Hi, You will find attached below three patches that are currently applied to the Debian package octave-statistics (upcoming version 1.1.0-1). They fix problems in the test suite for some functions.
The rationale and purpose of each patch is given in the respective patch header. Best,
Description: Fix normalise_distribution under Octave 3.6 Upstream applied (in r9016) a workaround for a regression in Octave core 3.4 (bug#34765 in savannah). In r9028, upstream made this workaround apply to Octave 3.4, but not to Octave 3.6, assuming that the bug has been fixed in Octave 3.6. But obviously this is not the case as of Octave 3.6.1, so we still need the workaround. Author: Sébastien Villemot <sebastien.ville...@ens.fr> Last-Update: 2012-04-13 --- This patch header follows DEP-3: http://dep.debian.net/deps/dep3/ --- a/inst/normalise_distribution.m +++ b/inst/normalise_distribution.m @@ -198,7 +198,7 @@ ## differently depending on octave version. This applies the fix for all ## 3.4 releases but it probably only appeared on 3.4.3 (can someone check?) ## See https://savannah.gnu.org/bugs/index.php?34765 - if (compare_versions (OCTAVE_VERSION, "3.4", "<") || compare_versions (OCTAVE_VERSION, "3.6", ">")) + if (compare_versions (OCTAVE_VERSION, "3.4", "<")) ## this is how it should work f_remap = @( k ) ( normal ( k ) ); normalised ( :, k ) = arrayfun ( f_remap, target_indices );
Description: Fix combnk with cell arrays
Without this patch, the following test fails on Octave 3.6:
%!test
%! c = combnk ({1, 2, 3}, 2);
%! assert (c, {1, 2; 1, 3; 2, 3});
Author: Sébastien Villemot <sebastien.ville...@ens.fr>
Last-Update: 2012-03-14
---
This patch header follows DEP-3: http://dep.debian.net/deps/dep3/
--- a/inst/combnk.m
+++ b/inst/combnk.m
@@ -51,7 +51,11 @@
else
## Process data
n = numel (data);
- retval = [];
+ if iscell (data)
+ retval = {};
+ else
+ retval = [];
+ endif
for j = 1:n
C = __combnk__ (data ((j+1):end), k-1);
C = cat (2, repmat (data (j), rows (C), 1), C);
Description: Clear last warning state in linkage test This patch clears the state of the last warning before running warning-based tests in the linkage function. This is necessary when we run the whole package testsuite in a single shot: if a previous test emits a warning, the test in linkage will fail. This may be the symptom of a bug in Octave, but I am not totally sure. Author: Sébastien Villemot <sebastien.ville...@ens.fr> Last-Update: 2012-03-14 --- This patch header follows DEP-3: http://dep.debian.net/deps/dep3/ --- a/inst/linkage.m +++ b/inst/linkage.m @@ -217,6 +217,7 @@ %!assert (cond (linkage (pdist (x), "complete")), 21.793345,t); %!assert (cond (linkage (pdist (x), "average")), 27.045012,t); %!assert (cond (linkage (pdist (x), "weighted")), 27.412889,t); +%!test lastwarn() % Clear last warning %!warning <monotonically> linkage (pdist (x), "centroid"); %!test warning off clustering %! assert (cond (linkage (pdist (x), "centroid")), 27.457477,t);
-- Sébastien Villemot Researcher in Economics & Debian Maintainer http://www.dynare.org/sebastien Phone: +33-1-40-77-84-04 - GPG Key: 4096R/381A7594
pgpxNTrm4B31i.pgp
Description: PGP signature
------------------------------------------------------------------------------ Virtualization & Cloud Management Using Capacity Planning Cloud computing makes use of virtualization - but cloud computing also focuses on allowing computing to be delivered as a service. http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________ Octave-dev mailing list Octave-dev@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/octave-dev
