[2.10] test: disable JIT in test/app/digest.test.lua

[ylobankov]

After we moved osx testing from the per-commit to nightly basis [1], we accidentally enabled running tarantool tests on macOS/M1 for 2.10. So we had some tests failed, and they were fixed by cherry-picking a few commits from the master branch [2]. Also, some tests were fixed by updating test-run to the new version [3] with needed fix [4]. So it's time to fix the last failed test (test/app/digest.test.lua) on macOS/M1 by disabling JIT in it due to the issue [5].

[1] #9571
[2] #9672
[3] #9685
[4] tarantool/test-run#422
[5] #6097

---

Tested on macOS/M1 as well:

...
Top 10 longest tests (seconds):
*  63.54 replication-luatest/quorum_orphan_test.lua
*  57.07 sql-tap/in2.test.lua:memtx
*  30.97 replication-luatest/gh_4669_applier_reconnect_test.lua
*  22.82 replication-luatest/quorum_misc_test.lua
*  22.43 sql-tap/trigger8.test.lua:memtx
*  20.72 box/gh-5422-broken_snapshot.test.lua
*  19.87 replication-luatest/gh_7318_box_wait_limbo_acked_test.lua
*  17.62 sql-tap/trigger8.test.lua:vinyl
*  15.89 engine-luatest/gh_6436_field_foreign_key_test.lua
*  15.05 vinyl/gh-4810-dump-during-index-build.test.lua
---------------------------------------------------------------------------------
Statistics:
* disabled: 73
* pass: 1583
* skip: 6

[coveralls]

coverage: 85.127% (+0.02%) from 85.109%
when pulling 34feeb3 on ylobankov/disable-digest.test.lua-test
into 71d5ff0
on release/2.10.

[Totktonada]

I've no objections. But the issue is Igor's one, it is better to ask him.

[ochaplashkin]

LGTM

[igormunkin]

@ylobankov, thanks for the patch! LGTM as trivial.