Add malloc align test

[krtab]

Add a test for malloc well-alignedness.

The test is pretty long to run (8 sec on my machine).

Depends on #104

[krtab]

Actually, rewriting the test as below (adding assumes) bring the run time to ~3 secs.

[zapashcanon]

Thanks! Do you want to try adding a dune rule yourself so that it becomes part of dune runtest or should I do it ?

[krtab]

I'm not sure we should add it given its large run time.

[filipeom]

Thanks for the test @krtab!! There's a bug regarding the Ptr type returned by owi_alloc. I just ran this locally, and with -w 8 owi is making 8200 calls to the solver, totaling 34s in the solver. This means that the expressions in owi_assert, i.e., stuff like this:

calling func : func owi_assert
stack        : [  ]
running instr: local.get 0
stack        : [ (i32.of_bool (i32.eq (i32.rem (Ptr (i32 8390672) (i32 0)) (i32 1)) (i32 0))) ]

Is being discharged to the solver instead of being solved concretely

[krtab]

ee66c6a is actually in #127 which should be merged first

[krtab]

Performance issue fixed in https://github.com/formalsec/encoding/pull/65

[zapashcanon]

Could you run dune fmt and commit the formatted dune file ? :-)

Otherwise, should I merge this ? It's all OK for me.

[krtab]

Otherwise, should I merge this ? It's all OK for me.

Yes, the slight performance issue will be fixed once we merge hashconsing in encoding so 🟢