|
|
|
last edited 17 years ago |
Edit detail for #295 PERM movedPoints wrong revision 1 of 1
| 1 | ||
|
Editor:
Time: 2007/11/17 22:22:39 GMT-8 |
||
| Note: fixed in Patch 50 | ||
changed: - 'movedPoints\$Permutation' computes the set of moved points incorrectly: \begin{axiom} p := coerceImages([1,2,3])$PERM INT movedPoints p \end{axiom} Note that 'p' really is the identity permutation, so the set of moved points should be empty... This bug was obviously known to the original authors, we read:: movedPoints p == construct p.1 --check on fixed points !! but they never bothered to fix it. In the line above, 'p.1' simply is the preimage of the permutation 'p', so there are two ways to fix the bug: * remove fixed points from preimage and image, always when permutations are generated * check on fixed points only in movedPoints This patch chooses the first option, which seems to be also what the original authors had in mind. Martin From kratt6 Sat Jun 3 12:50:45 -0500 2006 From: kratt6 Date: Sat, 03 Jun 2006 12:50:45 -0500 Subject: Message-ID: <20060603125045-0500@wiki.axiom-developer.org> Unfortunately, MathAction doesn't seem to like my patch: I can't upload it. Well it's on the mailing list anyway... Martin <a href="perm.spad.pamphlet.patch">perm.spad.pamphlet.patch</a> From billpage Sat Jun 3 16:03:21 -0500 2006 From: billpage Date: Sat, 03 Jun 2006 16:03:21 -0500 Subject: patch uploaded Message-ID: <20060603160321-0500@wiki.axiom-developer.org> The previous attempt to upload failed because of some over active anti-spam checking. It works now. From kratt6 Fri Oct 27 03:06:56 -0500 2006 From: kratt6 Date: Fri, 27 Oct 2006 03:06:56 -0500 Subject: fixed in Patch 50 Message-ID: <20061027030656-0500@wiki.axiom-developer.org> Status: open => closed
movedPoints$Permutation computes the set of moved points incorrectly:
fricas
(1) -> p := coerceImages([1,2, 3])$PERM INT
 | (1) |
Type: Permutation(Integer)
fricas
movedPoints p
 | (2) |
Type: Set(Integer)
Note that p really is the identity permutation, so the set of moved points should be empty...
This bug was obviously known to the original authors, we read:
movedPoints p == construct p.1 --check on fixed points !!
but they never bothered to fix it. In the line above, p.1 simply is the preimage of the permutation p, so there
are two ways to fix the bug:
- remove fixed points from preimage and image, always when permutations are generated
- check on fixed points only in movedPoints
This patch chooses the first option, which seems to be also what the original authors had in mind.
Martin
... --kratt6, Sat, 03 Jun 2006 12:50:45 -0500 reply
Unfortunately, MathAction doesn't seem to like my patch: I can't upload it. Well it's on the mailing list anyway...
Martin
The previous attempt to upload failed because of some over active anti-spam checking. It works now. Status: open => closed