ValueError: microsecond must be in 0..999999

Hello, all

I’ve got a new error after the v2.0.0 milestone, likely relating to #172. Here’s the error text:

  File "/Users/space/Documents/Programming/dt-blog-platform/api-server/db/objs.py", line 268, in commit
    self.updatedon = datetime.datetime.utcnow()
  File "/var/root/.local/share/virtualenvs/api-server-DI-zCP9m/lib/python3.6/site-packages/umongo/document.py", line 295, in __setattr__
    self._data.set(name, value, to_raise=AttributeError)
  File "/var/root/.local/share/virtualenvs/api-server-DI-zCP9m/lib/python3.6/site-packages/umongo/data_proxy.py", line 152, in set
    value = field._deserialize(value, name, None)
  File "/var/root/.local/share/virtualenvs/api-server-DI-zCP9m/lib/python3.6/site-packages/umongo/fields.py", line 245, in _deserialize
    return ret.replace(microsecond=round(ret.microsecond, -3))
ValueError: microsecond must be in 0..999999

My guess is something funky is going on either with my system, or maybe the round isn’t working properly? Not too sure, but this should be a simple one.

The way I did this is set a StrictDateTimeField equal to datetime.datetime.utcnow() in an overriden commit method. Ex:

@instance.register
class Foo(Document):
    name = fields.StringField()
    updatedon = fields.StrictDateTimeField()

    def commit(self, *args, **kwargs):
        self.updatedon = datetime.datetime.utcnow()
        return super(Foo, self).commit(*args, **kwargs)

This error happens when calling the commit method on the object.

Thanks in advance!

[enugentdt]

Can do. I’ll have a report for you tomorrow morning EDT

Thanks, Eamonn

Sent from my iPhone

On Apr 7, 2019, at 6:36 PM, Jérôme Lafréchoux [e-mail] wrote:

Thanks for reporting.

This is due to millisecond being rounded to 1000000. In this case, second should be incremented and millisecond set to 0.

This is my mistake. It already happened to me in another context and I forget about it when submitting #172. Sorry about that.

I submitted #190 to fix this. @enugentdt, do you think you could check if it solves your problem?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[lafrech]

I prefer to rely on comprehensive testing rather than release into the wild until someone complains about a broken app. But frankly, I’m confident about the patch.