Conversation
| stop("Unknown entry in getOption('repr.plot.backends'): ", o) | ||
| ) | ||
| stop( | ||
| "No plotting devide found in getOption('repr.plot.backends'): ", |
There was a problem hiding this comment.
Type: devide should be device.
Since this is an message intended for end users: maybe it could be even more explicit and say something like
stop(
"None of the backends specified in getOption('repr.plot.backends'): ",
toString(backends):
"seems to be available."
"\n Did you forget ..."
"Please specify a valid backend using 'options(repr.plot.backends = c(
)
(or, instead of available, you could say 'working' or whatever you think is the most accurate).
| png(tf, width, height, 'in', pointsize, bg, res, antialias = antialias) | ||
|
|
||
| dev.cb <- function(tf) get_device( | ||
| ragg = ragg::agg_png(tf, width, height, 'in', pointsize, bg, res), # scaling, bitsize |
There was a problem hiding this comment.
Looking at the code, at first I would have expected this to cause issues if the ragg package was not present on the end-user system, since I'd expect the namespace to not be known. Previously, for Cairo, you had it wrapped in if-else statemtents, so then the Cairo namespace would clearly only be called if the is_cairo_installed() returned true.
However, I initially didn't have ragg, and this line didn't produce any issues. I'm no R expert, so just want to check: are you sure this will work ok, even if some of the backends are not installed? And which function now imports this namespace? Before, I believe that was done by the requireNamespace call that was done by e.g. is_cairo_installed(), but now, it's not clear to me.
2 participants
Fixes IRkernel/IRkernel#604 … maybe