Same thing with this patch: https://chromium-review.googlesource.com/c/chromium/src/+/761557/3

Ouch, sorry for this. Looking.

Hm, in latest attempt, things appear to be working:

2017-11-09 15:10:07.044 UTC-7
[pid:40740 tid:140553849653056 infra_internal.services.cq.verification.try_job_utils:603] GerritPendingCommit(761420, 4) manually opts out from equivalent builders

2017-11-09 15:10:07.045 UTC-7
[pid:40740 tid:140553849653056 infra_internal.services.cq.verification.try_job:155] for GerritPendingCommit(761420, 4) computed Requirement(custom=False, no_try=False, no_presubmit=False, 25 builds {master.tryserver.chromium.android: [android_arm64_dbg_recipe, android_clang_dbg_recipe, android_compile_dbg, android_cronet, android_n5x_swarming_rel, cast_shell_android, linux_android_rel_ng], master.tryserver.chromium.chromiumos: [chromeos-amd64-generic-rel], master.tryserver.chromium.linux: [cast_shell_linux, chromeos_daisy_chromium_compile_only_ng, chromium_presubmit, fuchsia_x64, linux_chromium_asan_rel_ng, linux_chromium_chromeos_rel_ng, linux_chromium_compile_dbg_ng, linux_chromium_headless_rel, linux_chromium_rel_ng, linux_chromium_tsan_rel_ng], master.tryserver.chromium.mac: [ios-simulator, mac_chromium_compile_dbg_ng, mac_chromium_rel_ng], master.tryserver.chromium.win: [win10_chromium_x64_rel_ng, win7_chromium_rel_ng, win_chromium_compile_dbg_ng, win_clang]})

ah, i see now. The issue is that first description was like this:

"""
BUG=none
TEST=none
TBR=jam@chromium.org
No-Equivalent-Builders: true

Change-Id: I0f59ddc4ff0efca4fd6dd17ff7503522cb9b1be6
"""

This won't work, silently. 

(sent too early)
because 'No-Equivalent-Builders: true' must be in last paragraph.

There are 3 ways out:
1. Let beta-users suffer as is, because this WAI.
2. Add a check in CQ scanning for 'No-Equivalent-Builders: true' lines outside of last paragraph and immediately abort + warn users on incorrect usage.
3. Allow "No-Equivalent-Builders: true" everywhere, not just footer.

Personally, I think #1 is bad and #3 is against normalization of locations of CL metadata into last paragraph and eventually possibly outside of CL description.

i agree that (2) is better than others

Improvement: https://chrome-internal-review.googlesource.com/#/c/infra/infra_internal/+/501172

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/infra/infra_internal/+/5dda6d2fe5d82c044e6141c82014a0ec119fa764

commit 5dda6d2fe5d82c044e6141c82014a0ec119fa764
Author: Andrii Shyshkalov <tandrii@chromium.org>
Date: Sat Nov 11 00:08:11 2017

Tested end-to-end in https://chrome-internal-review.googlesource.com/c/infra/infra_internal/+/502792

Deployed to prod in https://chrome-internal-review.googlesource.com/c/infra/puppet/+/502812