Hey guys, Reading through the OSL manual, on page 20 there's this paragraph: """Functions may be polymorphic. That is, multiple functions may be defined to have the same name, as long as they have

I believe you are correct. Polymorphism should refer to types, what I was talking about is really "overloading." -- lg -- Larry Gritz l...@...

I think one of the reasons for this extensive explanation was meant to make the point that functions cannot be overloaded only by return type, in contrast to RSL :) This, along with many other

RSL and OSL both are able to overload functions by return type. So, do the terminology purists out there think that this makes it "polymorphic" or merely "overloaded"? -- lg -- Larry

Reading these two http://en.wikipedia.org/wiki/Type_polymorphism http://www.haskell.org/haskellwiki/Ad-hoc_polymorphism suggests that what OSL provides is ad-hoc polymorphism, or otherwise known as

"polymorphic" appears exactly once in the docs, I have changed to "overloaded." No biggie. -- Larry Gritz l...@...

Reviewers: , Description: Extend the inf/nan checking to also check (and print specific messages for) derivatives. Also, fixed a minor buglet that, for arrays, would only have checked the first

LGTM, except for the previously existing bug http://codereview.appspot.com/199063/diff/1/3 File src/liboslexec/exec.cpp

That works for me. Maybe throw in Const too? (just in case of some weird oso parsing regression?) http://codereview.appspot.com/199063/show

Sure, I guess I could. I figured that the constants from the oso file were probably beyond reproach, but since it doesn't cost anything, will do. -- lg -- Larry Gritz l...@...

Reviewers: , Description: In several string ops, I used a local var that just told whether any of the arguments were varying, to determine whether to assign to just the first element of Result, or

LGTM! You might want to hunt for this pattern in a few other places. get/setmessage come to mind http://codereview.appspot.com/198076/show

I'm going to commit this ASAP and then hunt for other possible occurrences. http://codereview.appspot.com/198076/show

Reviewers: , Description: In the same way as my last review for string ops, if getmessage had all uniform args but was called from within a varying conditional, it could fill in only the first value

LGTM But I still think there are additional corner cases to both setmessage and getmessage that might be broken. Overall the logic in these ops is quite hard to follow ... is there anything we could

Which other cases do you think are broken? I can try to simplify, but it's a fairly complicated set of moving parts. I'll see what I can do. -- lg -- Larry Gritz l...@...

Just wondering if you have thought about binding shader parameters to geometry like you can with primitive variables in RenderMan. Specifically - if you are going to do something like this - it

Thanks for the email, Simon! Yes, OSL already has primitive variables. I agree that it would be handy to have a built-in function that tells you if the particular value is the default, from a

OK, I've done some really interesting investigation the runflags issue. Chris Foster was very inspiring with his test program and span idea (which we've been debating versus Chris Kulla's similar

Next installment of "everything you wanted to know about runflags but were afraid to ask." So after sampling representative runflags from a frame with production shaders and assets, next I wrote a