[tor-bugs] #29280 [Core Tor/Tor]: Use Chutney in Tor's CI

[Tor Bug Tracker & Wiki]

#29280: Use Chutney in Tor's CI
-------------------------------------------------+-------------------------
 Reporter:  cohosh                               |          Owner:  (none)
     Type:  task                                 |         Status:
                                                 |  needs_revision
 Priority:  High                                 |      Milestone:  Tor:
                                                 |  0.4.2.x-final
Component:  Core Tor/Tor                         |        Version:
 Severity:  Normal                               |     Resolution:
 Keywords:  CI, PTs, 029-backport,               |  Actual Points:  1
  034-backport, 035-backport, 040-backport,      |
  network-team-roadmap-2019-Q1Q2, reviewer-was-  |
  teor-20190422, tor-ci, 041-deferred-20190530   |
Parent ID:  #29267                               |         Points:  1
 Reviewer:                                       |        Sponsor:
                                                 |  Sponsor19
-------------------------------------------------+-------------------------

Comment (by teor):

 Replying to [comment:25 nickm]:
 > Summarizing for my benefit, here are the changes I need to make:
 >   * Make a new test-network-ci target that behaves like test-network-
 all, but only runs the selected networks for CI (bridges-min, hs-v2-min,
 and hs-v3-min where supported).  This should happen after #30459 is
 merged, so that we can have chutney handle the skip logic.

 Let's just use test-network-all?
 All of the networks in test-network-all work fine in chutney's CI.

 Let's add a Linux job, and a macOS job (for IPv6).
 Once we're sure chutney is stable in Tor's CI, and we know how much time
 it takes, we can think about adding it to the end of one of the other
 jobs.
 (Or we can remove or combine some of the existing jobs.)

 >   * Open a ticket to disable "make check" for stem.  Do this after stem
 is allow_failure, to avoid conflects.
 >   * Re-do the merges to 0.3.5 and forward; drop plans to support 0.3.4.
 Make sure that I get the use of SKIP_MAKE_TEST right.
 >
 > I think that I am currently blocked on the allow_failure change in stem
 and the merge of #30459.

 I think they've both been merged, do you need stem allow_failures
 backported to 0.3.5?

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/29280#comment:26>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online