Waldek, I think in your patch, from line 49 + not(pv.v = qv.v) => + smaller?(leadingCoefficient(p), 0$R)$R => 1 + -- pv.v = qv.v
Should read + not(pv.v = qv.v) => + smaller?(leadingCoefficient(p), 0$R)$R => 1 + - 1 + -- pv.v = qv.v Or something like that. All tests pass with this modification. In CCL interpreter: (eql (the fixnum 1) (the fixnum nil)) issues a warning but returns nil whereas sbcl aborts computation. It could be that because from my point of view the triage routine can return "nothing" in this piece of code and the compiler should complain about that. >From my tests, distro.input (with )set break br) the error comes from smaller? and complains in triage(p,q) = 1 Apparently triage(p,q) returns nil sometimes and the eql_SI macro is used (src/lisp/primitives.lisp). Le mar. 18 juil. 2023 à 02:28, Grégory Vanuxem <[email protected]> a écrit : > > Le mar. 18 juil. 2023 à 00:44, Waldek Hebisch > <[email protected]> a écrit : > > > > On Mon, Jul 17, 2023 at 09:12:41PM +0200, Grégory Vanuxem wrote: > > > Hello Waldek, > > > > > > I just cloned fricas.git > > > (configure.ac : 2023-07-17 20:19:21.130766889 +0200 on my drive). > > > > > > And I guess your patch can not be used with recent Git version, am I > > > right? > > > > Thank for looking at this. I am not sure what you mean. Patch is > > with respect to recent git version, it can be applied to current > > trunk. > > Ok. > > In fact, I hesitated to ask for a simple test case where > FriCAS on sbcl with modified multpoly.spad dies sometimes with > Lisp errors. But since I had not upgraded FriCAS recently, > I decided to rebuild the latest git version and > was not expecting such a big _temporary_ regression. > > For sure, I see this patch as "for testing purposes". > > > Concerning "can be used": it can be used to reproduce the > > problem... > > I understand now :) > > > > With SBCL-2.3.6 no problem without your patch. But if I (successfully) > > > apply your patch here is the output of 'make check': > > > > > > ===================================== > > > 26 failing files: > > <snip> > > > > > > That should be reproducible I think, > > > > I do not know if all sbcl versions produce exactly the same > > set of failures. But there is a lot of failures, enough > > to reproduce the problem. I do not think that knowing > > exact set of failurs gives useful info. > > BTW, FriCAS built on Clozure CL does not exhibit this set of failures. > As ECL as you said. > > > -- > > 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 on the web visit > > https://groups.google.com/d/msgid/fricas-devel/ZLXEJd99sX2bj4zI%40fricas.math.uni.wroc.pl. -- 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 on the web visit https://groups.google.com/d/msgid/fricas-devel/CAHnU2dafJvH945bubxYcMoHtA6a95iCyoT%2BG9APAPxer7czuZA%40mail.gmail.com.
