Re: Potential infinite loop in sb-c::array-type-upgraded-element-type with sb-kernel:union-type
Stas Boukarev <[email protected]>
| Newsgroups | gmane.lisp.steel-bank.general |
|---|---|
| Message-ID | <CAF63=10DJuSJeXBN9gW4ZYW+BXABaLP+MWC1mn_KSNLqQN+hcg@mail.gmail.com> |
Submit the original form. On Tue, Aug 6, 2024 at 10:36 PM Shubhamkar Ayare via Sbcl-help < [email protected]> wrote: > I discovered the above function calls while trying to debug a > stuck-on-compilation issue in one of my libraries. > > Thanks for pointing out that this function call is not supposed to happen! > So, the bug must lie in whatever macroexpansion and compiler macroexpansion > my library is performing. I will debug that part. > > -------- Original message -------- > From: Douglas Katzman <[email protected]> > Date: 07/08/24 00:20 (GMT+05:30) > To: Shubhamkar Ayare <[email protected]> > Cc: [email protected] > Subject: Re: [Sbcl-help] Potential infinite loop in > sb-c::array-type-upgraded-element-type with sb-kernel:union-type > > In both usages you are calling the function in ways that it should not be > called by the compiler. > > NULL hits the member-type case and the (mapcar #'ctype-of) returns the > same object, so it recurses on the same arg that it was called with; i.e. > there is no base case to the recursion. > In the second example you're just plain calling it wrong. > > What point are you trying to make by exercising an internal function > incorrectly though? > > On Tue, Aug 6, 2024 at 2:36 PM Shubhamkar Ayare via Sbcl-help < > [email protected]> wrote: > >> >> On SBCL 2.4.7 (and also 2.4.4 and 2.4.5), the following seems to result >> in an >> infinite loop. It never returns despite running for 10 seconds. >> >> (sb-c::array-type-upgraded-element-type (sb-kernel:specifier-type >> '(or array null))) >> >> By contrast, the following returns almost immediately: >> >> (sb-c::array-type-upgraded-element-type '(or array null)) >> >> Is the difference of behavior a bug? >> >> Best, >> Shubhamkar / digikar >> >> >> _______________________________________________ >> Sbcl-help mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/sbcl-help >> > _______________________________________________ > Sbcl-help mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/sbcl-help > _______________________________________________ Sbcl-help mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/sbcl-help