Closed jordansread closed 7 years ago
Since it's not testable except for rendering, did you build any of the vignettes? I imagine we would see a change because the tcl.minor
default was 0.15 and should now be -0.25 if par('tcl')
remains unchanged.
I think having NA carry through to the final gsplot object makes sense to me. People can just look in the documentation to figure out what the actual value should be since it could change if pars change.
@lindsaycarr the tcl
is in the default config already as 0.3. I will keep that as-is. Working on merge conflicts for the vignettes now and that is coming w/ the next commit.
changes in the vignettes all seem to be due to the retina screen render. Note we don't have anything in the readme that calls out the tcl.minor, so we don't see changes there.
Doesn't it appear that the new fig changes are overplotting axes? Or is that just a visual trick?
Hard to say if overplotting or retina. I didn't change anything related to append vs replace for axis, so not sure why this would overplot if it didn't before this commit.
I added a change to the readme to force it to use the tcl.minor
default @lindsaycarr
for #437
I ended up doing this in a way that isn't testable except for the result of the render...
This is because I don't want the fraction of
par('tcl')
to be calculated until it is relevant - i.e., when the side is being rendered, so it is stuck into thedraw_axis
w/ the arg =NA
signifying to use this default. It is buried from the user, but the alternative solution that covers this and was discussed in #437 with the hierarchy is too complex for the value of this functionality. Additional thoughts welcome.