Hi,
Try
ldd ./bin/squeak
which will show you the libraries that the squeak binary must resolve before it even begins to start, ie, before main() is called.
In my case on x86-64
ldd squeak
linux-vdso.so.1 (0x00007ffddf20a000)
libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007fc93fdb1000)
libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007fc93fc62000)
libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007fc93fc5c000)
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007fc93fa6a000)
/lib64/ld-linux-x86-64.so.2 (0x00007fc940049000)
and on Arm32 PI
ldd local/squeak/squeak
linux-vdso.so.1 (0xbefce000)
/usr/lib/arm-linux-gnueabihf/libarmmem-${PLATFORM}.so => /usr/lib/arm-linux-gnueabihf/libarmmem-v7l.so (0xb6f12000)
libuuid.so.1 => /lib/arm-linux-gnueabihf/libuuid.so.1 (0xb6ee4000)
libutil.so.1 => /lib/arm-linux-gnueabihf/libutil.so.1 (0xb6ed1000)
libpthread.so.0 => /lib/arm-linux-gnueabihf/libpthread.so.0 (0xb6ea7000)
libm.so.6 => /lib/arm-linux-gnueabihf/libm.so.6 (0xb6e25000)
libdl.so.2 => /lib/arm-linux-gnueabihf/libdl.so.2 (0xb6e12000)
libc.so.6 => /lib/arm-linux-gnueabihf/libc.so.6 (0xb6cc4000)
/lib/ld-linux-armhf.so.3 (0xb6f27000)
And on Arm64
ldd local/squeak/squeak
linux-vdso.so.1 (0x0000007f86108000)
libpthread.so.0 => /lib/aarch64-linux-gnu/libpthread.so.0 (0x0000007f85e4b000)
libm.so.6 => /lib/aarch64-linux-gnu/libm.so.6 (0x0000007f85d92000)
librt.so.1 => /lib/aarch64-linux-gnu/librt.so.1 (0x0000007f85d7b000)
libdl.so.2 => /lib/aarch64-linux-gnu/libdl.so.2 (0x0000007f85d66000)
libc.so.6 => /lib/aarch64-linux-gnu/libc.so.6 (0x0000007f85c0d000)
/lib/ld-linux-aarch64.so.1 (0x0000007f860dc000)
These libraries have to be there to get started.
To minimize the length of this email I'll send a second one about debugging this.
cheers
bruce
On 2021-09-28T05:39:24.000+02:00, tim Rowledge <tim@rowledge.org> wrote:
On 2021-09-27, at 7:42 PM, David T. Lewis <lewis@mail.msen.com> wrote:
For the Unix VM, Ian designed a system of loadable VM modules. These
are analogous to the VM plugins (FilePlugin et al) except that they
are loaded directly by the VM executable itself before the image is
loaded. These modules are used to provide support for subsystems
such as sound and display.
Exactly - which was one of the reasons I got puzzled by the apparent insistence on loading libasound...
Try `./bin/squeak -vm-sound-null -h` instead, see below for explanation
of why this might work.
It's making no difference at all; I've even tried ./bin/squeak -vm-sound-oss -help to see if it makes any difference. It gets a bit weirder though. One the 'old' machine where things work decently
`./bin/squeak -vm-sound-oss -help
Usage: ./bin/squeak [<option>...] [<imageName> [<argument>...]]
./bin/squeak [<option>...] -- [<argument>...]
options begin with single -, but -- prefix is silently accepted
vm-sound-NAS tryLoading /srv/squeakbuildfactory/bin/vm-sound-NAS.so: dlopen: libaudio.so.2: cannot open shared object file: No such file or directory
ALSA <option>s:` etc etc
On the 'new ' set up
` ./bin/squeak -vm-sound-oss -help
./bin/squeak: error while loading shared libraries: libasound.so.2: cannot open shared object file: No such file or directory'
ie as before.
It's as if sometihng in whatever config/make environment created this particular VM had some clause explicitly linking to libasound. But even that doesn't provide any reason why one side complains about libasound and then runs anyway, whilst the other complains and dies!
I guess I'll try a new VM etc tomorrow, see if that makes any difference. Whodathunk simply moving a file could cause so much bother?
tim
--
tim Rowledge; tim@rowledge.org; www.rowledge.org/tim
Strange OpCodes: FLR: Flash Lights Randomly