Hans Moritz Günther

I did say in https://github.com/sherpa/sherpa/pull/1477#issuecomment-1228908661 that I think we should stick with this or at least do something very similar for writing the independent axis. Let me expand on this:...

That would also mean that `set_indep(get_indep())` would be a no-op. Should be finish discussing this before I re-review the entire PR? It's now > 15000 lines of code changed and...

So, for the 2D case, we have no choice but clear the filter, because we just don't have the ability to keep it. Does that means we want to clear...

@anetasie What do you think? See https://github.com/sherpa/sherpa/pull/1477#issuecomment-1245382892 and the following few comments. Should the filter always be cleared when the independent axis is changed? @DougBurke Even if we decide to...

Otherwise, I think the code changes are in good shape.

I like the old behavior for 1 D datasets (as I've argued above), but not the non-sense that happens in 2D datsets. The new warning says what happens. So, here...

Congratulations! This PR also brings us a significant reduction of the number of open issues - that's something to be proud of @DougBurke !

Is this still an issue?

@anetasie this ticket is ~5 sherpa versions old and even back in 2017 @DougBurke failed to re-produce the issue. Can you find an example that still fails today? ("improve datastack"...

One might say that changing the quality automatically removes the filter. In practice, one either uses the quality (and calls "ignore_bad" right after loading the data) or not, it's not...