- Added
as.gtable()S3 method (#97). - Add
clipargument togtable_col()andgtable_row()(#56) - Indexing a gtable with
NAwill now insert a zero-dimension row/column at the position of theNA-index (#13)
- Fixed partial matching issue when constructing viewport in
gtable()(#94) - General upkeep
- Fix package doc links
- Specify minimum rlang version
- General upkeep
-
Re-documented to fix HTML issues in
.Rd. -
gtable has been re-licensed as MIT (#85).
-
Made a range of internal changes to increase performance of gtable construction, these include:
- Use more performant
data.frameconstructor . - Treat layout data.frame as list when indexing and modifying it.
- Use length of
widthsandheightsfields instead ofncol()andnrow()internally. - Substitute
stopifnot(...)withif(!...) stop().
- Use more performant
-
Better documentation, including a new README, a vignette on performance profiling and a pkgdown site.
-
New logo
-
It is now an error to index into a gtable with non-increasing indices.
-
Dimnames are now inherited from the grobs data in
gtable_col(),gtable_row(), andgtable_matrix() -
gtable_trimnow works with empty gtables -
gtable_filternow has an invert argument to remove grops matching a name.
-
Switch from
preDrawDetails()andpostDrawDetails()methods tomakeContent()andmakeContext()methods (@pmur002, #50). This is a better approach facilitiated by changes in grid. Learn more at https://journal.r-project.org/archive/2013-2/murrell.pdf. -
Added a
NEWS.mdfile to track changes to the package. -
Partial argument matches have been fixed.
-
Import grid instead of depending on it.
print.gtablenow prints the z order of the grobs, and it no longer sort the names by z order. Previously, the layout names were sorted by z order, but the grobs weren't. This resulted in a mismatch between the names and the grobs. It's better to not sort by z by default, since that doesn't match how indexing works. Thezsortoption allows the output to be sorted by z.