:

new

[...]  If required programs (initdb and postmaster)
were not found, the function returns undef and sets
appropriate message to $Test::postgresql::errstr.

I think it's really only DBD::SQLite that will give you a complete package, from scratch.

( Probably one could create a package with all the bits to run a PostgreSQL instance from the binaries from the EnterpriseDB site [1], but that looks like quite an undertaking...:

  (cannot run as root)

 http://www.enterprisedb.com/products-services-training/pgbindownload
      "Binaries from installer version 9.2.1"

-- to use, unzip the pgsql directory somewhere, and cd into it:
      cd pgsql

-- run initdb:
      ./bin/initdb -D 9.2/data

-- start postgres:
      ./bin/pg_ctl/ -D 9.2/data -l logfile start

-- run tests...

-- to stop, close all connections, then:
      ./bin/pg_ctl -D 9.2/data -l logfile stop 

-- cleanup all  files

Hm, I don't know - possible but error-prone, probably...

)

[1] http://www.enterprisedb.com/products-services-training/pgbindownload

[jds17]

erix, thank you for pointing this out, I did not yet look at Test::postgresql (being busy handling SQLite support, until today I did not know just how different SQLite is from other DBMSs, in particular when trying to get meta data...). So I cannot use Test::postgresql for the stated purposes either.

[jds17]

I don't think it would be feasible to run such a PostgreSQL install during my module install, I will stick to SQLite for the time being, thank you for elaborating on this.

erix, please allow me a well-intentioned reminder: please mark updates to your messages as such, otherwise the thread may get confusing. Better still, post them as a new message (there are always exceptions, e.g. when a definite answer is found, I like to add an update to the root node containing the original question so the answer can be found with ease).