SF [2159499] Full bigfloat precision for Gamma after the second call



Raymond Toy wrote:
> Note that these differ quite a bit.  Just need to track down what's
> happening here....

Some additional info.  First, zot(z,fpprec) :=
bfloat(%pi*z/sin(%pi*z)); is enough to show the problem.

fpprec:64;
zot(bfloat(1/2),fpprec)$
:lisp bigfloat%pi
((BIGFLOAT SIMP 438)
 557478319480384710099680955233841104245484663623488188587759156163931561121116157458028029845525521685028647707875576434771246873456
 2)
:lisp (displa bigfloat%pi)
3.1415926535897932384626433832795028841971693993751058209749445923078164062862\
089986280348256812002485034480173261900300090541701233b0

So, for some reason, zot computed a 438-bit value of pi.  But compare
this with the "true" value of pi:

fpprec:256;  /* New value of pi computed */
3.1415926535897932384626433832795028841971693993751058209749445923078164062862\
089986280348253421170679821480865132823066470938446095505822317253594081284811\
174502841027019385211055596446229489549303819644288109756659334461284756482337\

The true value of pi (from www.joyofpi.com) is:

3.1415926535897932384626433832795028841971693993751058209749445923078164062862\
089986280348253421170679821480865132823066470938446095505822317253594081284811\
174502841027019385211055596446229489549303819644288109756659334461284756482337\

Note that 438 bit and 128 digit values of pi agree.  But the 256 bit
value is quite a bit different, in the 128 digits!

Perhaps the computation of pi is somehow broken?

Ray