* Cédric Krier: " Re: [tryton-dev] Scenario fail on MySQL" (Wed, 19 Feb 2014 00:50:21 +0100):
> On 06 Feb 01:00, Cédric Krier wrote: > > Hi, > > > > With the recent drop of Python 2.6, I re-wrote the scenario to get > > better error message like this [1]. It allowed to find some misbehavior > > between PostgreSQL and SQLite. > > The problem is that MySQL again doesn't play well [2] because we have to > > put constraint on the DECIMAL column [3] and MySQL doesn't succeed to > > return Decimal with the same precision as we send. > > I thought about using Decimal.normalize() [4] to get standard format to > > test in doctest but I don't like too much because it will hide other > > issues (like the one fixed with this change for SQLite). > > > > So I'm calling for ideas… > > So I'm thinking about skipping all doctest with MySQL backend because in > some way, it is just the unittest that really should check the internal > behavior of the code and scenario are just there for the big picture > workflow (and so they should not depend on the backend). As long as MySQL is a supported database, all tests should be run for this backend. If it can not comply, it is better to show the results instead of hiding them. There could also be other failures than the current Decimal precision. I for my share have less concerns to introduce backend specific doctests, clearly commenting the behavior of MySQL and not hiding issues for other backends. > > [1] http://hg.tryton.org/modules/account/rev/3f5a5a854341#l4.1 > > [2] http://tests.tryton.org/~test/mysql.html > > [3] > > http://hg.tryton.org/trytond/file/631515bc8c82/trytond/model/fields/numeric.py#l31 > > [4] http://docs.python.org/2/library/decimal.html#decimal.Decimal.normalize > -- Mathias Behrle MBSolutions Gilgenmatten 10 A D-79114 Freiburg Tel: +49(761)471023 Fax: +49(761)4770816 http://m9s.biz UStIdNr: DE 142009020 PGP/GnuPG key availabable from any keyserver, ID: 0x8405BBF6
signature.asc
Description: PGP signature
