Make simdjs tests redownload on a revision change.

Make simdjs tests redownload on a revision change.

LOG=N
BUG=None
TEST=manual,local
R=bbudge@chromium.org,machenbach@chromium.org

Committed: https://crrev.com/0c53c66938fa59330f65f778f71bfce168704940
Cr-Commit-Position: refs/heads/master@{#29780}

[bbudge, 2015-07-17 00:09:37]

LGTM
Still need to delete old data and tar.gz file?

[Dan Ehrenberg, 2015-07-17 00:44:51]

+1!

[bradn, 2015-07-17 00:57:39]

Not clear on what you guys are asking for?

You'd rather it not keep around the last version and/or the tar.gz?

[bbudge, 2015-07-17 01:04:33]

On 2015/07/17 00:57:39, bradn wrote:
> Not clear on what you guys are asking for?
> 
> You'd rather it not keep around the last version and/or the tar.gz?

That was just FYI. I don't know what best practices are here. I bet Dan is just
enthusiastic that we're switching the tests.

[bradn, 2015-07-18 22:13:59]

Ping machenbach for OWNERS?

[Dan Ehrenberg, 2015-07-21 00:47:27]

littledan@chromium.org changed reviewers:
+ littledan@chromium.org

[Dan Ehrenberg, 2015-07-21 00:47:27]

lgtm

[Michael Achenbach, 2015-07-21 06:58:18]

lgtm with comments:

Please always delete old archive files. You could e.g. do a clobber like here in
test262 (I'll refactor and extract this code at some point):
https://chromium.googlesource.com/v8/v8/+/5e5fadb7/test/test262-es6/testcfg.p...

An update of the test suite must be cleanly revertable. We've had problems with
test262 in the past when old archive files where lying around and a test262
update got rolled back.

https://codereview.chromium.org/1240453003/diff/1/test/simdjs/testcfg.py
File test/simdjs/testcfg.py (right):

https://codereview.chromium.org/1240453003/diff/1/test/simdjs/testcfg.py#newc...
test/simdjs/testcfg.py:88: checked_out_revision != revision):
Please add a print statement here that logs what's happening. Helps debugging
later.

[bradn, 2015-07-21 18:37:19]

https://codereview.chromium.org/1240453003/diff/1/test/simdjs/testcfg.py
File test/simdjs/testcfg.py (right):

https://codereview.chromium.org/1240453003/diff/1/test/simdjs/testcfg.py#newc...
test/simdjs/testcfg.py:88: checked_out_revision != revision):
On 2015/07/21 06:58:17, Michael Achenbach wrote:
> Please add a print statement here that logs what's happening. Helps debugging
> later.

Done.

[bradn, 2015-07-21 18:37:42]

The CQ bit was checked by bradnelson@google.com

[bradn, 2015-07-21 18:37:42]

The patchset sent to the CQ was uploaded after l-g-t-m from bbudge@chromium.org,
littledan@chromium.org, machenbach@chromium.org
Link to the patchset: https://codereview.chromium.org/1240453003/#ps20001
(title: "fix")

[commit-bot: I haz the power, 2015-07-21 18:37:50]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1240453003/20001

[commit-bot: I haz the power, 2015-07-21 19:02:33]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-07-21 19:02:42]

Patchset 2 (id:??) landed as
https://crrev.com/0c53c66938fa59330f65f778f71bfce168704940
Cr-Commit-Position: refs/heads/master@{#29780}

[Michael Achenbach, 2015-07-23 21:05:13]

Hmm - something is not right yet. It looks like it downloads on every build,
like:
http://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20arm%20-%20sim...

Same vm, same folder, same builder - but continues to redownload. Some
speculations below, could you take a look?

https://codereview.chromium.org/1240453003/diff/20001/test/simdjs/testcfg.py
File test/simdjs/testcfg.py (right):

https://codereview.chromium.org/1240453003/diff/20001/test/simdjs/testcfg.py#...
test/simdjs/testcfg.py:74: versionfile = os.path.join(self.root,
"CHECKED_OUT_VERSION")
Maybe this file needs to be added to the .gitignores file in the project root?
It's not there right now and maybe gets clobbered by bot_update?

https://codereview.chromium.org/1240453003/diff/20001/test/simdjs/testcfg.py#...
test/simdjs/testcfg.py:134: f.write(SIMDJS_ARCHIVE_MD5 + '\n')
Maybe the line splitting doesn't work as expected? Double line breaks or
something?

[bradn, 2015-07-23 22:41:24]

It's doing it due to the version file not being in .gitignore.
So the bots are reverting it on each run.
Sending out a change.