tlm@pur-phy (Timothy Lee Meisenheimer) (04/02/88)
In article <26370@cca.CCA.COM> jack@CCA.CCA.COM.UUCP (Jack Orenstein) writes: >In article <8803292144.AA02827@MATH.Tau.Ac.IL> <finkel%TAURUS.BITNET@CUNYVM.CUNY.EDU> writes: >I've run into another problem with the compiler trying to do arithmetic, >in Manx 3.6a. Any float expression that yields a constant, e.g. (float) >1 / (float) 2, causes the compiler to abort with the message "No ieee >library, Abort by user". However, something like one = 1; two = 2; >one_half = (float) one / (float) two; works since (I assume) >compile-time arithmetic is avoided. > >Jack Orenstein I have seen the same thing happen. It seems that if the ieee libraries are purged and the compiler needs it, it doesn't know how to suck them back in! When ever I see the above problem I run a program that needs them (they get sucked in) and then the compile works. Rebooting will take care of it too. It seems less stable in this rergards than 3.4 :-( On a side note: I was bemoaning the fact that I thought I had a lemon 2000 a few weeks ago where my floppies were always taking hits when about half full (especially on downloads) and with someone's wise suggestion I started checking to see if it was a delay(0) bug or something of that ilk. Well, I found it! I at first blamed Matts shell, or popcli etc. but nnnooooooo, it was CBM's clock program!!!! It this very old news??? Or does no one else like to see an analog clock? :-) I have stopped running this program and have not had one problem after about 5MBytes of transfers. Enlightenment comes slowly to a few of us. tlm@newton.physics.purdue.edu