MuPAD ... About the bug-fixed release 1.2.2a


The bug-fixed version MuPAD 1.2.2a was released on 02. Feb. 1996.

MuPAD version 1.2.2 was released on 24. July 1995.


Last Date of Update: 21/12/1995

Dear MuPAD friend,

this files describes bug fixes and changes of MuPAD 1.2.2a
with respect to the first release of MuPAD 1.2.2, published
in July 1995.

It contains two parts:

I)  Bug Fixes of the MuPAD Kernel
---------------------------------

    You will find a short description about the bug and
    its status (if it is fixed or not).


II) Bug Fixes and Changes of the Library 
----------------------------------------

    You will find each function which was changed, the file
    it is contained in and its location in the library
    (relatively to the directory $MuPAD_ROOT_PATH/share/lib)
    and a short description of the changes.

    To include the fixes into your existing library you may
    either:

    (i)  replace the changed files by the new ones,

    (ii) or create your own directory (let us call it "PATCH"),
         and set your LIB_PATH to

                  LIB_PATH:= "PATCH", LIB_PATH;

         Note that the directory "PATCH" must be placed in
	 front of the original LIB_PATH, so that the changed
	 files are loaded instead of the old ones.

         The above statement could be placed either into each
	 users .mupadinit file or in the .mupadsysinit file,
	 which is located in $MuPAD_ROOT_PATH/share/lib.
 

If you have any questions or remarks, please send an e-mail to 

	MuPAD-bugs@uni-paderborn.de 

We will answer you as soon as possible.


Happy MuPAD computing!

The MuPAD Development Team.


========================================


I) Bug fixes of the MuPAD kernel
--------------------------------

+ Bug in the Pretty-Printer:
  If TEXTWIDTH was an even integer, it could happend that
  characters of long outputs were lost (or replaced by ^A 
  under Linux).

  Fixed.

+ Matrix and Table elements bigger than TEXTWIDTH are printed
  with PRETTY_PRINT = FALSE

+ In some cases missing parentheses inserted

+ To have more than one library path (given in LIB_PATH) have caused
  problems. This bug is now fixed.

=========================================


II) Bug fixes and changes of the library
----------------------------------------

DOMAINS/DOMAIN/ExprFld.mu (ExpressionField)
    bug in gcd fixed

DOMAINS/DOMAIN/Matrix.mu (Matrix)
    bug in the method gaussElim fixed

INTLIB/int.mu (int)
    bug in int(1/(x^2+k^2),x) fixed

INTLIB/solveRde.mu
    bug in int((sin(y/b)+cos(y))^2,y) fixed, bugs fixed
    (was producing Illegal argument [nthmonomial])

INTLIB/weaknorm.mu (int)
    local variables 's' and 't' were not specified

INTLIB/hermite.mu (int)
    bug fixed

INTLIB/NUMERIC/oscillate.mu was renamed to INTLIB/NUMERIC/osc.mu

LIBFILES/intlib.mu
    missing file, renamed file

LIBFILES/Network.mu 
    renamed files

LIBFILES/stdlib.mu
    renamed file

LIBFILES/module.mu
    renamed file

LINALG/concat.mu (concatMatrix)
    was extended in order to allow only one argument

LINALG/eigenvec.mu (eigenVectors)
    system of eigenvectors can now be computed only over fields.
    
LINALG/extrmat.mu (extractMatrix)
    bug fixed: dimension of submatrix (extracted by lists) was wrong.

LINALG/linsolve.mu (linearSolve)
    case linsolve(A,b,l): l must now be either a set or an array with
    *exactly* ncol(A) elements (see documentation of linalg).

LINALG/nulspace.mu (nullSpace)
    use a new backsubsitution process.

LINALG/gausjordan.mu (gaussJordan)
    if option 'All' is used, additionally a set of the characteristic
    column indices is returned (as the fourth operand of the list).

LINALG/stack.mu (stackMatrix)
    was extended in order to allow only one argument

MISC/subFrVar.mu (subsFreeVars)
    bug in func fixed 
        [ func(1/(1+t^2)^2 + x, t,x)(2,3) was producing an error ]

MODULE/expr2text was renamed to MODULE/expr2txt.mu

NETWORK/blockflow.mu was renamed to NETWORK/blckflow.mu

NETWORK/random.mu (Random)
    bug fixed

NETWORK/rdijkstra.mu was renamed to NETWORK/rdijkstr.mu

NETWORK/showgr.mu (ShowGraph)
    line width changed

SERIES/gseries.mu (gseries)
    returns an error when sum or product of series cannot be computed
    (necessary to avoid problems in series (e.g. sin(exp(-1/x)+x)))

SPECFUNC/abs.mu (abs)
    abs(x) now considers the value of sign(x)

SPECFUNC/acosh.mu (acosh)
    bug fixed (acosh was evaluated to cosh)

SPECFUNC/erf.mu (erf)
    bug fixed with undeclared local variable (could cause loading problems)

SPECFUNC/igamma.mu (igamma)
    added the "diff" attribute

SPECFUNC/sign.mu (sign)
    bug fixed

SPECFUNC/sqrt.mu (sqrt)
    bug in evaluating sqrt(a^(1/2)) fixed

STDLIB/asympt.mu (asympt)
    changed proc into purefunc to avoid name spacing problem

STDLIB/bernoulli.mu was renamed to STDLIB/bernoull.mu

STDLIB/D.mu (D)
    bug in concatenation of D fixed
    pb with infinite loop fixed

STDLIB/ReIm.mu (Re)
    Re(Re(x)) now returns Re(x)
    Bug in cases of symbolic complex numbers fixed

STDLIB/infinity.mu (infinity)
    bug in subs fixed [ subs(-infinity,infinity=infinity) ]

STDLIB/limit.mu (limit)
    limit now handles the gamma function if its
      argument do not tend to -infinity or +infinity
    bug in (stdlib::limit)::sign fixed

STDLIB/solve.mu (linsolve)
    bug fixed (for undetermined systems)

STDLIB/factor.mu (factor, Factor)
    bug fixed: factor did not (in special cases) return polynomials of the 
         same domain


Installed by: MuPAD-Webmaster
Last updated: 06. Feb. 1996