Skip to content

Commit e853f22

Browse files
author
ripley
committed
RNG changes
update solve section git-svn-id: https://svn.r-project.org/R-dev-web/trunk@1256 c52295ea-58df-0310-926a-d16021944841
1 parent 7f5057f commit e853f22

File tree

1 file changed

+16
-6
lines changed

1 file changed

+16
-6
lines changed

170update.txt

+16-6
Original file line numberDiff line numberDiff line change
@@ -55,6 +55,11 @@ only distinction between the class given by data.class() and that
5555
given by class() and used for dispatch by UseMethod is that
5656
data.class() reports "numeric" for integer vectors.)
5757

58+
[If you want code to be compatible with < 1.7.0 versions of R, use
59+
60+
if(is.null(attr(x, "class"))) attr(x, "class") <- data.class(x)
61+
62+
]
5863

5964
Unless the package makes use of setMethod (methods), loading can be
6065
speeded up by having an object called .noGenerics (with any value) in
@@ -77,12 +82,15 @@ documentation is in `Writing R Extensions'.
7782
Other changes
7883
-------------
7984

85+
The default random number generators have been changed.
86+
87+
8088
solve.default() now uses LAPACK by default, and works as documented,
81-
on square matrices `a'. Some packages have been using it for QR
82-
decompositions (which was documented in S but not in R: use qr.solve)
83-
or even for length-1 vectors (just divide). QR decompositions are
84-
accepted with a warning. One package was using it for over-determined
85-
rectangular systems: use qr() and qr.solve() for those.
89+
on _square matrices_ `a'. Some packages have been using it for QR
90+
decompositions (which was documented in S but not in R: use solve()
91+
but not solve.default()) or even for length-1 vectors (just divide).
92+
One package was using it for over-determined rectangular systems: use
93+
qr() and qr.solve() for those.
8694

8795

8896
Packages should not execute code with side effects in their .R files.
@@ -91,13 +99,15 @@ files, please move them to .First.lib. This is necessary for the
9199
experiments with saving objects in databases, a feature planned for
92100
1.8.x.
93101

102+
94103
A full double-precision LAPACK library Rlapack.{so,sl,dll,dynlib} is
95104
installed with R, unless a high-performance LAPACK library was found
96105
at configure time. If your package makes use of LAPACK, please
97106
convert it to use this via
98107

99108
$(LAPACK_LIBS) $(BLAS_LIBS)
100109

101-
in PKG_LIBS.
110+
in PKG_LIBS in Makevars.
111+
102112

103113

0 commit comments

Comments
 (0)