facFirstWeyl computes wrong result in Master (but not in SW)...

[Oleksandr]

The test Long/ncfactor_lee_l.tst from the SW' test-suite:

ring R = 0,(x,d),dp; def r = nc_algebra(1,1);setring(r); 
poly L = 10x5d4+26x4d5+47x5d2-97x4d3;facFirstWeyl(L);

Current Master Singular returns wrong output (according to testNCfac and entry [2] is plainly wrong!):

[1]:
   [1]:
      1
   [2]:
      10x4d2+26x3d3+47x4-117x3d-78x2d2+117x2+156xd-156
   [3]:
      x
   [4]:
      d
   [5]:
      d
[2]:
   [1]:
      1
   [2]:
      x
   [3]:
      x
   [4]:
      d
   [5]:
      d
[3]:
   [1]:
      1
   [2]:
      xd-3
   [3]:
      10x3d2+26x2d3+47x3-117x2d-52xd2+52d
   [4]:
      xd-1
[4]:
   [1]:
      1
   [2]:
      xd-3
   [3]:
      x
   [4]:
      x
[5]:
   [1]:
      1
   [2]:
      10x4d2+26x3d3+47x4-117x3d-78x2d2+117x2+156xd-156
   [3]:
      d
   [4]:
      xd-1
[6]:
   [1]:
      1
   [2]:
      xd-3
   [3]:
      x
   [4]:
      x
   [5]:
      x
   [6]:
      10xd3+26d4+47xd-107d2-47
[7]:
   [1]:
      1
   [2]:
      x
   [3]:
      xd-2
   [4]:
      x
   [5]:
      x
   [6]:
      10xd3+26d4+47xd-107d2-47
[8]:
   [1]:
      1
   [2]:
      x
   [3]:
      x
   [4]:
      xd-1
   [5]:
      x
   [6]:
      10xd3+26d4+47xd-107d2-47
[9]:
   [1]:
      1
   [2]:
      x
   [3]:
      x
   [4]:
      x
   [5]:
      x
   [6]:
      d
   [7]:
      10xd3+26d4+47xd-107d2-47
[10]:
   [1]:
      1
   [2]:
      x
   [3]:
      x
   [4]:
      x
   [5]:
      x
   [6]:
      10xd2+26d3+47x-97d
   [7]:
      d
[11]:
   [1]:
      1
   [2]:
      x
   [3]:
      x
   [4]:
      x
   [5]:
      x
   [6]:
      10xd2+26d3+47x-97d
   [7]:
      d
   [8]:
      d

Master Build GIT-ID is 79c9b876326058a6d1c21eccb9b6080c1ae3a93e (with a BS and FLINT related fix). It was build as follows:

Singular for x86_64-Linux version 3-1-7 (3170)  Aug 27 2013 16:55:57
with
	factory(@(#) factoryVersion = 3.1.6),libfac(3.1.6,December 2012),
	GMP(5.1),NTL(5.5.2),FLINT(2.3),64bit,static readline,Plural,DBM,
	dynamic modules,dynamic p_Procs,OM_CHECK=0,OM_TRACK=0,random=1378012097
	CC= gcc -O2 -w -fomit-frame-pointer -pipe -DNDEBUG -DOM_NDEBUG -Dx86_64_Linux -DHAVE_CONFIG_H,
	CXX= g++ -O2 -w -fomit-frame-pointer -fno-rtti -I.. -I/GITHUB/w/MASTER -pipe -DNDEBUG -DOM_NDEBUG -Dx86_64_Linux -DHAVE_CONFIG_H (4.8.1 20130725 (prerelease))

[Oleksandr]

Note that the current SW gives short and correct (according to testNCfac) output:

[1]:
   [1]:
      1
   [2]:
      10x4d2+26x3d3+47x4-117x3d-78x2d2+117x2+156xd-156
   [3]:
      x
   [4]:
      d
   [5]:
      d
[2]:
   [1]:
      1
   [2]:
      10x4d2+26x3d3+47x4-117x3d-78x2d2+117x2+156xd-156
   [3]:
      d
   [4]:
      xd-1
[3]:
   [1]:
      1
   [2]:
      xd-3
   [3]:
      10x3d2+26x2d3+47x3-117x2d-52xd2+52d
   [4]:
      xd-1
[4]:
   [1]:
      1
   [2]:
      xd-3
   [3]:
      x
   [4]:
      x
   [5]:
      x
   [6]:
      10xd3+26d4+47xd-107d2-47
[5]:
   [1]:
      1
   [2]:
      x
   [3]:
      xd-2
   [4]:
      x
   [5]:
      x
   [6]:
      10xd3+26d4+47xd-107d2-47
[6]:
   [1]:
      1
   [2]:
      x
   [3]:
      x
   [4]:
      xd-1
   [5]:
      x
   [6]:
      10xd3+26d4+47xd-107d2-47
[7]:
   [1]:
      1
   [2]:
      x
   [3]:
      x
   [4]:
      x
   [5]:
      x
   [6]:
      d
   [7]:
      10xd3+26d4+47xd-107d2-47
[8]:
   [1]:
      1
   [2]:
      x
   [3]:
      x
   [4]:
      x
   [5]:
      x
   [6]:
      10xd2+26d3+47x-97d
   [7]:
      d
   [8]:
      d

[mlee]

I've send Albert Heinle (the author of ncfactor.lib) an email to check 6358cfe22f06e105e3de6a6241edfa4cb812aa2e again as SW and master's ncfactor.lib differ by this commit.

[Albert Heinle]

Hello @all,

Martin Lee sent me the Bug description, and I could locate and fix the bug. I sent the fixed version of the code to him.

Short description of the problem: I found a way to increase the versatility of the results and added a function called "checkForHomogInhomogInterchangability". I submitted it recently, thus in 3-1-6 is everything okay still, as outlined above by motsak. In this function, there was sort of an off-by-one error at a block, that was rarely visited. Thus I did not catch it before (sorry for that)...

But now everything is okay again. Thank your for checking my library.

With best wishes,

Albert Heinle