Re: Build failure on FreeBSD/PowerPC 32-bit

Looking back to the last version I could build (1.4.7), I had to increase
"rehash-size" from 1.5 to 2.0 for a successful build.

However, the code has changed.  Is it possible to change "rehash-size"?  If
so, where do I make the change?

On Sat, Jan 11, 2025 at 1:46 PM Curtis Hamilton <[email protected]> wrote:

> Not sure where to go from here!  This is probably a dead end, but I'm
> going to attempt a build on LE.
>
> On Sat, Jan 11, 2025 at 12:24 PM Stas Boukarev <[email protected]> wrote:
>
>> SB-IMPL::EXPECTED-N-EMPTY-BUCKETS calls EXP, that's probably where it
>> picks up a NAN or something.
>>
>> On Sat, Jan 11, 2025 at 8:22 PM Stas Boukarev <[email protected]> wrote:
>> >
>> > Which also probably means it's actually a NAN or an infinity. Is the
>> > FPU alright? Is it affected by the different ABI? Are there any
>> > trigonometric functions involved?
>> >
>> > On Sat, Jan 11, 2025 at 8:20 PM Stas Boukarev <[email protected]>
>> wrote:
>> > >
>> > > That seems like the error is in ERROR itself. I don't think that
>> > > matters, as really there should be no single-float to bignum
>> > > conversions.
>> > >
>> > > On Sat, Jan 11, 2025 at 8:16 PM Curtis Hamilton <[email protected]>
>> wrote:
>> > > >
>> > > > Per your suggestion, here's the results:
>> > > >
>> > > > internal error too early in init, can't recover
>> > > >
>> > > > Internal error #10 "Object is of the wrong type." at 0x5077f70c
>> > > >     SC: 0, Offset: 30  0x0000004e: other immediate 1, unbound-marker
>> > > >     SC: 3, Offset: 9 $1=  0x50776c73: list pointer
>> > > > Welcome to LDB, a low-level debugger for the Lisp runtime
>> environment.
>> > > > ldb> $1=  0x50776c73: list pointer
>> > > > $2=    car: 0x5062d78f: OR
>> > > > $3=    cdr: 0x50776c6b: ($4=STRING $5=FUNCTION $6=SYMBOL
>> $7=CONDITION ...)
>> > > > ldb> $5=  0x5062a28f: other pointer
>> > > >    header: 0x00000026: symbol
>> > > > ldb> $4=  0x5063042f: other pointer
>> > > >    header: 0x00000026: symbol
>> > > > $8=    fun: 0x517fcd3d: #<ptr to simplefun>
>> > > > ldb> $6=  0x50630c2f: other pointer
>> > > >    header: 0x00000026: symbol
>> > > > ldb> Really quit? [y]       550.96 real         0.05 user
>>  0.11 sys
>> > > >
>> > > > On Sat, Jan 11, 2025 at 9:04 AM Stas Boukarev <[email protected]>
>> wrote:
>> > > >>
>> > > >> Protip, instead of "p pointer", do "p $1" (or whichever number
>> after $ it is).
>> > > >>
>> > > >> On Sat, Jan 11, 2025 at 5:02 PM Douglas Katzman via Sbcl-help
>> > > >> <[email protected]> wrote:
>> > > >> >
>> > > >> >
>> > > >> >
>> > > >> > On Sat, Jan 11, 2025 at 8:25 AM Curtis Hamilton <
>> [email protected]> wrote:
>> > > >> >>
>> > > >> >> Could the issue be related to the ABI?  FreeBSD uses ABI 2 on
>> PowerPC both big and little endian, which is the same ABI Linux uses for
>> little endian.  According to “make-config.sh” little endian is never 32-bit.
>> > > >> >>
>> > > >> > Yes, I think this would be a problem for calling from Lisp into
>> C. However, the failure you're getting implies that at least _some_  calls
>> into C are working.  The fact that it is able to print the list of files
>> means it successfully invoked write() in libc.
>> > > >> >
>> > > >> > When you get this error:
>> > > >> >    SC: 0, Offset: 30  0x0000004e: other immediate 1,
>> unbound-marker
>> > > >> >     SC: 3, Offset: 9 $1=  0x50775c6b: list pointer
>> > > >> > Can you try 'p 0x50775c6b' (substituting in the actual value of
>> that pointer if it is not this) to get a little more information? I'm
>> slightly confused why UNARY-TRUNCATE-SINGLE-FLOAT-TO-BIGNUM is trying to
>> signal a type error about an object the name of whose type is a list.
>> Also, if I purposely call unary-truncate-single-float-to-bignum with
>> unbound-marker, then I get a fairly hard crash (which is reasonable since
>> it is system-internal), and not a type error.
>> > > >> >
>> > > >> > _______________________________________________
>> > > >> > 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