Unknown command '_'. Skipping.

[kalmard0]

While working on my (fairly standard) ink script, the inkcpp compiler keeps outputting the message from the title.

After some research, I found that inkle/ink@ee008e5 (in 2017!) added the _ command as a new name for Negate, calling ~ Invert.

I don't fully understand the distinctions of various commands, but I assume Negate means the unary multiply by -1 operator.

Inkcpp still calls ~ Negate and does not handle _ at all. Surprisingly, so far I haven't noticed actual errors related to this while running my ink script, but maybe I just wasn't paying attention.

I'm also surprised that this hasn't been noticed yet, since the command has been part of ink for a very long time. Is there a reason this is not part of inkcpp? I would be happy to give it a shot and add it, but I'm not sure what the exact steps are to do it correctly.

[kalmard0]

In the meantime I realized that my specific code that triggered this behavior had an issue: I used the
{ -binary_expression: result } format for single-line conditionals, instead of { binary_expression: result }.

So this question is not so urgent for me, but I think it may be still generally valid.

[JBenda]

Funny, never encountered it. We implemented ink pp based on their documentation of the json format and ink proof.

At some point we implemented both and through we were finished.

I will look into it soon about the implications for the engine

[kalmard0]

Seems like this is a genuine issue. To reproduce, use this ink file:

VAR boolvar = true
{ -boolvar: first boolvar }
{
  -boolvar: second boolvar
}
-> END

Output from ink editor:

first boolvar
second boolvar
End of story

Output from inkcpp:

ERROR: Unknown command '_'. Skipping.
Cancelling play mode. Errors detected in compilation

[JBenda]

VAR boolvar = true
{boolvar}
{-boolvar}
{ -boolvar: first boolvar }
{
  -boolvar: second boolvar
}
-> END

I extended the example to the above to show a funny side effect, negate is not implemented for boolean, therefore, a -bool is a -static_cast<int>(bool) therefore -true=-1 kind a funny.

Expected output:

true
-1
first boolvar
seond boolvar

Now we support invert and negate, but invert appears to be unused currently. We will see what will happen.

The change is also in #110 since it was quite minor.

[JBenda]

#110