Howdy,

So I thought I'd try kick off my aarch64 build of SBBS again (knowing you were working on a MAC version) - but I'm on a Pi.

I was super pleased to see it build and complete successfully :)

Wasnt pleased when it core dumped when I started it.

Is it too early for a Pi version?

[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/aarch64-linux-gnu/libthread_db.so.1".
Core was generated by `sbbs'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0 0x0000ffff9dc35fe0 in js_GetClassPrototype(JSContext*, JSObject*, JSProtoKey, JSObject**, js::Class*) () from /opt/sbbs/exec/libsbbs.so
[Current thread is 1 (Thread 0xffff837fe1c0 (LWP 72))]

(gdb) bt
#0 0x0000ffff9dc35fe0 in js_GetClassPrototype(JSContext*, JSObject*, JSProtoKey, JSObject**, js::Class*) () from /opt/sbbs/exec/libsbbs.so
#1 0x0000ffff9dbff710 in js_NewFunction(JSContext*, JSObject*, int (*)(JSContext*, unsigned int, js::Value*), unsigned int, unsigned int, JSObject*, JSAtom*) ()
from /opt/sbbs/exec/libsbbs.so
#2 0x0000ffff9dc02194 in js_DefineFunction(JSContext*, JSObject*, long, int (*)(JSContext*, unsigned int, js::Value*), unsigned int, unsigned int) () from /opt/sbbs/exec/libsbbs.so
#3 0x0000ffff9dbb0dc4 in JS_DefineFunctions () from /opt/sbbs/exec/libsbbs.so
#4 0x0000ffff9dc38b60 in js::DefineConstructorAndPrototype(JSContext*, JSObject*, JSProtoKey, JSAtom*, JSObject*, js::Class*, int (*)(JSContext*, unsigned int, js::Value*), unsigned int, JSPropertySpec*, JSFunctionSpec*, JSPropertySpec*, JSFunctionSpec*) () from /opt/sbbs/exec/libsbbs.so
#5 0x0000ffff9dc39790 in js_InitClass(JSContext*, JSObject*, JSObject*, js::Class*, int (*)(JSContext*, unsigned int, js::Value*), unsigned int, JSPropertySpec*, JSFunctionSpec*, JSPropertySpec*, JSFunctionSpec*) () from /opt/sbbs/exec/libsbbs.so
#6 0x0000ffff9dbffe4c in js_InitFunctionClass(JSContext*, JSObject*) () from /opt/sbbs/exec/libsbbs.so
#7 0x0000ffff9dbb0b9c in js_InitFunctionAndObjectClasses(JSContext*, JSObject*) () from /opt/sbbs/exec/libsbbs.so
#8 0x0000ffff9dbb0c38 in JS_InitStandardClasses () from /opt/sbbs/exec/libsbbs.so
#9 0x0000ffff9dac7294 in js_CreateGlobalObject () from /opt/sbbs/exec/libsbbs.so
#10 0x0000ffff9d84ae14 in js_initcx () from /opt/sbbs/exec/libservices.so
#11 0x0000ffff9d84c094 in js_static_service_thread () from /opt/sbbs/exec/libservices.so
#12 0x0000ffff9d817648 in start_thread (arg=0xffff837fdac0) at pthread_create.c:477
#13 0x0000ffff9d76dc9c in thread_start () at ../sysdeps/unix/sysv/linux/aarch64/clone.S:78


...ëîåï

---
þ Synchronet þ AnsiTEX bringing back videotex but with ANSI

Re: aarch64 build
By: deon to Digital Man on Tue Jan 21 2025 07:31 pm

> Howdy,
>
> So I thought I'd try kick off my aarch64 build of SBBS again (knowing you
> were working on a MAC version) - but I'm on a Pi.
>
> I was super pleased to see it build and complete successfully :)
>
> Wasnt pleased when it core dumped when I started it.
>
> Is it too early for a Pi version?

No, there's a lot of sysop running on rPis.

> [Thread debugging using libthread_db enabled]
> Using host libthread_db library "/lib/aarch64-linux-gnu/libthread_db.so.1".
> Core was generated by `sbbs'.
> Program terminated with signal SIGSEGV, Segmentation fault.
> #0 0x0000ffff9dc35fe0 in js_GetClassPrototype(JSContext*, JSObject*,
> JSProtoKey, JSObject**, js::Class*) () from /opt/sbbs/exec/libsbbs.so
> [Current thread is 1 (Thread 0xffff837fe1c0 (LWP 72))]

I think you just got unlucky in your timing of cloning the git repo. That bug was pretty quickly reverted today. Try again.
--
digital man (rob)

This Is Spinal Tap quote #31:
Viv Savage: Quite exciting, this computer magic!
Norco, CA WX: 51.9øF, 14.0% humidity, 1 mph NE wind, 0.01 inches rain/24hrs
---
þ Synchronet þ Vertrauen þ Home of Synchronet þ [vert/cvs/bbs].synchro.net

Re: aarch64 build
By: Digital Man to deon on Tue Jan 21 2025 12:53 am

Howdy,

> > Is it too early for a Pi version?
>
> No, there's a lot of sysop running on rPis.

On aarch64 with a 64bit build? Running 32bit SBBS on aarch64 worked for me before, but not 64 bit.

Anyway, I recompiled and confirmed I'm running master/d84a380bf, but the same segfault.

I'll recompile a DEBUG build for a proper backtrace...


...ëîåï

---
þ Synchronet þ AnsiTEX bringing back videotex but with ANSI

Re: aarch64 build
By: deon to Digital Man on Tue Jan 21 2025 10:00 pm

Howdy,

> I'll recompile a DEBUG build for a proper backtrace...

Here is a backtrace

Using host libthread_db library "/lib/aarch64-linux-gnu/libthread_db.so.1". [7/4929]
Core was generated by `sbbs'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0 0x0000ffffb57ba258 in JSObject::getClass (this=0x7fffa9e04100) at jsobj.h:427
427 jsobj.h: No such file or directory.
[Current thread is 1 (Thread 0xffffaaffd1c0 (LWP 541))]
(gdb) bt full
#0 0x0000ffffb57ba258 in JSObject::getClass (this=0x7fffa9e04100) at jsobj.h:427
No locals.
#1 0x0000ffffb57baed0 in JSObject::isFunction (this=0x7fffa9e04100) at jsfun.h:312
No locals.
#2 0x0000ffffb588c67c in js::IsFunctionObject (v=...) at jsfun.h:342
No locals.
#3 js::FindClassPrototype (cx=0xffff8c018240, scopeobj=0xffffa9e03048, protoKey=JSProto_Function, protop=0xffffaaff9c80, clasp=0xffffb5c39b28 <js_FunctionClass>) at jsobj.cpp:6168
v = {data = {asBits = 18445618172357787904, debugView = {payload47 = 140736043434240, tag = JSVAL_TAG_OBJECT}, s = {payload = {i32 = -1444921088, u32 = 2850046208,
why = 2850046208, word = 18445618172357787904}}, asDouble = -nan(0xbffffa9e04100), asPtr = 0xfffbffffa9e04100}}
#4 0x0000ffffb588c8e0 in js_GetClassPrototype (cx=0xffff8c018240, scopeobj=0xffffa9e03048, protoKey=JSProto_Function, protop=0xffffaaff9c80, clasp=0xffffb5c39b28 <js_FunctionClass>)
at jsobj.cpp:6212
No locals.
#5 0x0000ffffb5827d78 in js::FindProto (proto=0xffffaaff9c80, parent=0xffffa9e04080, clasp=0xffffb5c39b28 <js_FunctionClass>, cx=0xffff8c018240) at jsobjinlines.h:1053
protoKey = JSProto_Function
#6 js::detail::NewObject<false, true> (kind=js::gc::FINALIZE_OBJECT2, parent=0xffffa9e04080, proto=0x0, clasp=0xffffb5c39b28 <js_FunctionClass>, cx=0xffff8c018240)
at jsobjinlines.h:1070
obj = 0xffffb6a11150
#7 js::NewFunction (parent=0xffffa9e04080, cx=0xffff8c018240) at jsobjinlines.h:1114
No locals.
#8 js_NewFunction (cx=0xffff8c018240, funobj=0x0, native=0xffffb5825954 <fun_toSource(JSContext*, uintN, js::Value*)>, nargs=0, flags=0, parent=0xffffa9e04080, atom=0xffffa9e01640)
at jsfun.cpp:2729
fun = 0xffffa9e04080
#9 0x0000ffffb5828630 in js_DefineFunction (cx=0xffff8c018240, obj=0xffffa9e04080, id=281473531778624, native=0xffffb5825954 <fun_toSource(JSContext*, uintN, js::Value*)>, nargs=0,
attrs=0) at jsfun.cpp:2960
gop = 0xffffb57b05e8 <JS_PropertyStub(JSContext*, JSObject*, jsid, jsval*)>
sop = 0xffffb57b0608 <JS_StrictPropertyStub(JSContext*, JSObject*, jsid, JSBool, jsval*)>
fun = 0x8
wasDelegate = true

My build command was:
make -f install-sbbs.mk DEBUG=1 SBBSDIR=/opt/sbbs


...ëîåï

---
þ Synchronet þ AnsiTEX bringing back videotex but with ANSI

Re: aarch64 build
By: deon to Digital Man on Tue Jan 21 2025 10:00 pm

> > > Is it too early for a Pi version?
>
> > No, there's a lot of sysop running on rPis.
>
> On aarch64 with a 64bit build? Running 32bit SBBS on aarch64 worked for me
> before, but not 64 bit.

I *thought* so at least. Are you using the libmozjs from the sbbs git repo or a package for your platform/OS?

> Anyway, I recompiled and confirmed I'm running master/d84a380bf, but the
> same segfault.

Okay, I guess it's a different issue. Filing an issue at gitlab.synchro.net or discussing in #synchronet at irc.synchro.net will get a solution quicker if you're interest.
--
digital man (rob)

Sling Blade quote #5:
Karl Childers (to father): You ought not killed my little brother...
Norco, CA WX: 68.9øF, 7.0% humidity, 1 mph E wind, 0.00 inches rain/24hrs
---
þ Synchronet þ Vertrauen þ Home of Synchronet þ [vert/cvs/bbs].synchro.net

Re: aarch64 build
By: deon to Digital Man on Tue Jan 21 2025 11:23 pm

> Re: aarch64 build
> By: deon to Digital Man on Tue Jan 21 2025 10:00 pm
>
> Howdy,
>
> > I'll recompile a DEBUG build for a proper backtrace...
>
> Here is a backtrace
>
> Using host libthread_db library "/lib/aarch64-linux-gnu/libthread_db.so.1".
> [7/4929]
> Core was generated by `sbbs'.
> Program terminated with signal SIGSEGV, Segmentation fault.
> #0 0x0000ffffb57ba258 in JSObject::getClass (this=0x7fffa9e04100) at
> jsobj.h:427
> 427 jsobj.h: No such file or directory.
> [Current thread is 1 (Thread 0xffffaaffd1c0 (LWP 541))]
> (gdb) bt full
> #0 0x0000ffffb57ba258 in JSObject::getClass (this=0x7fffa9e04100) at
> jsobj.h:427
> No locals.
> #1 0x0000ffffb57baed0 in JSObject::isFunction (this=0x7fffa9e04100) at
> jsfun.h:312
> No locals.
> #2 0x0000ffffb588c67c in js::IsFunctionObject (v=...) at jsfun.h:342
> No locals.
> #3 js::FindClassPrototype (cx=0xffff8c018240, scopeobj=0xffffa9e03048,
> protoKey=JSProto_Function, protop=0xffffaaff9c80, clasp=0xffffb5c39b28
> <js_FunctionClass>) at jsobj.cpp:6168
> v = {data = {asBits = 18445618172357787904, debugView = {payload47 =
> 140736043434240, tag = JSVAL_TAG_OBJECT}, s = {payload = {i32 = -1444921088,
> u32 = 2850046208,
> why = 2850046208, word = 18445618172357787904}}, asDouble =
> -nan(0xbffffa9e04100), asPtr = 0xfffbffffa9e04100}}
> #4 0x0000ffffb588c8e0 in js_GetClassPrototype (cx=0xffff8c018240,
> scopeobj=0xffffa9e03048, protoKey=JSProto_Function, protop=0xffffaaff9c80,
> clasp=0xffffb5c39b28 <js_FunctionClass>)
> at jsobj.cpp:6212
> No locals.
> #5 0x0000ffffb5827d78 in js::FindProto (proto=0xffffaaff9c80,
> parent=0xffffa9e04080, clasp=0xffffb5c39b28 <js_FunctionClass>,
> cx=0xffff8c018240) at jsobjinlines.h:1053
> protoKey = JSProto_Function
> #6 js::detail::NewObject<false, true> (kind=js::gc::FINALIZE_OBJECT2,
> parent=0xffffa9e04080, proto=0x0, clasp=0xffffb5c39b28 <js_FunctionClass>,
> cx=0xffff8c018240)
> at jsobjinlines.h:1070
> obj = 0xffffb6a11150
> #7 js::NewFunction (parent=0xffffa9e04080, cx=0xffff8c018240) at
> jsobjinlines.h:1114
> No locals.
> #8 js_NewFunction (cx=0xffff8c018240, funobj=0x0, native=0xffffb5825954
> <fun_toSource(JSContext*, uintN, js::Value*)>, nargs=0, flags=0,
> parent=0xffffa9e04080, atom=0xffffa9e01640)
> at jsfun.cpp:2729
> fun = 0xffffa9e04080
> #9 0x0000ffffb5828630 in js_DefineFunction (cx=0xffff8c018240,
> obj=0xffffa9e04080, id=281473531778624, native=0xffffb5825954
> <fun_toSource(JSContext*, uintN, js::Value*)>, nargs=0,
> attrs=0) at jsfun.cpp:2960
> gop = 0xffffb57b05e8 <JS_PropertyStub(JSContext*, JSObject*, jsid,
> jsval*)>
> sop = 0xffffb57b0608 <JS_StrictPropertyStub(JSContext*, JSObject*,
> jsid, JSBool, jsval*)>
> fun = 0x8
> wasDelegate = true
>
> My build command was:
> make -f install-sbbs.mk DEBUG=1 SBBSDIR=/opt/sbbs

So that would build/use the libmozjs from the sbbs repo. One thing to try would be to install/use the libmozjs package for your OS/platform (or similar), instead:
https://wiki.synchro.net/howto:raspbian_install


Those instructions previously included teh JSINCLUDE and JSLIB options to point to the installed/system libs and header files for libmozjs.
--
digital man (rob)

Sling Blade quote #25:
Karl: they seen fit to put me in here and here I've been a great long while.
Norco, CA WX: 68.9øF, 7.0% humidity, 1 mph E wind, 0.00 inches rain/24hrs
---
þ Synchronet þ Vertrauen þ Home of Synchronet þ [vert/cvs/bbs].synchro.net

Re: aarch64 build
By: Digital Man to deon on Tue Jan 21 2025 12:02 pm

Howdy,

> So that would build/use the libmozjs from the sbbs repo. One thing to try
> would be to install/use the libmozjs package for your OS/platform (or
> similar), instead:
> https://wiki.synchro.net/howto:raspbian_install

>
> Those instructions previously included teh JSINCLUDE and JSLIB options to
> point to the installed/system libs and header files for libmozjs.

Been here (many) times before.

That just ultimately fails (libffi if I recall) not knowing what to do on aarch64...


...ëîåï

---
þ Synchronet þ AnsiTEX bringing back videotex but with ANSI