Hmm, so it sounds like there are three unrelated bugs here:
1) Crash in FRED when trying to use a list modifier in the SEXP tree in Events Editor.
2) Using a new container in FRED's Events Editor without saving the mission first causes a crash.
3) Multidimensionality (parsing the result from a container lookup/operation as a reference into another container) is broken.
For 1 and 2, I don't understand your steps to repro. I think there are some unstated/implied actions/steps that I'm not aware of. Keep in mind that I haven't used FRED since 2012. I also don't know how to use the Containers feature in FRED; I just know some things about how containers work under the hood.
As such, can you please provide more thorough steps to reproduce for 1 and 2?
I think I understand the repro steps for 3, but a more thorough list of steps wouldn't be a bad idea.
Re: bug 1, I noticed some code that got lost in the rebase in sexp_tree that's related to assigning operators, so maybe fixing that will help fix this.
Re: bug 2, this might be fixed now that I removed the "raw_data" thing from the edit containers dialog. All edits made in the dialog UI now affect the edit_sexp_containers variable (now called m_containers) directly. The edited containers still aren't saved to the global Sexp_containers until the user clicks OK, though.
I'd like to provide an updated build, but it's a bit tricky in that I'm still working out issues in my revised version of the edit containers dialog. As I mentioned on Discord, the UI isn't changed, mostly internals, plus a few small changes in user-facing behavior.
I'll see what fixes I have tonight and provide an experimental build that may or may not fix some of these things.
If you can provide more complete repro steps though that'd help a lot.