Klaus Greff

@JarnoRFB There was some change in the way github handles webhooks which caused the docs build to fail. But that should be fixed now.

Hi Evgeny, unfortunately the answer to both is "not yet". In an earlier version of the current release I had a prototype `Assistant` that could do an early version of...

Yes, this is an important and recurring usecase. We should support continuing experiments, either from aborted or even finished runs. There are some details that need to be figured out...

Wow, thanks @bouthilx for jumping in with such a detailed proposal! I really like the idea of having standalone final versions with references to parents. That would allow a lot...

Not sure I fully understand, but it does sound like a minor extension to the `SQLObserver`. So if you prepare a PR, I'll have a look and will likely merge...

This is a recurring non-deterministic issue, and honestly, I have no idea why it happens. I'm thoroughly flushing stdout and stderr [here](https://github.com/IDSIA/sacred/blob/master/sacred/stdout_capturing.py#L128). And in general I've tried to be very...

Several issues with stdout capturing have been resolved recently, so I'd be surprised if the failure rate is still 40%. But there remains a problem somewhere. (see also #289)

I agree, that the current way is too inflexible. My only concern is that storing results as JSON strings would loose the support for sorting and range querying based on...

SQL alchemy does [support the JSON datatype](http://docs.sqlalchemy.org/en/latest/core/type_basics.html#sqlalchemy.types.JSON). I don't have enough experience with SQL to know how easy this datatype is to work with. But in any case it does...

Ohh wow I see. Many thanks for tracking this down! Your "quick-and-dirty" fix seems appropriate, since we don't want to convert `\r` to `\n` (basically nobody uses `\r` as a...