Release 0.3-2:
  * Minor change for CRAN.

Release 0.3-1:
  * Fixed compiler check in autoconf.
  * Added iparam2stage and others to internal LAPACK build.
  * Fixed remaining non-void declared parameterless functions.
  * Fixed link flags in system BLAS/LAPACK configure tests.
  * Updated citation file to use bibentry().

Release 0.3-0:
  * Updated internals to LAPACK 3.10.0.
  * Fixed LAPACK bug on gcc-12.
  * Package does a better job of searching for single precision BLAS/LAPACK
  symbols on the system on *NIX systems.

Release 0.2-6:
  * Fixed build for clang 13.

Release 0.2-5:
  * Fixed bug in norm() with type=="F" (max modulus).

Release 0.2-4:
  * Fix codes warned by "-Wargument-mismatch" because gcc10 gives warning/error
  with "-fallow-argument-mismatch" or not. The file includes
  "src/lapack/slapack3.f".

Release 0.2-3:
  * Renamed 'len_t' in float/float32.h to 'float_len_t'.
  * Fixed several compiler warnings.
  * Fixed header issue in float/slapack.h that affects C++.
  * Changed F77 to FC in Makevars.win for R >= 3.6.0.

Release 0.2-2:
  * Added sign().
  * Fixed bug with linking on OS X.
  * Fixed some minor (mostly harmless) internal issues.

Release 0.2-1:
  * Better linking/flags.
  * Use static library on Windows.
  * Add support for non-logical center/scale arguments in scale().
  * Fixed protection issue flagged by rchk in R_scale_spm().

Release 0.2-0:
  * Fixed protect stack imbalance in R_init_NA().
  * Hid 'restrict' keyword from C++ in inst/include headers.
  * Fixed control logic in various [ calls.
  * Created float:::ldflags() for dynamic linking.
  * Added some skeleton support for names+row/col-names.
  * Fixed performance bug with [.
  * Fixed NA overflow bug with length() for large matrices.
  * Fixed str() printing bug for matrices.
  * Added rep() method.
  * Added as.numeric(), as.double() and as.integer() methods.
  * Added float() (like double(), integer(), and numeric()).
  * Added build fix for some clang/OpenMP configurations.

Release 0.1-1:
  * Added str() method.
  * Added replacer [ method.
  * Remove \dontrun{} blocks from manual.

Release 0.1-0:
  * Added fl() and dbl() casters.
  * Added random constructors flrunif(), flrnorm(), and flrand().
  * Added backsolve() and forwardsolve() methods.
  * Added binary arithmetic methods.
  * Added rbind() and cbind() methods.
  * Added [ method.
  * Added c() method.
  * Added chol() and chol2inv() methods.
  * Added colSums(), rowSums(), colMeans(), and rowMeans() methods.
  * Added rcond() method.
  * Added crossprod() and tcrossprod() methods.
  * Added diag() method.
  * Added eigen() method (symmetric case only).
  * Added min(), max(), which.min(), and which.max() methods.
  * Added isSymmetric() method.
  * Added math methods.
  * Added %*% method.
  * Added is.na() and na.omit() methods.
  * Added norm() method.
  * Added qr(), qr.Q(), qr.R(), qr.qy(), and qr.qty() methods.
  * Added scale() method.
  * Added solve() method.
  * Added sum() method.
  * Added svd() and La.svd() methods.
  * Added sweep() method (+, -, *, and / only).
  * Added t() method.
  * Added float32() constructor for developers.