Post new topic Reply to topic  [ 3 posts ] 
Author Message
 Post subject: ksReducePoly and loop construct [SOLVED]
PostPosted: Thu Jul 07, 2011 4:45 pm 

Joined: Mon Jun 27, 2011 5:01 pm
Posts: 4
Location: Aachen
Hello,
i have two technical questions concerning some code from the singular kernel.

1.)
In kInline.cc kspoly.cc a function named ksOldSpolyRed invokes the function ksReducePoly with 3 Arguments:
Code:
//extracted from kInline.cc
KINLINE poly ksOldSpolyRed(poly p1, poly p2, poly spNoether)
{
  LObject L(p2);
  TObject T(p1);

  ksReducePoly(&L, &T, spNoether);

  return L.GetLmCurrRing();
}
,
but i can find no definition of ksReducePoly, which need less than 5 Arguments.
The only definition i can find (in kspoly.cc) needs 5 Argument (and none of it is optional).

Where can i find the appropriate definition needed for ksOldSpolyRed?

(I am sorry for spamming with code here, i don't really believe it will be necessary for you to review it, but i give it for the sake of completeness.)
Code:
//extracted from kspoly.cc
/***************************************************************
*
* Reduces PR with PW
* Assumes PR != NULL, PW != NULL, Lm(PW) divides Lm(PR)
*
***************************************************************/
int ksReducePoly(LObject* PR,
                 TObject* PW,
                 poly spNoether,
                 number *coef,
                 kStrategy strat)
{
#ifdef KDEBUG
  red_count++;
#ifdef TEST_OPT_DEBUG_RED
  if (TEST_OPT_DEBUG)
  {
    Print("Red %d:", red_count); PR->wrp(); Print(" with:");
    PW->wrp();
  }
#endif
#endif
  int ret = 0;
  ring tailRing = PR->tailRing;
  kTest_L(PR);
  kTest_T(PW);

  poly p1 = PR->GetLmTailRing();   // p2 | p1
  poly p2 = PW->GetLmTailRing();   // i.e. will reduce p1 with p2; lm = LT(p1) / LM(p2)
  poly t2 = pNext(p2), lm = p1;    // t2 = p2 - LT(p2); really compute P = LC(p2)*p1 - LT(p1)/LM(p2)*p2
  assume(p1 != NULL && p2 != NULL);// Attention, we have rings and there LC(p2) and LC(p1) are special
  p_CheckPolyRing(p1, tailRing);
  p_CheckPolyRing(p2, tailRing);

  pAssume1(p2 != NULL && p1 != NULL &&
           p_DivisibleBy(p2,  p1, tailRing));

  pAssume1(p_GetComp(p1, tailRing) == p_GetComp(p2, tailRing) ||
           (p_GetComp(p2, tailRing) == 0 &&
            p_MaxComp(pNext(p2),tailRing) == 0));

#ifdef HAVE_PLURAL
  if (rIsPluralRing(currRing))
  {
    // for the time being: we know currRing==strat->tailRing
    // no exp-bound checking needed
    // (only needed if exp-bound(tailring)<exp-b(currRing))
    if (PR->bucket!=NULL)  nc_kBucketPolyRed(PR->bucket, p2,coef);
    else
    {
      poly _p = (PR->t_p != NULL ? PR->t_p : PR->p);
      assume(_p != NULL);
      nc_PolyPolyRed(_p, p2,coef);
      if (PR->t_p!=NULL) PR->t_p=_p; else PR->p=_p;
      PR->pLength=0; // usaully not used, GetpLength re-comoutes it if needed
    }
    return 0;
  }
#endif

  if (t2==NULL)           // Divisor is just one term, therefore it will
  {                       // just cancel the leading term
    PR->LmDeleteAndIter();
    if (coef != NULL) *coef = n_Init(1, tailRing);
    return 0;
  }

  p_ExpVectorSub(lm, p2, tailRing); // Calculate the Monomial we must multiply to p2

  if (tailRing != currRing)
  {
    // check that reduction does not violate exp bound
    while (PW->max != NULL && !p_LmExpVectorAddIsOk(lm, PW->max, tailRing))
    {
      // undo changes of lm
      p_ExpVectorAdd(lm, p2, tailRing);
      if (strat == NULL) return 2;
      if (! kStratChangeTailRing(strat, PR, PW)) return -1;
      tailRing = strat->tailRing;
      p1 = PR->GetLmTailRing();
      p2 = PW->GetLmTailRing();
      t2 = pNext(p2);
      lm = p1;
      p_ExpVectorSub(lm, p2, tailRing);
      ret = 1;
    }
  }

  // take care of coef buisness
  if (! n_IsOne(pGetCoeff(p2), tailRing))
  {
    number bn = pGetCoeff(lm);
    number an = pGetCoeff(p2);
    int ct = ksCheckCoeff(&an, &bn);    // Calculate special LC
    p_SetCoeff(lm, bn, tailRing);
    if ((ct == 0) || (ct == 2))
      PR->Tail_Mult_nn(an);
    if (coef != NULL) *coef = an;
    else n_Delete(&an, tailRing);
  }
  else
  {
    if (coef != NULL) *coef = n_Init(1, tailRing);
  }


  // and finally,
  PR->Tail_Minus_mm_Mult_qq(lm, t2, PW->GetpLength() - 1, spNoether);
  assume(PW->GetpLength() == pLength(PW->p != NULL ? PW->p : PW->t_p));
  PR->LmDeleteAndIter();

  // the following is commented out: shrinking
#ifdef HAVE_SHIFTBBA_NONEXISTENT
  if ( (currRing->isLPring) && (!strat->homog) )
  {
    // assume? h->p in currRing
    PR->GetP();
    poly qq = p_Shrink(PR->p, currRing->isLPring, currRing);
    PR->Clear(); // does the right things
    PR->p = qq;
    PR->t_p = NULL;
    PR->SetShortExpVector();
  }
#endif
 
#if defined(KDEBUG) && defined(TEST_OPT_DEBUG_RED)
  if (TEST_OPT_DEBUG)
  {
    Print(" to: "); PR->wrp(); Print("\n");
  }
#endif
  return ret;
}


My second question is less important, i just ask out of curiosity.
2.)
In various function in the kernel of Singular i found a strange loop construct, which i suspect not to be in the language specification of C. Is it a macro? Where is it defined? Here is an example extracted from the bba function in kstd2.cc:
Code:
loop
{
  if (j>=k) break;
  clearS(strat->S[j],strat->sevS[j],&k,&j,strat);
  j++;
}


edit:
I forgot to say, that i use Singular Version 3-1-3.

Background:
I am currently adapting code from the singular kernel for a new letterplace approach using Distance Vectors as shift invariant representations for monomials.
I am a Hiwi of Viktor Levandovskyy Lehrstuhl D fuer Mathematik, Rwth Aachen and working together with Grischa Studzinski.

Thanks in advance for any answers, especially with respect to my first question.
Benjamin Schnitzler


Last edited by bschnitz on Mon Jul 11, 2011 10:56 am, edited 1 time in total.

Report this post
Top
 Profile  
Reply with quote  
 Post subject: Re: ksReducePoly and loop construct
PostPosted: Fri Jul 08, 2011 2:57 pm 

Joined: Wed May 25, 2005 4:16 pm
Posts: 275
1.) there are 2 functions ksReducePoly (botyh declared in kernel/kutil.h)
one with 5 argument (the last 3 are optional), one with
4 argument (the last is optional)
Remember, that optional argument can only be given at exact one place, which is usulaly the declaration, not the implementation

2.) loop is indeed a construct which is missing in C,
but quite useful as Modula2 shows.
We emulated it by:
Quote:
#define loop for(;;)

in structs.h


Report this post
Top
 Profile  
Reply with quote  
 Post subject: Re: ksReducePoly and loop construct [SOLVED]
PostPosted: Mon Jul 11, 2011 10:56 am 

Joined: Mon Jun 27, 2011 5:01 pm
Posts: 4
Location: Aachen
Hi,
thank you for your help. For some unknown reason, i can't find the declarations ksReducePoly through my by ctags generated tag file (:ts ksReducePoly in vim only showed me the definition in kspoly.cc), but now you gave me the header file, i could find it!
Ben


Report this post
Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 3 posts ] 

You can post new topics in this forum
You can reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

It is currently Fri May 13, 2022 11:02 am
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group