Wrap download_from_google_storage.py in a cwd-independent script

Wrap download_from_google_storage.py in a cwd-independent script

gclient runhooks runs all of the hooks with a working directory set to
the location of the .gclient file (i.e. parent of src/) and
download_from_google_storage.py relies on this to find the sha1. This
wraps download_from_google_storage.py in a script that works from anywhere.

R=viettrungluu@chromium.org

Committed: https://chromium.googlesource.com/external/mojo/+/d36bbce7d68a32722180280bcc553648bb983299

[jamesr, 2015-10-22 23:42:02]

jamesr@chromium.org changed reviewers:
+ viettrungluu@chromium.org

[jamesr, 2015-10-22 23:42:02]

I really hate this, but don't have a better idea.

download_from_google_storage is a sha1 comparison + wget + a bunch of logic we
never use so I'd like to rip it out eventually

[jamesr, 2015-10-28 20:50:48]

Description was changed from

==========
Add script to set cwd explicitly when running gclient hooks

gclient runhooks runs all of the hooks with a working directory set to
the location of the .gclient file (i.e. parent of src/) and several of
the hooks, most nostably download_from_google_storage, rely on this to
resolve relative paths. jiri hooks don't set the cwd (and shouldn't, imo)
so this adds a wrapper script for our hooks that sets the dir manually
and wraps the download scripts.
==========

to

==========
Wrap download_from_google_storage.py in a cwd-independent script

gclient runhooks runs all of the hooks with a working directory set to
the location of the .gclient file (i.e. parent of src/) and
download_from_google_storage.py relies on this to find the sha1. This
wraps download_from_google_storage.py in a script that works from anywhere.
==========

[jamesr, 2015-10-28 20:51:33]

This seems better.  We can over time change this from being a wrapper to just
doing wget + validating the sha1 since that's the only form of binary we grab.

[viettrungluu, 2015-10-28 22:37:33]

https://codereview.chromium.org/1410783004/diff/20001/DEPS
File DEPS (right):

https://codereview.chromium.org/1410783004/diff/20001/DEPS#newcode214
DEPS:214: '--no_resume',
I wonder if we shouldn't include "--no_resume", "--quiet", "--no_auth"
automatically in our wrapper.

[jamesr, 2015-10-28 22:42:05]

On 2015/10/28 at 22:37:33, viettrungluu wrote:
> https://codereview.chromium.org/1410783004/diff/20001/DEPS
> File DEPS (right):
> 
> https://codereview.chromium.org/1410783004/diff/20001/DEPS#newcode214
> DEPS:214: '--no_resume',
> I wonder if we shouldn't include "--no_resume", "--quiet", "--no_auth"
automatically in our wrapper.

Maybe, but one step at a time IMO

[viettrungluu, 2015-10-28 23:15:48]

On 2015/10/28 22:42:05, jamesr wrote:
> On 2015/10/28 at 22:37:33, viettrungluu wrote:
> > https://codereview.chromium.org/1410783004/diff/20001/DEPS
> > File DEPS (right):
> > 
> > https://codereview.chromium.org/1410783004/diff/20001/DEPS#newcode214
> > DEPS:214: '--no_resume',
> > I wonder if we shouldn't include "--no_resume", "--quiet", "--no_auth"
> automatically in our wrapper.
> 
> Maybe, but one step at a time IMO

OK, LGTM.

[jamesr, 2015-10-28 23:23:10]

Committed patchset #2 (id:20001) manually as
d36bbce7d68a32722180280bcc553648bb983299 (presubmit successful).