Avoid crashing when we ask for running bridges, but UseBridges is 0

Instead, log a bug, and return the obvious answer: "there are no running bridges".

changed milestone to %Tor: 0.3.0.x-final

added 030-backport 031-backport actualpoints::0.2 component::core tor/tor dos-resistance milestone::Tor: 0.3.0.x-final owner::teor parent::24392 points::0.1 priority::medium resolution::fixed severity::normal status::closed type::defect version::tor 0.3.0.1-alpha labels

Please see my branch bug23524.

This bug can't be triggered in existing code, so there's no need to backport it unless we think we'll backport extra calls to this function.

Trac:
Keywords: N/A deleted, dos-resistance added
Status: new to needs_review

I added an additional commit to make sure the bug warning isn't triggered by the code I added in #23347 (moved).

Nice; thank you!

Merging now.

Trac:
Resolution: N/A to fixed
Status: needs_review to closed

This was a bug on #23347 (moved), but now we're replacing uses of any_bridge_descriptors_known() iwith num_bridges_usable() > 0, we need to make this change there instead.

Trac:
Status: closed to reopened
Parent: N/A to #24367 (moved)
Resolution: fixed to N/A
Keywords: 030-backport, 031-backport deleted, N/A added

This change is in my bug24367_032 branch.

Trac:
Status: reopened to needs_review

Creating review-group-26.

Trac:
Keywords: N/A deleted, review-group-26 added

setting owner

Trac:
Owner: N/A to teor
Status: needs_review to assigned

Trac:
Status: assigned to needs_review

Teor, did you mean "bug24367"? I don't see "bug24367_032".

Oops, I might have forgotten to push it. They are the same.

Now I've pushed bug24367_032, the relevant commit for this ticket is d7833c9d27.

Yes, this seems okay; merging to 0.3.2. Should part of this be a backport candidate? If not, please close.

Trac:
Status: needs_review to merge_ready
Milestone: Tor: 0.3.2.x-final to Tor: 0.3.1.x-final

I spoke too soon -- the unit tests break when I merge this branch:

  FAIL src/test/test_entrynodes.c:618: assert(gs_br OP_EQ NULL): 0x5585741f34b0 vs (nil)entrynodes/parse_from_state_full: 
  [parse_from_state_full FAILED]
entrynodes/get_guard_selection_by_name: 
  FAIL src/test/test_entrynodes.c:802: assert(gs3 == get_guard_selection_info())
  [get_guard_selection_by_name FAILED]

  FAIL src/test/test_entrynodes.c:2258: assert(r OP_EQ 0): -1 vs 0entrynodes/drop_guards: 
  [drop_guards FAILED]
entrynodes/outdated_dirserver_exclusion: 
  FAIL src/test/test_entrynodes.c:2777: expected log to contain "No primary or confirmed guards available."  Captured logs:
    1. info: "Not setting md restriction: only 0 usable guards.\n"
    2. info: "Trying to sample a reachable guard: We know of 0 in the USABLE_FILTERED set.\n"
    3. info: "  (That isn\'t enough. Trying to expand the sample.)\n"
    4. info: "Expanding the sample guard set. We have 0 guards in the sample, and 0 eligible guards to extend it with.\n"
    5. info: "Not expanding the guard sample any further; just ran out of eligible guards\n"
    6. info: "  (After filters [b], we have 0 guards to consider.)\n"
    7. info: "Trying to sample a reachable guard: We know of 0 in the USABLE_FILTERED set.\n"
    8. info: "  (That isn\'t enough. Trying to expand the sample.)\n"
    9. info: "Expanding the sample guard set. We have 0 guards in the sample, and 0 eligible guards to extend it with.\n"
   10. info: "Not expanding the guard sample any further; just ran out of eligible guards\n"
   11. info: "  (After filters [7], we have 0 guards to consider.)\n"
   12. info: "Absolutely no sampled guards were available. Marking all guards for retry and starting from top again.\n"
   13. info: "No router found for microdescriptor fetch; falling back to dirserver list.\n"
   14. notice: "While fetching directory info, no running dirservers known. Will try again later. (purpose 19)\n"

  [outdated_dirserver_exclusion FAILED]
4/1031 TESTS FAILED. (31 skipped)

Trac:
Milestone: Tor: 0.3.1.x-final to Tor: 0.3.2.x-final
Status: merge_ready to needs_revision

Yes, #24392 (moved) is assigned to you (nickm) to fix the unit tests, because I don't understand the guard code well enough. I split it off from #24367 (moved) because that ticket deals with a lot of different issues.

Trac:
Parent: #24367 (moved) to #24392 (moved)

Okay, I found the problem: the entrynode tests were depending on some global-state that some of the earlier tests were munging. I haven't tracked down what global state exactly, though -- only enough to fix this.

See my branch bug24367_032, based on yours.

Trac:
Status: needs_revision to merge_ready

Ok, I think we can do the reviews of this in #24367 (moved).

On this commit, I'm mainly waiting to know whether Roger thinks that the merge should be blocked because of #24486 (moved)

Trac:
Keywords: N/A deleted, review-group-27 added

This was fixed in #24392 (moved) in master and 0.3.2. We might also need to backport this to 0.3.0 and 0.3.1.

Trac:
Milestone: Tor: 0.3.2.x-final to Tor: 0.3.0.x-final
Status: merge_ready to needs_revision
Keywords: review-group-26, review-group-27 deleted, 030-backport, 031-backport added

I backported this in bug24367_030. I am happy for us to handle this in #24392 (moved).

Trac:
Status: needs_revision to closed
Actualpoints: 0.1 to 0.2
Resolution: N/A to fixed

closed

changed time estimate to 48m

added 1h 36m of time spent

mentioned in issue #24392 (moved)

mentioned in issue #25493 (moved)

moved to tpo/core/tor#23524 (closed)

mentioned in issue tpo/core/tor#25493 (closed)