On Thu, Aug 15, 2024 at 3:11 PM Eduard Zingerman <eddyz87@xxxxxxxxx> wrote:
>
> On Thu, 2024-08-15 at 15:07 -0700, Andrii Nakryiko wrote:
>
> [...]
>
> > > > Isn't that a bit counter-intuitive and potentially dangerous behavior
> > > > for checking disassembly? If my assumption is correct, maybe we should
> > > > add some sort of `__jit_x86("...")` placeholder to explicitly mark
> > > > that we allow some amount of lines to be skipped, but otherwise be
> > > > strict and require matching line-by-line?
> > >
> > > This is a valid concern.
> > > What you suggest with "..." looks good.
> >
> > I'd add just that for now. _not and _next might be useful in the
> > future, but meh.
>
> If we commit to "..." now and decide to add _not and _next in the
> future this would make __jit macro special. Which is not ideal, imo.
> (on the other hand, tests can always be adjusted).
>
It is already special with a different flavor of regex. And I assume
we won't have that many jit-testing tests, so yeah, could be adjusted,
if necessary. But just in general, while __msg() works with large
verifier logs, __jit() is much more narrow-focused, so even if it
behaves differently from __msg() I don't really see much difference.
But we also have __xlated() with similar semantics, so I'd say we
should keep __jit() and __xlated() behaving similarly.
> [...]
>
