All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=737399

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDgru2E6QsM


Bot(s) for this bug's original alert(s):

chromium-rel-mac12-mini-8gb

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975578945949773824

=== BISECT JOB RESULTS ===
NO Perf regression found

Bisect Details
  Configuration: mac_10_12_mini_8gb_perf_bisect
  Benchmark    : loading.desktop
  Metric       : timeToFirstContentfulPaint_avg/pcv1-warm/Kenh14

Revision             Result                  N
chromium@480308      129.312 +- 72.4654      21      good
chromium@480311      134.676 +- 93.4648      21      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=Kenh14 loading.desktop

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8975578945949773824

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=4962149895503872


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975566116723158096

=== BISECT JOB RESULTS ===
Perf regression found but unable to narrow commit range

Build failures prevented the bisect from narrowing the range further.


Bisect Details
  Configuration: mac_10_12_mini_8gb_perf_bisect
  Benchmark    : loading.desktop
  Metric       : timeToFirstContentfulPaint_avg/pcv1-warm/Kenh14
  Change       : 12.98% | 142.048571429 -> 123.613428571

Suspected Commit Range
  3 commits in range
  https://chromium.googlesource.com/chromium/src/+log/573cfaa63593c27d3702a90881a95ef8c0abee0b..e438353b8b9ecd93216aca5313dc5446c4095102


Revision             Result                  N
chromium@480308      142.049 +- 93.6136      14       good
chromium@480309      ---                     ---      build failure
chromium@480310      ---                     ---      build failure
chromium@480311      123.613 +- 48.3319      14       bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=Kenh14 loading.desktop

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8975566116723158096

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=4962149895503872


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

Trying a wider range.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975515597834979824

=== BISECT JOB RESULTS ===
Bisect was unable to run to completion

Error: INFRA_FAILURE

The bisect was able to narrow the range, you can try running with:
  good_revision: f67c99422c89423538b74def4c562a7df324fc1a
  bad_revision : cd4e730743d522affeae07a1280792359f7856b1

If failures persist contact the team (see below) and report the error.


Bisect Details
  Configuration: mac_10_12_mini_8gb_perf_bisect
  Benchmark    : loading.desktop
  Metric       : timeToFirstContentfulPaint_avg/pcv1-warm/Kenh14

Revision             Result                  N
chromium@480299      133.467 +- 89.8928      21      good
chromium@480328      130.843 +- 37.225       6       good
chromium@480343      164.606 +- 24.7359      9       bad
chromium@480357      161.103 +- 18.9978      14      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=Kenh14 loading.desktop

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8975515597834979824

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5884921471565824


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975498043095874720

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975494829066718976

Kicked off another bisect with the wider range, since that repro-ed.

=== BISECT JOB RESULTS ===
Bisect was unable to run to completion

Error: INFRA_FAILURE

The bisect was able to narrow the range, you can try running with:
  good_revision: 1960d03cec25ebdf6d8d14a06e2eeebf19debcc2
  bad_revision : cd4e730743d522affeae07a1280792359f7856b1

If failures persist contact the team (see below) and report the error.


Bisect Details
  Configuration: mac_10_12_mini_8gb_perf_bisect
  Benchmark    : loading.desktop
  Metric       : timeToFirstContentfulPaint_avg/pcv1-warm/Kenh14

Revision             Result                  N
chromium@480299      117.099 +- 4.46681      6      good
chromium@480328      125.659 +- 16.0074      6      good
chromium@480336      124.697 +- 45.6335      9      good
chromium@480343      163.242 +- 18.262       6      bad
chromium@480357      159.915 +- 14.8529      6      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=Kenh14 loading.desktop

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8975498043095874720

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5884921471565824


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

=== BISECT JOB RESULTS ===
Bisect was unable to run to completion

Error: INFRA_FAILURE

The bisect was able to narrow the range, you can try running with:
  good_revision: d718153981830addd3136882bd536a67c6fa738f
  bad_revision : cd4e730743d522affeae07a1280792359f7856b1

If failures persist contact the team (see below) and report the error.


Bisect Details
  Configuration: mac_10_12_mini_8gb_perf_bisect
  Benchmark    : loading.desktop
  Metric       : timeToFirstContentfulPaint_avg/pcv1-warm/Kenh14

Revision             Result                  N
chromium@480299      130.756 +- 44.363       9       good
chromium@480321      131.01 +- 66.6578       14      good
chromium@480332      156.969 +- 334.822      14      good
chromium@480343      160.99 +- 30.583        14      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=Kenh14 loading.desktop

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8975494829066718976

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5848238155890688


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8975475329083953200

=== BISECT JOB RESULTS ===
Bisect was unable to run to completion

Error: INFRA_FAILURE

The bisect was able to narrow the range, you can try running with:
  good_revision: d718153981830addd3136882bd536a67c6fa738f
  bad_revision : cd4e730743d522affeae07a1280792359f7856b1

If failures persist contact the team (see below) and report the error.


Bisect Details
  Configuration: mac_10_12_mini_8gb_perf_bisect
  Benchmark    : loading.desktop
  Metric       : timeToFirstContentfulPaint_avg/pcv1-warm/Kenh14

Revision             Result                  N
chromium@480299      132.695 +- 66.4315      14      good
chromium@480321      135.378 +- 70.5984      14      good
chromium@480332      142.41 +- 217.92        21      good
chromium@480343      162.736 +- 26.6009      14      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests --story-filter=Kenh14 loading.desktop

Debug Info
  https://chromeperf.appspot.com/buildbucket_job_status/8975475329083953200

Is this bisect wrong?
  https://chromeperf.appspot.com/bad_bisect?try_job_id=5848238155890688


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with component Speed>Bisection.  Thank you!

Weird bisect issues, filed  bug 737971 .

bashi, kouhei: I'm really sorry about the late notice here, but it looks like bisect reproduced this regression from a month ago from https://chromium-review.googlesource.com/538457.

However, this is just one single page and one single device (a mac mini with a spinning disk drive). kouhei: is it worth investigating further, or should we WontFix?

I'm leaning toward WONTFIX, but maybe below is worth trying.
I don't think we should revert the change as the CL fixed the spec violation iiuc. 

Skimming through the CL, I think its object lifetime change may have prolonged context in this specific case accidentally.
bashi: would it be possible to change the Bind in DataTransferItem::getAsString to capture context as weakPersistent?

Sure, I'll give it a try.

bashi, any update here?

bashi, did you try the idea in #18? Should we close this?

I'm sorry that I was not responding this. Let me close this as WontFix. I tried kouhei@'s suggestion on trybot but I wasn't able to get reliable results.