rybern

#522 addressed items 3 and 4. More work on items 1 and 2 here might still be beneficial. There may well still be edge cases.

#522 addressed items 3 and 4. More work on items 1 and 2 here might still be beneficial.

#522 addressed item 3, and at least showed that DCE wasn't causing problems on the corpus in 4. I think it's likely that DCE could be more aggressive, I did...

I'd say that #522 addressed items 3 and 4. I did some substantial work on function inlining correctness in that patch, but I'm doing more work on items 1 and...

It looks like that's generated in [Statement_gen.ml](https://github.com/stan-dev/stanc3/blob/master/src/stan_math_backend/Statement_gen.ml#L58)

And it looks like the sequence for reading in a parameter is in [Transform_Mir.ml](https://github.com/stan-dev/stanc3/blob/master/src/stan_math_backend/Transform_Mir.ml#L189), where `decl` is the Decl constructor (that's turned into the declaration and assignment later in Statement_gen)...

Yeah, interesting question. One thing I should mention is that I did some [refactoring](https://github.com/stan-dev/stanc3/pull/561/files#diff-a455b46ddb916d08703b8e11c5054054) of those declaration rules in #561. Those changes should substantially reduce their redundancy, and certainly we...

#522 addressed items 3 and 4, I haven't payed more attention to it

Yeah, I figured - this really isn't time sensitive, since what's implemented now works, it's just might be a bit more complex and expensive than ideal.

Hi Jonah, I agree the warning doesn't make much sense for this model. It's happening because `x ~ normal(0, sigma);` is identified as a prior on `sigma`. That's happening because:...