| NEWS | R Documentation |
Importing OPJ files containing unsupported objects with
tree = TRUE now avoids an error in the R code.
When decoding strings in the OPJ files, the target encoding is now set to UTF-8, not the native encoding of the R session. This makes it possible to decode strings not representable in the R session encoding, i.e. Windows-1252 files on a Windows-1251 machine and vice versa. Realistically, this is only important on Windows versions older than Windows 10 build 1903.
A spurious MinGW warning concerning the use of strftime
with format strings %F and %T has been worked around
by patching out the call using the preprocessor. (The call wasn't
reachable from R code, and the UCRT library currently used by R is
documented to support these format strings.)
Updated bundled liborigin, fixing rare memory access
issues.
A compilation error in ‘tree.hh’ has been fixed by adding
an #include to provide the missing global type definition.
Update bundled liborigin, fixing issues related to newer
project files.
Update liborigin from 3.0.0 with patches to the
officially released version 3.0.1, fixing some correctness
issues.
Use the ‘comment’ attribute (as in comment()
function) for spreadsheet imports in addition to ‘comments’
attribute (as before)
Pass the ... arguments from read.opj to
as.data.frame, making it possible to set
stringsAsFactors and other parameters
New tree parameter to return the objects in a tree
structure instead of a flat list
Prevent truncation of long tables when importing large files
Fix a memory leak possibly resulting from misuse of
tree.hh library
Decode strings using Riconv interface instead of
Rcpp::Function call, making it much faster
Added support for multi-spreadsheets, matrices and notes
Correctly decode byte strings read from the file
Fix undefined variable access found by UBSan