TimeStamp __repr__ returns a double quoted string

[do3cc]

The test creates the string the same way so it does not fail. Relevant commits: https://github.com/zopefoundation/persistent/commit/db732a81b0ce4e69bceee1014843b24a21b80fe7 https://github.com/zopefoundation/persistent/commit/e692af8281466fa309aae9273864039dcb287383

and a traceback in the wild where it bit me: https://github.com/ploneintranet/ploneintranet.todo/commit/7f215f69f945b692ae1004f7533c04311c9709fa#commitcomment-8752965

[mgedmin]

I don't understand what the problem is. Current persistent master, on Python 2:

>>> from persistent.TimeStamp import TimeStamp
>>> print repr(TimeStamp(2014, 2, 3))
'\x03\xa4\x8a\xa0\x00\x00\x00\x00'

and on Python 3:

>>> from persistent.TimeStamp import TimeStamp
>>> print(repr(TimeStamp(2014, 2, 3)))
b'\x03\xa4\x8a\xa0\x00\x00\x00\x00'

I don't see any double-quoted strings here.

Also, what does that traceback have to do with this issue? How is __repr__ involved in that ValueError?

(Also, hint: if you edit the comment and wrap the traceback with ..., it'll be more readable.)

[do3cc]

At first I could not reproduce you calling arg, I'll create a different bug report for this. but going back to your example, you print the representation, and it starts with a "'" that is the inner quote. The traceback does not help in understand how this is a problem... The problem occurs more or less here:

https://github.com/zopefoundation/ZODB/blob/master/src/ZODB/utils.py#L147 Just that I am looking at ZODB3 3.10.5 In ZODB3 3.10.5 there is no ts.raw() but a ts

So my problem is caused by a combination of incompatible package versions :-(

[do3cc]

Since ZODB is not a dependency on persistent, there is no way to mark this package incompatible with ZODB 3.10. I close this issue

[mgedmin]

Ah, the return value of __repr__ changed in some version of persistent (but this is not mentioned in the changelog), and that makes older versions of ZODB fail when used with new versions of persistent?

At the very least we should add a changelog notice in persistent, warning people about this (and mention the exact exception this incompatibility causes so it'll come up in Google searches).

(FWIW I don't have the energy to do this.)

[tseaver]

Older versions of ZODB bundle their own version of persistent.

[mgedmin]

Hm. So if the user ends up with a mixture of packages that include an older ZODB and a newer persistent, it's a matter of chance which persistent ends up first on sys.path?

[do3cc]

Yes, I think so

[tseaver]

Just don't install both persistent and ZODB3: persistent should only be used with the separately released ZODB (note the trailing "3" on the "omnibus" release).