Is control flow represented correctly?

[MKLau]

From B Lerner in #61:

I wonder if want to distinguish between definitely-depends-on and might-depend-on (and similarly for produce). The distinction happens if the code contains an if-statement or other control construct. For example:

if (x > 0) df <- read.csv("file1.csv")
else df <- read.csv("file2.csv")

If we are not going inside if-statements, the simplest thing would be to claim that the output depended on both files even though that is not correct.

[MKLau]

My sense here is that we don't need an ambiguous designation of "might" if we have df depend on the control statement if ... else which itself uses file1.csv and file2.csv.