Added test for ntiles. Fixed q2c parsing of integer/double subcommands

[pspp-builds.git] / src / ChangeLog

diff --git a/src/ChangeLog b/src/ChangeLog
index e8850ced7eaa8a8da127bdb319f5eab8dbee0f03..69c756d48587955f95709b3ad627a9b977e1f45d 100644 (file)
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,32 @@
+Mon Nov  1 12:46:17 WST 2004 John Darrington <john@darrington.wattle.id.au>
+
+       * q2c.c frequencies.q set.q t-test.q  Fixed the q2c parsing of DBL 
+       subcommand types.  Changed frequencies.q to use it rather then the 
+       custom parser.  Dealt with the consequences.  Added a test for NTILES
+       subcommand of frequencies.
+
+Sat Oct 30 09:16:29 WST 2004 John Darrington <john@darrington.wattle.id.au>
+
+       * oneway.q   Fixed up the behaviour when given missing values
+
+       * levene.c oneway.q Fixed a buglet with the levene statistic and
+       incorporated the levene test into the oneway command.
+
+        * group.h  t-test.q  Moved the CMP_EQ and CMP_LE symbols out of 
+       global scope, since they're only relevant to T-TEST
+
+Fri Oct 29 17:39:03 WST 2004 John Darrington <john@darrington.wattle.id.au>
+
+        * group.c group.h group_proc.h levene.c oneway.q t-test.q
+
+       Made the t-test more consistent
+       with the way it handles groups.  That is, it now uses a hash instead
+       of an array of 2.  Also, made the levene.c file independent of the 
+       implementation of the t-test.  So now levene should be fine for both
+       t-test and anova.
+
+       * Added an oneway.q file for one way anova
+
 Wed Jun  2 22:08:02 2004  Ben Pfaff  <blp@gnu.org>
 
        * descript.c: (cmd_descriptives) Remove harmless but bogus test in