Convert sig_short_exp acceptance to use Bazel

[scrye]

This reduces the worst case running time of the test from
about 8 minutes to about 30 seconds.

Also:

• sig_short_exp no longer runs a control plane, instead
  using a fake sciond
• removes the old sig_short_exp acceptance test from
  the main acceptance pipeline

---

This change is 

[lukedirtwalker]

Reviewed 19 of 19 files at r1.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @scrye)

---

acceptance/sig_short_exp_time/test, line 86 at r1 (raw file):

    echo "Shutting down path A"
    docker-compose -f acceptance/sig_short_exp_time/docker-compose.yml stop patha

Is that even required?

---

acceptance/sig_short_exp_time/testdata/1-ff00_0_110/sig/fake_sciond.json, line 21 at r1 (raw file):

                    "next_hop": "242.254.100.3:50000",
                    "ia": "1-ff00:0:111",
                    "expiry_seconds": 10

Shouldn't this be 5 here? since it's already 5 seconds in?
Ah no looking at the implementation it doesn't seem like it. Ok

[scrye]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @lukedirtwalker)

---

acceptance/sig_short_exp_time/test, line 86 at r1 (raw file):

Previously, lukedirtwalker (Lukas Vogel) wrote…

Is that even required?

I'd say it is, otherwise a SIG that doesn't never checks for expiry would pass the test.

This way, the test checks that the SIG switched ahead of time to path B due to path A's imminent expiration.

---

acceptance/sig_short_exp_time/testdata/1-ff00_0_110/sig/fake_sciond.json, line 21 at r1 (raw file):

Previously, lukedirtwalker (Lukas Vogel) wrote…

Shouldn't this be 5 here? since it's already 5 seconds in?
Ah no looking at the implementation it doesn't seem like it. Ok

I was fearing the semantics would be confusing. Hopefully the new field names are more intuitive.

[lukedirtwalker]

Reviewed 8 of 8 files at r2.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @scrye)

---

acceptance/sig_short_exp_time/test, line 74 at r2 (raw file):

    # Set up forward route on network stack 1 and 2 through sig tunnel device
    # If this fails, the test is not stopped.

Maybe explain that the containers use the same network context thus the route is persisted.

---

acceptance/sig_short_exp_time/testdata/1-ff00_0_110/sig/fake_sciond.json, line 21 at r1 (raw file):

Previously, scrye (Sergiu Costea) wrote…

I was fearing the semantics would be confusing. Hopefully the new field names are more intuitive.

Seems more clear 👍

[scrye]

Reviewable status: 18 of 19 files reviewed, 1 unresolved discussion (waiting on @lukedirtwalker)

---

acceptance/sig_short_exp_time/test, line 74 at r2 (raw file):

Previously, lukedirtwalker (Lukas Vogel) wrote…

Maybe explain that the containers use the same network context thus the route is persisted.

Done.

[lukedirtwalker]

Reviewed 1 of 1 files at r3.
Reviewable status: complete! all files reviewed, all discussions resolved