[9front] 5e

[9front] 5e

[Stuart Morrow]

Everyone has to have thought this before, and just not asked. (But
it's probably been asked on IRC in the last ~ten years.)

Why is it 8i, vi, ..., 5e?

Re: [9front] 5e

[Jacob Moody]

On 5/19/24 15:07, Stuart Morrow wrote:
> Everyone has to have thought this before, and just not asked. (But
> it's probably been asked on IRC in the last ~ten years.)
> 
> Why is it 8i, vi, ..., 5e?

They are different designs really.
The ?i programs are basically a db(1) that also
emulates the system itself. They share a lot of
code designs with each other.

5e is more of a "userspace emulator" and it
was written from the ground up by aiju. It
has more in common with the gba emulator than
it does any of the ?i programs. 5e also emulates
a lot more, ie it's possible to run rio in 5e
but ?i doesn't even emulate fork.

To me the name change represented a break in the
code design and interface. I would suggest any
future additions of emulators like these also
should adopt the ?e notation and share more
with the 5e code then the ?i code.


Thanks,
moody

Re: [9front] 5e

[Stuart Morrow]

Ah. I understand this distinction. ?i is the type of emulation used
for like, shellcode analysis by security people.

Also, if I had read the man pages first I wouldn't have had to ask.
One thing I've just learned is there is actually such a program as 5i.
The existence of both for the same cputype instantly suggests two
different uses, instead of the newest interpreter just having the
wrong name.

One last random observation:

>     BUGS
>          The code generated by the compilers is well supported, but
>          some unusual instructions are unimplemented.  Some Plan 9

Qemu-386 actually can't run 9vx. It creates a window and then dies at
the point where vx32 starts doing stuff.