On Thu, Jun 11, 2026 at 07:16:46AM +0800, Qian Yun wrote:
> On 6/11/26 5:27 AM, Waldek Hebisch wrote:
> > On Wed, Jun 10, 2026 at 07:08:53PM +0800, Qian Yun wrote:
> >> I have verified from the generated lsp file that these
> >> 4 functions are not called. They are overwritten by
> >> functions defined later in the source file.
> >
> > Right, those 4 versions should not be used. But is seems that
> > there are other troubles related to those functions, like:
> >
> > uP := UPXS(EXPR(INT), z, 0)
> > acsc(1/sin(z::uP))
> >
> > >> Error detected within library code:
> > acsc of function with singularity
> >
> > I am not sure if you want to look at this deeper. If not just commit
> > this fix.
> >
> Is it related that at z=0, the function value need to be computed
> via limit?
There is loose connection. But main thing is that various special
definition of power series operations try to compute result when
it makes sense, even if naive definition fails. We have
'csc(y) = 1/sin(y)' so to compute 'acsc(z)' we essentially should
do 'asin(1/z)' which leads to quite different conditions than
current code.
I actually started looking at this because normally we prefer
general definition othere special cases. So why we need special
'acoshIfCan'? We need it because we need special version
of 'acosh' to correctly handle things like 'acosh(1 + z)' (which
needs square root). Such special handling is needed for most
of inverse trigonometric functions and most of inverse hyperbolic
functions. And then I noticed that condition in 'acsc' and
few others looks fishy.
--
Waldek Hebisch
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion visit
https://groups.google.com/d/msgid/fricas-devel/ain4KCR_lz9WVt02%40fricas.org.
