[westoncb]

I hear this talked about quite a bit, but it still doesn't quite make sense to me why it's such a big deal: I know this situation where debugging-fix attempt loop is far more efficient if you don't lose state, but it's nowhere near a pervasive situation: when it comes up, I can generally write something quickly to re-create the necessary parts of the state automatically.

If I was doing that every day, or even every week, I could see it being a huge advantage to incorporate a solution in the language—but at least for me it just doesn't appear to be an obstacle that often (incidentally my background was initially in game programming, too).

[jedimastert]

> I can generally write something quickly to re-create the necessary parts of the state automatically.

I think this is where the trip up is. In a game, for example, that state is often exceptionally complex and getting everything right back to where you came from is usually only possible running the full game and recreating the same state.

[westoncb]

It just doesn’t come up that often though.

In most cases it’s not hard to narrow down which aspects of state are relevant; you really only need to preserve everything for rare exceedingly subtle/deep bugs, which is a special case not a general usage kinda thing—and yet this feature is often discussed as revolutionary for programming in general.

[edgyquant]

I wonder what it is that prevents e.g. python from having a similar “persistent state in development” tool/repl? I’ve never really used lisp other than a MAL I did a few years ago for kicks so I’m not well versed in its dev tooling.

[twobitshifter]

https://www.libhunt.com/posts/72311-reloadr-hot-code-reloadi...

[Y_Y]

You can definitely do something similar with e.g. ipdb

  from ipdb import launch_ipdb_on_exception
  with launch_ipdb_on_exception():
    foo()