odd behaviour when colouring based on a numeric vector

[tim-salabim]

I encountered some rather odd behaviour when using colur_values_rgb on a numeric vector. The vector gets modified by reference, see below.

# integer works as expected
set.seed(23)
v = sample.int(100, 10)
v
# [1] 58 23 33 69 79 41 91 94 78 93
clrs = colour_values_rgb(v, include_alpha = FALSE) / 255
v
# [1] 58 23 33 69 79 41 91 94 78 93

# numeric gets mumbled
set.seed(23)
v = as.numeric(sample.int(100, 10))
v
# [1] 58 23 33 69 79 41 91 94 78 93
clrs = colour_values_rgb(v, include_alpha = FALSE) / 255
v
# [1] 0.4929577 0.0000000 0.1408451 0.6478873 0.7887324 0.2535211 0.9577465 1.0000000 0.7746479 0.9859155

Any ideas?
The colours are identical, btw.

[SymbolixAU]

Yes, that's unexpected. The numeric v is returned scaled [0,1] , so I might need to put a clone() in there somewhere.

But that doesn't explain why the integer works though.

I'll investigate tomorrow

[SymbolixAU]

Ok, a bit of digging,

All my Rcpp method signatures for "numbers" are Rcpp::NumericVector. So, when passing in an integer vector (from R), it has to be coerced to a NumericVector, thus a copy is made, and hence the difference between the integer and numeric results.

References:

• https://stackoverflow.com/a/38104905/5977215
• https://stackoverflow.com/a/11300707/5977215

Fix is incoming

[tim-salabim]

Thanks! Just tested and behaves as expected now.