Remove dependency of NaCl untarred toolchain from NaCl SDK Tests.

Remove dependency of NaCl untarred toolchain from NaCl SDK Tests.

Nothing on the chromium side should depend on the untarred toolchains
within the NaCl tree. Have NaCl SDK tests untar its own copy of what it
needs before proceeding.

R=binji@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=263791

[David Yen, 2014-04-14 21:50:51]

This is what we discussed last week about removing the dependency on the NaCl
toolchain for the NaCl SDK tests. Most of test_all.py will be replaced, once
this CL lands I will put the changes necessary inside of this CL:

https://codereview.chromium.org/235603004/

[binji, 2014-04-14 22:17:26]

lgtm, please run trybots.

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/test_a...
File native_client_sdk/src/test_all.py (right):

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/test_a...
native_client_sdk/src/test_all.py:8: import shutil
alphabetize

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/test_a...
native_client_sdk/src/test_all.py:14: TOOLS_DIR = os.path.join(SCRIPT_DIR,
'tools')
TODO remove this stuff too (like below)?

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/test_a...
native_client_sdk/src/test_all.py:25: TOOLCHAIN_OUT =
os.path.join(build_paths.OUT_DIR, 'sdk_tests', 'toolchain')
move this into build_paths (or does it go away after we have
package_version.py)?

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/test_a...
native_client_sdk/src/test_all.py:51: # TODO(dyen): This function is basically
all temporary code. After the DEPS
Add a bug for this, unless this is all going to happen very soon.

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/tools/...
File native_client_sdk/src/tools/tests/create_nmf_test.py (right):

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/tools/...
native_client_sdk/src/tools/tests/create_nmf_test.py:17: CHROME_SRC =
os.path.dirname(os.path.dirname(os.path.dirname(TOOLS_DIR)))
unused?

[David Yen, 2014-04-14 22:31:05]

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/test_a...
File native_client_sdk/src/test_all.py (right):

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/test_a...
native_client_sdk/src/test_all.py:8: import shutil
On 2014/04/14 22:17:26, binji wrote:
> alphabetize

Done.

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/test_a...
native_client_sdk/src/test_all.py:14: TOOLS_DIR = os.path.join(SCRIPT_DIR,
'tools')
On 2014/04/14 22:17:26, binji wrote:
> TODO remove this stuff too (like below)?

These will stay, only sys.path.append(TOOLS_DIR) is no longer needed.

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/test_a...
native_client_sdk/src/test_all.py:25: TOOLCHAIN_OUT =
os.path.join(build_paths.OUT_DIR, 'sdk_tests', 'toolchain')
On 2014/04/14 22:17:26, binji wrote:
> move this into build_paths (or does it go away after we have
> package_version.py)?

This is the output directory for the nacl_x86_glibc toolchain so it won't go
away, but build_paths is a general library for build_tools so it doesn't really
make sense to put this there. If we really want to share the code we could make
another test_paths.py or something?

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/test_a...
native_client_sdk/src/test_all.py:51: # TODO(dyen): This function is basically
all temporary code. After the DEPS
On 2014/04/14 22:17:26, binji wrote:
> Add a bug for this, unless this is all going to happen very soon.

This is going to happen very soon, I will add the change to a CL right after
this lands.

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/tools/...
File native_client_sdk/src/tools/tests/create_nmf_test.py (right):

https://codereview.chromium.org/237883002/diff/1/native_client_sdk/src/tools/...
native_client_sdk/src/tools/tests/create_nmf_test.py:17: CHROME_SRC =
os.path.dirname(os.path.dirname(os.path.dirname(TOOLS_DIR)))
On 2014/04/14 22:17:26, binji wrote:
> unused?

It's used in the line below to get MOCK_DIR. I've removed it from the other test
file though.

[David Yen, 2014-04-14 23:46:29]

The CQ bit was checked by dyen@chromium.org

[commit-bot: I haz the power, 2014-04-14 23:46:46]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dyen@chromium.org/237883002/20001

[commit-bot: I haz the power, 2014-04-15 05:30:23]

Change committed as 263791

[binji, 2014-04-15 16:35:27]

On 2014/04/15 05:30:23, I haz the power (commit-bot) wrote:
> Change committed as 263791

This broke the windows bots, as you can see from the win_nacl_sdk and
win_nacl_sdk_build trybot runs. Unfortunately, the sdk builders still do not
block the CQ so you will have to be more careful.

[David Yen, 2014-04-15 16:39:39]

On 2014/04/15 16:35:27, binji wrote:
> On 2014/04/15 05:30:23, I haz the power (commit-bot) wrote:
> > Change committed as 263791
> 
> This broke the windows bots, as you can see from the win_nacl_sdk and
> win_nacl_sdk_build trybot runs. Unfortunately, the sdk builders still do not
> block the CQ so you will have to be more careful.

Ah okay, I assumed it would block the CQ if something was my fault. Do you have
any idea why it would break?

[binji, 2014-04-15 16:41:55]

On 2014/04/15 16:39:39, David Yen wrote:
> On 2014/04/15 16:35:27, binji wrote:
> > On 2014/04/15 05:30:23, I haz the power (commit-bot) wrote:
> > > Change committed as 263791
> > 
> > This broke the windows bots, as you can see from the win_nacl_sdk and
> > win_nacl_sdk_build trybot runs. Unfortunately, the sdk builders still do not
> > block the CQ so you will have to be more careful.
> 
> Ah okay, I assumed it would block the CQ if something was my fault. Do you
have
> any idea why it would break?

Not really, it doesn't have any useful output. See
http://build.chromium.org/p/tryserver.chromium/builders/win_nacl_sdk_build/bu...

You'll probably have to try it on a Windows machine.