sqlalchemy-bot

**Alessio Bogon ([@youtux](https://github.com/youtux)) wrote:** oh, it is described in the sessionmaker(), I see. Unfortunately, this trick would not work with autoflush=False.

**Michael Bayer ([@zzzeek](https://github.com/zzzeek)) wrote:** the way backrefs work is that they receive events as above and then actually mutate the collections on the other side. if you have the stomach...

**Michael Bayer ([@zzzeek](https://github.com/zzzeek)) wrote:** a hybrid property would be tricky because you'd have to create a SQL expression for that collection of collections also, I don't know how to do...

**Alessio Bogon ([@youtux](https://github.com/youtux)) wrote:** Also, I tried the solution you proposed with events, but in case the models is being initialized from scratch, object_session() will return None, so we have...

**Changes by Michael Bayer ([@zzzeek](https://github.com/zzzeek)):** * changed **title** from "Relationship using "composite secondary join" do n" to "add note on using events to sync two relationships"

**Changes by Michael Bayer ([@zzzeek](https://github.com/zzzeek)):** * added labels: **documentation**

**Changes by Michael Bayer ([@zzzeek](https://github.com/zzzeek)):** * set **milestone** to "1.x.xx"

**Michael Bayer ([@zzzeek](https://github.com/zzzeek)) wrote:** Issue #3990 was marked as a duplicate of this issue.

**Michael Bayer ([@zzzeek](https://github.com/zzzeek)) wrote:** test case for mutually dependent objects, if doing just an UPDATE, post_update should not be needed: ``` from sqlalchemy import * from sqlalchemy.ext.declarative import declarative_base from...

**Michael Bayer ([@zzzeek](https://github.com/zzzeek)) wrote:** as tihs would also fix #3496 I'm motivated to work on this, but I'm struggling. https://gerrit.sqlalchemy.org/427