Fuzzer now deduplicates on the analysis side instead of the download side

Fuzzer now deduplicates on the analysis side instead of the download side

This will drastically decrease the stored fuzzes and reduce re-analysis time.

There were some refactorings in this, e.g. putting the logic to download all fuzzes in common/storage, moving frontend/data to data and putting the logic to download data.FuzzReports in a centralized location.

This also has a tiny UI bug that makes details-summary-sk notify when expanded is clicked.

Finally, a pipe "|" separated list of all fuzzes is uploaded to GCS when a commit is "closed", i.e. when we move from revision 'abc' to 'def', we store the names of the bad and grey fuzzes in 'abc' to speed up and make more accurate fuzz counting.
This will be built upon in a follow-on CL.

BUG=skia:4438
Committed: https://skia.googlesource.com/buildbot/+/37adeea89b1ed1b4e9466a849173399b5cf20a44

[kjlubick, 2016-02-11 18:24:38]

Description was changed from

==========
Fuzzer now deduplicates on the analysis side instead of the download side

This will drastically decrease the stored fuzzes and reduce re-analysis time.

BUG=skia:4438
==========

to

==========
Fuzzer now deduplicates on the analysis side instead of the download side

This will drastically decrease the stored fuzzes and reduce re-analysis time.

There were some refactorings in this, e.g. putting the logic to download all
fuzzes in common/storage, moving frontend/data to data and putting the logic to
download data.FuzzReports in a centralized location.

This also has a tiny UI bug that makes details-summary-sk notify when expanded
is clicked.

BUG=skia:4438
==========

[kjlubick, 2016-02-11 18:26:22]

Description was changed from

==========
Fuzzer now deduplicates on the analysis side instead of the download side

This will drastically decrease the stored fuzzes and reduce re-analysis time.

There were some refactorings in this, e.g. putting the logic to download all
fuzzes in common/storage, moving frontend/data to data and putting the logic to
download data.FuzzReports in a centralized location.

This also has a tiny UI bug that makes details-summary-sk notify when expanded
is clicked.

BUG=skia:4438
==========

to

==========
Fuzzer now deduplicates on the analysis side instead of the download side

This will drastically decrease the stored fuzzes and reduce re-analysis time.

There were some refactorings in this, e.g. putting the logic to download all
fuzzes in common/storage, moving frontend/data to data and putting the logic to
download data.FuzzReports in a centralized location.

This also has a tiny UI bug that makes details-summary-sk notify when expanded
is clicked.

Finally, a pipe "|" separated list of all fuzzes is uploaded to GCS when a
commit is "closed", i.e. when we move from revision 'abc' to 'def', we store the
names of the bad and grey fuzzes in 'abc' to speed up and make more accurate
fuzz counting.

BUG=skia:4438
==========

[kjlubick, 2016-02-11 18:26:45]

Description was changed from

==========
Fuzzer now deduplicates on the analysis side instead of the download side

This will drastically decrease the stored fuzzes and reduce re-analysis time.

There were some refactorings in this, e.g. putting the logic to download all
fuzzes in common/storage, moving frontend/data to data and putting the logic to
download data.FuzzReports in a centralized location.

This also has a tiny UI bug that makes details-summary-sk notify when expanded
is clicked.

Finally, a pipe "|" separated list of all fuzzes is uploaded to GCS when a
commit is "closed", i.e. when we move from revision 'abc' to 'def', we store the
names of the bad and grey fuzzes in 'abc' to speed up and make more accurate
fuzz counting.

BUG=skia:4438
==========

to

==========
Fuzzer now deduplicates on the analysis side instead of the download side

This will drastically decrease the stored fuzzes and reduce re-analysis time.

There were some refactorings in this, e.g. putting the logic to download all
fuzzes in common/storage, moving frontend/data to data and putting the logic to
download data.FuzzReports in a centralized location.

This also has a tiny UI bug that makes details-summary-sk notify when expanded
is clicked.

Finally, a pipe "|" separated list of all fuzzes is uploaded to GCS when a
commit is "closed", i.e. when we move from revision 'abc' to 'def', we store the
names of the bad and grey fuzzes in 'abc' to speed up and make more accurate
fuzz counting.
This will be built upon in a follow-on CL.

BUG=skia:4438
==========

[kjlubick, 2016-02-11 18:26:45]

kjlubick@google.com changed reviewers:
+ jcgregorio@google.com

[jcgregorio, 2016-02-12 14:12:25]

lgtm

[kjlubick, 2016-02-12 14:14:34]

The CQ bit was checked by kjlubick@google.com

[commit-bot: I haz the power, 2016-02-12 14:14:39]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1691893002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1691893002/1

[commit-bot: I haz the power, 2016-02-12 14:17:47]

Description was changed from

==========
Fuzzer now deduplicates on the analysis side instead of the download side

This will drastically decrease the stored fuzzes and reduce re-analysis time.

There were some refactorings in this, e.g. putting the logic to download all
fuzzes in common/storage, moving frontend/data to data and putting the logic to
download data.FuzzReports in a centralized location.

This also has a tiny UI bug that makes details-summary-sk notify when expanded
is clicked.

Finally, a pipe "|" separated list of all fuzzes is uploaded to GCS when a
commit is "closed", i.e. when we move from revision 'abc' to 'def', we store the
names of the bad and grey fuzzes in 'abc' to speed up and make more accurate
fuzz counting.
This will be built upon in a follow-on CL.

BUG=skia:4438
==========

to

==========
Fuzzer now deduplicates on the analysis side instead of the download side

This will drastically decrease the stored fuzzes and reduce re-analysis time.

There were some refactorings in this, e.g. putting the logic to download all
fuzzes in common/storage, moving frontend/data to data and putting the logic to
download data.FuzzReports in a centralized location.

This also has a tiny UI bug that makes details-summary-sk notify when expanded
is clicked.

Finally, a pipe "|" separated list of all fuzzes is uploaded to GCS when a
commit is "closed", i.e. when we move from revision 'abc' to 'def', we store the
names of the bad and grey fuzzes in 'abc' to speed up and make more accurate
fuzz counting.
This will be built upon in a follow-on CL.

BUG=skia:4438

Committed:
https://skia.googlesource.com/buildbot/+/37adeea89b1ed1b4e9466a849173399b5cf2...
==========

[commit-bot: I haz the power, 2016-02-12 14:17:48]

Committed patchset #1 (id:1) as
https://skia.googlesource.com/buildbot/+/37adeea89b1ed1b4e9466a849173399b5cf2...