state equality check is too weak

[franzlst]

To check if two states are identical, their str() representation is compared.
This could be ambiguous. We should define a stronger method!

Why did we overwrite the __eq__ function anyways?

Open for discussion.

Originally created by @sebastian-brunner ([email protected]) at 2018-05-29 06:16:14+00:00 (moved from RMC internal repository)

[franzlst]

We should also keep
8c21340a022166e429
and
6d3804540690a479
in mind.

Originally created by @sebastian-brunner ([email protected]) at 2018-05-29 06:19:09+00:00 (moved from RMC internal repository)

[franzlst]

Ok some information that for all, pretty much surprising for me:
As you probably know:
a == b (1)
and
a is b (2)
are two different things.

(2) should be always used to check if the object are identical i.e. same reference
(1) is used to compare if two states are the same, e.g. after a copy.

a in list uses (2)

a in dict.values() uses (1)

Originally created by @sebastian-brunner ([email protected]) at 2018-05-29 09:06:49+00:00 (moved from RMC internal repository)

[franzlst]

The title is maybe missleading. After discussion with @sebastian-brunner we came to the conclusion that the __eq__ method which was implemented (before there were no proper hash-methods implemented) is simply wrong. Especially because it is the __str__ method what is wrong concerning the current use in __eq__.

We concluded that we will let the __eq__ method use the hash value which also is tested much better.

As mentioned by @sebastian-brunner he has had already a discussion with @franzlst. So an implementation could be started to fix the methods.

Originally created by @Rbelder at 2018-05-29 11:40:41+00:00 (moved from RMC internal repository)