[Android] Fix stack symbolization when packed relocations are on.

[Android] Fix stack symbolization when packed relocations are on.

We had been looking in lib.stripped for the library when symbolizing
debuggerd traces. The libraries in lib.stripped are not the ones with
packed relocations, though, so the resulting symbolization was wrong.

This CL generates a wrapper script around the stack symbolization
script that points to the correct library to use for symbolization.
It also contains a few tests that intentionally crash in various ways.
These are all disabled by default and should only ever be run manually
for testing out instrumentation test crash tooling.

BUG=712265
Review-Url: https://codereview.chromium.org/2840193003
Cr-Commit-Position: refs/heads/master@{#468206}
Committed: https://chromium.googlesource.com/chromium/src/+/276cc56b7c04cc3a28ec6315feb00f056ea8327b

[jbudorick, 2017-04-26 17:18:19]

jbudorick@chromium.org changed reviewers:
+ agrieve@chromium.org

[jbudorick, 2017-04-26 17:18:20]

Noticed this while working on multidex tests.

I think there's room for further work here on both:
 - improving and/or entirely replacing stack + potentially integrating support
for {break,crash}pad dump symbolization
 - a generic build-time script binding mechanism

I experimented with the latter a bit but don't have the time to pursue it
further right now.

[agrieve, 2017-04-26 18:19:44]

lgtm. 

Nice to have this actually work. I know there have been the odd complaint in the
past. Planning on making the bots use the generated script?

https://codereview.chromium.org/2840193003/diff/20001/build/config/android/in...
File build/config/android/internal_rules.gni (right):

https://codereview.chromium.org/2840193003/diff/20001/build/config/android/in...
build/config/android/internal_rules.gni:663: stack_script =
"//third_party/android_platform/development/scripts/stack"
nit: prefix with _

https://codereview.chromium.org/2840193003/diff/20001/build/print_python_deps.py
File build/print_python_deps.py (right):

https://codereview.chromium.org/2840193003/diff/20001/build/print_python_deps...
build/print_python_deps.py:41: if path.endswith('c') and
os.path.exists(path[:-1]):
nit: can you put this back to ".pyc" just so that it's a bit more apparent what
it's looking for?

[jbudorick, 2017-04-26 19:34:22]

I do want the bots to run this in some capacity. There are a few possible ways
that could go, though.

https://codereview.chromium.org/2840193003/diff/20001/build/config/android/in...
File build/config/android/internal_rules.gni (right):

https://codereview.chromium.org/2840193003/diff/20001/build/config/android/in...
build/config/android/internal_rules.gni:663: stack_script =
"//third_party/android_platform/development/scripts/stack"
On 2017/04/26 18:19:44, agrieve wrote:
> nit: prefix with _

Done for this and generated_script.

https://codereview.chromium.org/2840193003/diff/20001/build/print_python_deps.py
File build/print_python_deps.py (right):

https://codereview.chromium.org/2840193003/diff/20001/build/print_python_deps...
build/print_python_deps.py:41: if path.endswith('c') and
os.path.exists(path[:-1]):
On 2017/04/26 18:19:44, agrieve wrote:
> nit: can you put this back to ".pyc" just so that it's a bit more apparent
what
> it's looking for?

The pyc files only end in '.pyc' if the corresponding python file ends in '.py'.
stack's pyc file is 'stackc', hence this change. I've tuned the check a bit,
ptal.

[jbudorick, 2017-04-26 19:34:31]

The CQ bit was checked by jbudorick@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-04-26 19:35:59]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-04-26 21:00:26]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-04-26 21:00:27]

Dry run: This issue passed the CQ dry run.

[agrieve, 2017-04-27 00:30:18]

lgtm

https://codereview.chromium.org/2840193003/diff/20001/build/print_python_deps.py
File build/print_python_deps.py (right):

https://codereview.chromium.org/2840193003/diff/20001/build/print_python_deps...
build/print_python_deps.py:41: if path.endswith('c') and
os.path.exists(path[:-1]):
On 2017/04/26 19:34:22, jbudorick wrote:
> On 2017/04/26 18:19:44, agrieve wrote:
> > nit: can you put this back to ".pyc" just so that it's a bit more apparent
> what
> > it's looking for?
> 
> The pyc files only end in '.pyc' if the corresponding python file ends in
'.py'.
> stack's pyc file is 'stackc', hence this change. I've tuned the check a bit,
> ptal.

did not know that!

[jbudorick, 2017-04-27 00:40:06]

On 2017/04/27 00:30:18, agrieve wrote:
> lgtm
> 
>
https://codereview.chromium.org/2840193003/diff/20001/build/print_python_deps.py
> File build/print_python_deps.py (right):
> 
>
https://codereview.chromium.org/2840193003/diff/20001/build/print_python_deps...
> build/print_python_deps.py:41: if path.endswith('c') and
> os.path.exists(path[:-1]):
> On 2017/04/26 19:34:22, jbudorick wrote:
> > On 2017/04/26 18:19:44, agrieve wrote:
> > > nit: can you put this back to ".pyc" just so that it's a bit more apparent
> > what
> > > it's looking for?
> > 
> > The pyc files only end in '.pyc' if the corresponding python file ends in
> '.py'.
> > stack's pyc file is 'stackc', hence this change. I've tuned the check a bit,
> > ptal.
> 
> did not know that!

I didn't either, and it took me a minute to figure out why 'stackc' was in the
pydeps file rather than 'stack' :)

[jbudorick, 2017-04-27 00:45:53]

jbudorick@chromium.org changed reviewers:
+ dpranke@chromium.org, tedchoc@chromium.org

[jbudorick, 2017-04-27 00:45:54]

+dpranke for PRESUBMIT.py
+tedchoc for IntentionalCrashTest

[Dirk Pranke, 2017-04-27 00:49:40]

lgtm

[Ted C, 2017-04-27 16:09:45]

lgtm

https://codereview.chromium.org/2840193003/diff/40001/chrome/android/javatest...
File
chrome/android/javatests/src/org/chromium/chrome/test/crash/IntentionalCrashTest.java
(right):

https://codereview.chromium.org/2840193003/diff/40001/chrome/android/javatest...
chrome/android/javatests/src/org/chromium/chrome/test/crash/IntentionalCrashTest.java:23:
@CommandLineFlags.Add("disable-fre")
ChromeSwitches.DISABLE_FIRST_RUN_EXPERIENCE

[jbudorick, 2017-04-28 23:00:14]

https://codereview.chromium.org/2840193003/diff/40001/chrome/android/javatest...
File
chrome/android/javatests/src/org/chromium/chrome/test/crash/IntentionalCrashTest.java
(right):

https://codereview.chromium.org/2840193003/diff/40001/chrome/android/javatest...
chrome/android/javatests/src/org/chromium/chrome/test/crash/IntentionalCrashTest.java:23:
@CommandLineFlags.Add("disable-fre")
On 2017/04/27 16:09:45, Ted C wrote:
> ChromeSwitches.DISABLE_FIRST_RUN_EXPERIENCE

Done.

[jbudorick, 2017-04-28 23:00:21]

The CQ bit was checked by jbudorick@chromium.org

[jbudorick, 2017-04-28 23:00:21]

The patchset sent to the CQ was uploaded after l-g-t-m from
tedchoc@chromium.org, dpranke@chromium.org, agrieve@chromium.org
Link to the patchset: https://codereview.chromium.org/2840193003/#ps60001
(title: "tedchoc comment")

[commit-bot: I haz the power, 2017-04-28 23:01:04]

CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-04-29 01:35:02]

CQ is committing da patch.

Bot data: {"patchset_id": 60001, "attempt_start_ts": 1493420421405340,
"parent_rev": "b6da435540cdb1774a04fcf9a45c98158caf299c", "commit_rev":
"276cc56b7c04cc3a28ec6315feb00f056ea8327b"}

[commit-bot: I haz the power, 2017-04-29 01:35:13]

Description was changed from

==========
[Android] Fix stack symbolization when packed relocations are on.

We had been looking in lib.stripped for the library when symbolizing
debuggerd traces. The libraries in lib.stripped are not the ones with
packed relocations, though, so the resulting symbolization was wrong.

This CL generates a wrapper script around the stack symbolization
script that points to the correct library to use for symbolization.
It also contains a few tests that intentionally crash in various ways.
These are all disabled by default and should only ever be run manually
for testing out instrumentation test crash tooling.

BUG=712265
==========

to

==========
[Android] Fix stack symbolization when packed relocations are on.

We had been looking in lib.stripped for the library when symbolizing
debuggerd traces. The libraries in lib.stripped are not the ones with
packed relocations, though, so the resulting symbolization was wrong.

This CL generates a wrapper script around the stack symbolization
script that points to the correct library to use for symbolization.
It also contains a few tests that intentionally crash in various ways.
These are all disabled by default and should only ever be run manually
for testing out instrumentation test crash tooling.

BUG=712265

Review-Url: https://codereview.chromium.org/2840193003
Cr-Commit-Position: refs/heads/master@{#468206}
Committed:
https://chromium.googlesource.com/chromium/src/+/276cc56b7c04cc3a28ec6315feb0...
==========

[commit-bot: I haz the power, 2017-04-29 01:35:18]

Committed patchset #4 (id:60001) as
https://chromium.googlesource.com/chromium/src/+/276cc56b7c04cc3a28ec6315feb0...