For seven years and two days, the algorithm for ascent and descent has had an off-by-one error, and started at the wrong level.
Also, the message for failed descent has apparently never been right.
I figured I'd follow Java's conventions (as I see them) and have a body for the loop, so instead of incrementing the counter within the loop condition, I refactored the loop and put in a break statement. I prefer the shorter version, but POLA wins in a foreign code base.
Added the option format to config files (config.yml + worldedit.properties). By default, current date and time are displayed.
The formatting function is implemented by analogy with the class java.util.logging.SimpleFormatter
* Sign texts not starting with { are never valid json objects, so don't try to parse
* Else try to parse as json, but fall back to read sign text as string if not parseable
For example, //snow didn't work on Oak leaves while it did work as you would expect on Dark Oak leaves. Fixed by adding leaves2 to translucent block list and checking if the block to put snow on is a leaf block when executing //snow.
Fix EditSession being created for commands that don't need it
Fix error from schematic list (when there are files in sub directories)
Fix dispatcher redundantly checking a huge number of permissions
Due to odd rounding and math issues, using //center on a region with an edge of even
length in quadrants with negative coordinates would only fill one of the two center blocks.
Thanks to Griffrez and up201406036 for the original PR.
Fixes WORDLEDIT-2986.
For example, the getChunks method returns a Set<Vector2D> which is
actually comprised of BlockVector2D, so using set.contains(some vector)
will always return false.
There's unlikely to be multiple entities on the same block, or above
y=256, so using the int coords is better.
Java 8's Kernel statically loads the AWT library, which
can apparently cause some issues on some systems, for some reason.
Either way, there is no need to load AWT if it's not needed (Java pls).
