Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8972523541592376352

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8972523371265384448

=== Auto-CCing suspected CL author aseemgarg@google.com ===

Hi aseemgarg@google.com, the bisect results pointed to your CL, please take a look at the
results.


=== BISECT JOB RESULTS ===
Test failure found with culprit

Suspected Commit
  Author : Aseem Garg
  Commit : eb65f35e96a417b7792400a3867560248d9809f2
  Date   : Tue Jul 25 22:32:39 2017
  Subject: [wasm] redirect wasm calls to js functions through a GCed table

Bisect Details
  Configuration: android_one_perf_bisect
  Benchmark    : jetstream
  Metric       : Score/Score

Revision                           Exit Code      N
chromium@489599                    0 +- N/A       2      good
chromium@489628                    0 +- N/A       2      good
chromium@489630                    0 +- N/A       2      good
chromium@489630,v8@4fe1d71509      0 +- N/A       2      good
chromium@489630,v8@eb65f35e96      1 +- N/A       2      bad       <--
chromium@489630,v8@60c0ba3285      1 +- N/A       2      bad
chromium@489631                    1 +- N/A       2      bad
chromium@489632                    1 +- N/A       2      bad
chromium@489635                    1 +- N/A       2      bad
chromium@489642                    1 +- N/A       2      bad
chromium@489656                    1 +- N/A       2      bad
chromium@489713                    1 +- N/A       2      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests jetstream

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8972523541592376352


For feedback, file a bug with component Speed>Bisection

=== BISECT JOB RESULTS ===
Test failure found with culprit

Suspected Commit
  Author : Aseem Garg
  Commit : eb65f35e96a417b7792400a3867560248d9809f2
  Date   : Tue Jul 25 22:32:39 2017
  Subject: [wasm] redirect wasm calls to js functions through a GCed table

Bisect Details
  Configuration: android_nexus6_perf_bisect
  Benchmark    : jetstream
  Metric       : Score/Score

Revision                           Exit Code      N
chromium@489277                    0 +- N/A       5      good
chromium@489518                    0 +- N/A       5      good
chromium@489578                    0 +- N/A       5      good
chromium@489608                    0 +- N/A       5      good
chromium@489623                    0 +- N/A       5      good
chromium@489627                    0 +- N/A       5      good
chromium@489629                    0 +- N/A       5      good
chromium@489630                    0 +- N/A       5      good
chromium@489630,v8@4fe1d71509      0 +- N/A       5      good
chromium@489630,v8@eb65f35e96      1 +- N/A       5      bad       <--
chromium@489630,v8@60c0ba3285      1 +- N/A       5      bad
chromium@489631                    1 +- N/A       5      bad
chromium@489638                    1 +- N/A       5      bad
chromium@489758                    1 +- N/A       5      bad
chromium@490239                    1 +- N/A       5      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests jetstream

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8972523371265384448


For feedback, file a bug with component Speed>Bisection

Revert: https://chromium-review.googlesource.com/c/596268

@nednguyen Can you point me to what the failure is? I am having a hard time trying to parse the logs. 

The following revision refers to this bug:
  https://chromium.googlesource.com/v8/v8.git/+/25f03308a7c34df996e93178ba4d29de8949349a

commit 25f03308a7c34df996e93178ba4d29de8949349a
Author: Annie Sullivan <sullivan@chromium.org>
Date: Tue Aug 01 21:34:25 2017

Revert "[wasm] redirect wasm calls to js functions through a GCed table"

This reverts commit eb65f35e96a417b7792400a3867560248d9809f2.

Reason for revert: Broke jetstream benchmark on android.

BUG= chromium:750828 

Original change's description:
> [wasm] redirect wasm calls to js functions through a GCed table
> 
> With this patch, rather than embedding the JSReceiver address directly
> in the WasmToJS wrappers, we put that in a fixed array with global handle
> scope and instead embed the location of the handle and the index in the
> wrapper. This ensures that the wrapper doesn't need to be patched if the
> GC kicks in. This is needed to get the WASM code off the GCed heap.
> 
> R=​mtrofin@chromium.org
> 
> Bug: 
> Change-Id: Ie5a77a78cdecec51b04f702c63b8e4285e6a2d8d
> Reviewed-on: https://chromium-review.googlesource.com/581682
> Commit-Queue: Aseem Garg <aseemgarg@chromium.org>
> Reviewed-by: Mircea Trofin <mtrofin@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#46884}

TBR=mtrofin@chromium.org,aseemgarg@google.com,aseemgarg@chromium.org,clemensh@chromium.org

# Not skipping CQ checks because original CL landed > 1 day ago.

Change-Id: I26f49ee0a1fe73cc5d8852ded87b56638be39ebf
Reviewed-on: https://chromium-review.googlesource.com/596268
Commit-Queue: Aseem Garg <aseemgarg@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47059}
[modify] https://crrev.com/25f03308a7c34df996e93178ba4d29de8949349a/src/compiler/wasm-compiler.cc
[modify] https://crrev.com/25f03308a7c34df996e93178ba4d29de8949349a/src/compiler/wasm-compiler.h
[modify] https://crrev.com/25f03308a7c34df996e93178ba4d29de8949349a/src/wasm/module-compiler.cc
[modify] https://crrev.com/25f03308a7c34df996e93178ba4d29de8949349a/src/wasm/wasm-interpreter.cc
[modify] https://crrev.com/25f03308a7c34df996e93178ba4d29de8949349a/src/wasm/wasm-module.cc
[modify] https://crrev.com/25f03308a7c34df996e93178ba4d29de8949349a/src/wasm/wasm-objects.h
[modify] https://crrev.com/25f03308a7c34df996e93178ba4d29de8949349a/test/cctest/wasm/test-run-wasm-js.cc
[modify] https://crrev.com/25f03308a7c34df996e93178ba4d29de8949349a/test/cctest/wasm/test-wasm-stack.cc
[modify] https://crrev.com/25f03308a7c34df996e93178ba4d29de8949349a/test/cctest/wasm/wasm-run-utils.h

Landed the revert. Taking a look at underlying problem and re land the CL

Log in https://chromium-swarm.appspot.com/task?id=3798c7d5e4971110&refresh=10&show_raw=1

From some reason, Chrome native stack is missing:

Stack from Logcat
	Reading native crash info from stdin
	Unknown Android release, consider passing --packed-lib.
	Reading Android symbols from: /b/swarming/w/ir
	Searching for Chrome symbols from within: /b/swarming/w/ir/out/Release/lib.unstripped:/b/swarming/w/ir/out/Release/lib:/b/swarming/w/ir/out/Release


John: you have any idea why the native crash would be missing here?

Multiple re-runs are passing these bots. Probably issue with some other commit that got fixed with rebase. Marking as not reproducible.
https://chromium-review.googlesource.com/c/597010

The following revision refers to this bug:
  https://chromium.googlesource.com/v8/v8.git/+/649b1e70e73dbdc98fe9655f454f0ace0113c990

commit 649b1e70e73dbdc98fe9655f454f0ace0113c990
Author: Aseem Garg <aseemgarg@google.com>
Date: Sat Aug 05 01:57:25 2017

[wasm] Reland "[wasm] redirect wasm calls to js functions through a GCed table"

This reverts commit 25f03308a7c34df996e93178ba4d29de8949349a.

Reason for revert: Fix the cause of bot failure and reland

Original change's description:
> Revert "[wasm] redirect wasm calls to js functions through a GCed table"
> 
> This reverts commit eb65f35e96a417b7792400a3867560248d9809f2.
> 
> Reason for revert: Broke jetstream benchmark on android.
> 
> BUG= chromium:750828 
> 
> Original change's description:
> > [wasm] redirect wasm calls to js functions through a GCed table
> > 
> > With this patch, rather than embedding the JSReceiver address directly
> > in the WasmToJS wrappers, we put that in a fixed array with global handle
> > scope and instead embed the location of the handle and the index in the
> > wrapper. This ensures that the wrapper doesn't need to be patched if the
> > GC kicks in. This is needed to get the WASM code off the GCed heap.
> > 
> > R=​mtrofin@chromium.org
> > 
> > Bug: 
> > Change-Id: Ie5a77a78cdecec51b04f702c63b8e4285e6a2d8d
> > Reviewed-on: https://chromium-review.googlesource.com/581682
> > Commit-Queue: Aseem Garg <aseemgarg@chromium.org>
> > Reviewed-by: Mircea Trofin <mtrofin@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#46884}
> 
> TBR=mtrofin@chromium.org,aseemgarg@google.com,aseemgarg@chromium.org,clemensh@chromium.org
> 
> # Not skipping CQ checks because original CL landed > 1 day ago.
> 
> Change-Id: I26f49ee0a1fe73cc5d8852ded87b56638be39ebf
> Reviewed-on: https://chromium-review.googlesource.com/596268
> Commit-Queue: Aseem Garg <aseemgarg@chromium.org>
> Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#47059}

R=mtrofin@chromium.org,aseemgarg@google.com,aseemgarg@chromium.org,clemensh@chromium.org,sullivan@chromium.org

Change-Id: I29ef35f6e612a706d9f571da3e7beb1da8b5052b
Bug:  chromium:750828 
Reviewed-on: https://chromium-review.googlesource.com/597010
Commit-Queue: Aseem Garg <aseemgarg@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47177}
[modify] https://crrev.com/649b1e70e73dbdc98fe9655f454f0ace0113c990/src/compiler/wasm-compiler.cc
[modify] https://crrev.com/649b1e70e73dbdc98fe9655f454f0ace0113c990/src/compiler/wasm-compiler.h
[modify] https://crrev.com/649b1e70e73dbdc98fe9655f454f0ace0113c990/src/wasm/module-compiler.cc
[modify] https://crrev.com/649b1e70e73dbdc98fe9655f454f0ace0113c990/src/wasm/wasm-interpreter.cc
[modify] https://crrev.com/649b1e70e73dbdc98fe9655f454f0ace0113c990/src/wasm/wasm-module.cc
[modify] https://crrev.com/649b1e70e73dbdc98fe9655f454f0ace0113c990/src/wasm/wasm-objects.h
[modify] https://crrev.com/649b1e70e73dbdc98fe9655f454f0ace0113c990/test/cctest/wasm/test-run-wasm-js.cc
[modify] https://crrev.com/649b1e70e73dbdc98fe9655f454f0ace0113c990/test/cctest/wasm/test-wasm-stack.cc
[modify] https://crrev.com/649b1e70e73dbdc98fe9655f454f0ace0113c990/test/cctest/wasm/wasm-run-utils.h

You unreverted this, and jetstream started failing again at this point:
https://build.chromium.org/p/chromium.perf/builders/Android%20One%20Perf/builds/361

It contains a v8-autoroll@chromium.org CL. 
https://chromium.googlesource.com/v8/v8/+log/3b568044..277ee04e

That roll contains your CL. What devices did you try to reproduce this on locally, and with what versions of android? I'll revert your CL again.

I can keep an eye on this bot, and if I am wrong it should be pretty obvious and the test will continue to fail, and it just happened to start failing again when that CL re-landed I will unrevert this re-revert. But if it starts passing again, we need to figure out why.

Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8971902157972497552

I didn't try to reproduce locally. I triggered the chromium perf bots for all-android

tools/perf/run_benchmark try --repo_path v8 all-android blink_perf.bindings

=== Auto-CCing suspected CL author aseemgarg@google.com ===

Hi aseemgarg@google.com, the bisect results pointed to your CL, please take a look at the
results.


=== BISECT JOB RESULTS ===
Test failure found with culprit

Suspected Commit
  Author : Aseem Garg
  Commit : 649b1e70e73dbdc98fe9655f454f0ace0113c990
  Date   : Sat Aug 05 01:57:25 2017
  Subject: [wasm] Reland "[wasm] redirect wasm calls to js functions through a GCed table"

Bisect Details
  Configuration: android_one_perf_bisect
  Benchmark    : jetstream
  Metric       : 3d-raytrace/http___browserbench.org_JetStream_

Revision                           Exit Code      N
chromium@492231                    0 +- N/A       2      good
chromium@492231,v8@57af86a16e      0 +- N/A       2      good
chromium@492231,v8@649b1e70e7      1 +- N/A       2      bad       <--
chromium@492231,v8@90e8624ea0      1 +- N/A       2      bad
chromium@492232                    1 +- N/A       2      bad
chromium@492233                    1 +- N/A       2      bad
chromium@492234                    1 +- N/A       2      bad
chromium@492236                    1 +- N/A       2      bad

To Run This Test
  src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests jetstream

More information on addressing performance regressions:
  http://g.co/ChromePerformanceRegressions

Debug information about this bisect:
  https://chromeperf.appspot.com/buildbucket_job_status/8971902157972497552


For feedback, file a bug with component Speed>Bisection

Its slightly disheartening to hear that a perf tryjob came back successful, when the waterfall and bisect both show failures.

I dont know that much about tryjobs. I think simonhatch@ might be the right person to rope in on that front? CCing them. I see you are in SFO, I have an A1 device you can borrow if you need to test locally. 

Sorry, what's the issue here? It looks like jetstream is failing? That bisect is on return_code, and seems to have identified a culprit.

I tried running this earlier (before re-landing the CL) and it passed. I guess will try to repro locally on a device. @Simon, I have a Nexus 6 device and will try to repro on that. If that doesn't work, then will grab the A1 device from you. Thanks

The following revision refers to this bug:
  https://chromium.googlesource.com/v8/v8.git/+/aee1e1fb8df597bedf41281e4733e8d398b43bf1

commit aee1e1fb8df597bedf41281e4733e8d398b43bf1
Author: Aseem Garg <aseemgarg@google.com>
Date: Tue Aug 08 19:37:45 2017

Revert "[wasm] Reland "[wasm] redirect wasm calls to js functions through a GCed table""

This reverts commit 649b1e70e73dbdc98fe9655f454f0ace0113c990.

Reason for revert: A1 Jetstream bots are still failing.

Original change's description:
> [wasm] Reland "[wasm] redirect wasm calls to js functions through a GCed table"
>
> This reverts commit 25f03308a7c34df996e93178ba4d29de8949349a.
>
> Reason for revert: Fix the cause of bot failure and reland
>
> Original change's description:
> > Revert "[wasm] redirect wasm calls to js functions through a GCed table"
> >
> > This reverts commit eb65f35e96a417b7792400a3867560248d9809f2.
> >
> > Reason for revert: Broke jetstream benchmark on android.
> >
> > BUG= chromium:750828 
> >
> > Original change's description:
> > > [wasm] redirect wasm calls to js functions through a GCed table
> > >
> > > With this patch, rather than embedding the JSReceiver address directly
> > > in the WasmToJS wrappers, we put that in a fixed array with global handle
> > > scope and instead embed the location of the handle and the index in the
> > > wrapper. This ensures that the wrapper doesn't need to be patched if the
> > > GC kicks in. This is needed to get the WASM code off the GCed heap.
> > >
> > > R=​mtrofin@chromium.org
> > >
> > > Bug:
> > > Change-Id: Ie5a77a78cdecec51b04f702c63b8e4285e6a2d8d
> > > Reviewed-on: https://chromium-review.googlesource.com/581682
> > > Commit-Queue: Aseem Garg <aseemgarg@chromium.org>
> > > Reviewed-by: Mircea Trofin <mtrofin@chromium.org>
> > > Cr-Commit-Position: refs/heads/master@{#46884}
> >
> > TBR=mtrofin@chromium.org,aseemgarg@google.com,aseemgarg@chromium.org,clemensh@chromium.org
> >
> > # Not skipping CQ checks because original CL landed > 1 day ago.
> >
> > Change-Id: I26f49ee0a1fe73cc5d8852ded87b56638be39ebf
> > Reviewed-on: https://chromium-review.googlesource.com/596268
> > Commit-Queue: Aseem Garg <aseemgarg@chromium.org>
> > Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#47059}
>
> R=​mtrofin@chromium.org,aseemgarg@google.com,aseemgarg@chromium.org,clemensh@chromium.org,sullivan@chromium.org
>
> Change-Id: I29ef35f6e612a706d9f571da3e7beb1da8b5052b
> Bug:  chromium:750828 
> Reviewed-on: https://chromium-review.googlesource.com/597010
> Commit-Queue: Aseem Garg <aseemgarg@chromium.org>
> Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#47177}

TBR=bradnelson@chromium.org,sullivan@chromium.org,mtrofin@chromium.org,aseemgarg@google.com,aseemgarg@chromium.org,clemensh@chromium.org

Bug:  chromium:750828 
Change-Id: I04b12c0eb0705ad809822a7d7461423be77d942a
Reviewed-on: https://chromium-review.googlesource.com/606867
Commit-Queue: Aseem Garg <aseemgarg@chromium.org>
Reviewed-by: Aseem Garg <aseemgarg@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47231}
[modify] https://crrev.com/aee1e1fb8df597bedf41281e4733e8d398b43bf1/src/compiler/wasm-compiler.cc
[modify] https://crrev.com/aee1e1fb8df597bedf41281e4733e8d398b43bf1/src/compiler/wasm-compiler.h
[modify] https://crrev.com/aee1e1fb8df597bedf41281e4733e8d398b43bf1/src/wasm/module-compiler.cc
[modify] https://crrev.com/aee1e1fb8df597bedf41281e4733e8d398b43bf1/src/wasm/wasm-interpreter.cc
[modify] https://crrev.com/aee1e1fb8df597bedf41281e4733e8d398b43bf1/src/wasm/wasm-module.cc
[modify] https://crrev.com/aee1e1fb8df597bedf41281e4733e8d398b43bf1/src/wasm/wasm-objects.h
[modify] https://crrev.com/aee1e1fb8df597bedf41281e4733e8d398b43bf1/test/cctest/wasm/test-run-wasm-js.cc
[modify] https://crrev.com/aee1e1fb8df597bedf41281e4733e8d398b43bf1/test/cctest/wasm/test-wasm-stack.cc
[modify] https://crrev.com/aee1e1fb8df597bedf41281e4733e8d398b43bf1/test/cctest/wasm/wasm-run-utils.h

 Issue 753480  has been merged into this issue.

Mircea was trying to understand whether his new patch has fixed the problem. (He has tested locally.) But now this test is still failing on Android One Perf. And the log does not help much because it is just saying Android kills Chrome.

Ned, is there any way to figure this out?

Ned is on vacation until 8/30. So Mircea ran this on an A1 perf try bot, it passed, and then he submitted it and it failed on the waterfall?

I think the case is that he didn't know if his test caused its failing, because it has been failing for a long time and the log does not give much useful information. Mircea can give you more details.

I think this is not failing due to Mircea's commit. I tried to revert Mircea's change and run locally on a device and it still crashed.

Just for reference, I have a change that failed this. This is a run without my patch:
https://build.chromium.org/p/tryserver.chromium.perf/builders/android_one_perf_bisect/builds/2648/steps/Running%20WITHOUT%20patch.Performance%20Test%20%28Without%20Patch%29%201%20of%201

This is the perf_bisect run:
https://build.chromium.org/p/tryserver.chromium.perf/builders/android_one_perf_bisect/builds/2648

Expanding steps 20 and 22 show the failures.

Log from my local run with Mircea's change reverted.

Deleted: log_verbose.txt 65.8 KB

log_verbose.txt 65.8 KB View Download

This is my fault. See https://bugs.chromium.org/p/chromium/issues/detail?id=758309#c3 for the reason.

Ah, never mind: my linked-to reason was only responsible for the recent jetstream failures. Sorry for the noise.

Assem, in #c16, you said that you tried to reproduce the problem on the perf try bots with:

tools/perf/run_benchmark try --repo_path v8 all-android blink_perf.bindings

However, it was jetstream that was failing, not blink_perf.bindings. Is it possible that the issue was that you launched a perf try job for the wrong benchmark?

Yeah. But tried with Jetstream too. It's failing both with and without my patch it seems..

The following revision refers to this bug:
  https://chromium.googlesource.com/v8/v8.git/+/5f3a2def8245bea904fa4ed0ed0a04d167b3f8a2

commit 5f3a2def8245bea904fa4ed0ed0a04d167b3f8a2
Author: Aseem Garg <aseemgarg@chromium.org>
Date: Fri Sep 08 04:36:12 2017

[wasm] redirect wasm calls to js functions through a GCed table

This is revert of commit aee1e1fb8df597bedf41281e4733e8d398b43bf1 with the fix for A1 and N6 jetstream failure.

R=bradnelson@chromium.org,mtrofin@chromium.org,clemensh@chromium.org
Bug:  chromium:750828 

Change-Id: Id38896af51315f76a0667ace32c77a2ba7287eec
Reviewed-on: https://chromium-review.googlesource.com/607092
Commit-Queue: Mircea Trofin <mtrofin@chromium.org>
Reviewed-by: Mircea Trofin <mtrofin@chromium.org>
Reviewed-by: Brad Nelson <bradnelson@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47910}
[modify] https://crrev.com/5f3a2def8245bea904fa4ed0ed0a04d167b3f8a2/src/compiler/wasm-compiler.cc
[modify] https://crrev.com/5f3a2def8245bea904fa4ed0ed0a04d167b3f8a2/src/compiler/wasm-compiler.h
[modify] https://crrev.com/5f3a2def8245bea904fa4ed0ed0a04d167b3f8a2/src/wasm/module-compiler.cc
[modify] https://crrev.com/5f3a2def8245bea904fa4ed0ed0a04d167b3f8a2/src/wasm/wasm-interpreter.cc
[modify] https://crrev.com/5f3a2def8245bea904fa4ed0ed0a04d167b3f8a2/src/wasm/wasm-objects-inl.h
[modify] https://crrev.com/5f3a2def8245bea904fa4ed0ed0a04d167b3f8a2/src/wasm/wasm-objects.h
[modify] https://crrev.com/5f3a2def8245bea904fa4ed0ed0a04d167b3f8a2/test/cctest/wasm/test-run-wasm-js.cc
[modify] https://crrev.com/5f3a2def8245bea904fa4ed0ed0a04d167b3f8a2/test/cctest/wasm/test-wasm-stack.cc
[modify] https://crrev.com/5f3a2def8245bea904fa4ed0ed0a04d167b3f8a2/test/cctest/wasm/wasm-run-utils.cc
[modify] https://crrev.com/5f3a2def8245bea904fa4ed0ed0a04d167b3f8a2/test/cctest/wasm/wasm-run-utils.h

Aseem, when you say "with the fix for A1 and N6 jetstream failure", does that mean that you figured out what was going on?

yes. We had formed a cycle of references with a global rooting my new table that caused the wasm instance to not get GCed. Fixed that by changing the global to be a weak reference.

Great! Glad that you were able to identify the problem. Any idea why it wasn't reproducing on the trybots?

So, it was. I was looking in the wrong place. I just saw the perf-bisects to be green and assumed that all went fine. But if I expanded the with patch and without patch parts, it was failing with patch and not without (at some point both started failing due to another CL that created a similar problem. Both are fixed now). 

Ahhh, okay. Sorry about the terrible UI: in the current world, that UI is about as decent a one as we can come up with to display try job results. dtu@ is working on a new UI that will be a lot more intuitive.