[Octave-bug-tracker] [bug #53543] regression: new gammainc inaccurate for small real inputs

[Mike Miller]

Follow-up Comment #1, bug #53543 (project octave):

I saw a case in gammainc.m:


  elseif (strcmp (tail, "scaledlower"))
    if (abs (x) < 1/2)
      y = expm1 (x) ./ x;
    else
      y = (exp (x) - 1) ./ x;
    endif


But there is no equivalent for the case a few lines above


  if (strcmp (tail, "lower"))
    y = 1 - exp (-x);


which is the case that this bug report is about.

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?53543>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/