Makes analyzer always output static_libraries that have changed

Makes analyzer always output static_libraries that have changed

This is to deal with static_libraries not being compile dependencies.
In particular if we only output the topmost ancestor that depends upon
the static library then the static library may not get built if the
ancestor is a static_library itself.

BUG=gyp:493
TEST=covered by test
R=thakis@chromium.org

Committed: https://chromium.googlesource.com/external/gyp/+/fdc7b812f99e48c00e9a487bd56751bbeae07043

[sky, 2015-06-08 17:31:38]

sky@chromium.org changed reviewers:
+ thakis@chromium.org

[Nico, 2015-06-08 17:37:56]

https://codereview.chromium.org/1142323006/diff/20001/pylib/gyp/generator/ana...
File pylib/gyp/generator/analyzer.py (right):

https://codereview.chromium.org/1142323006/diff/20001/pylib/gyp/generator/ana...
pylib/gyp/generator/analyzer.py:436: not target.is_or_has_executable_ancestor)):
you care about "is_or_has_linked_ancestor", right? (in particular, dlls /
shared_library targets are linked but not executable)

[sky, 2015-06-08 17:46:26]

https://codereview.chromium.org/1142323006/diff/20001/pylib/gyp/generator/ana...
File pylib/gyp/generator/analyzer.py (right):

https://codereview.chromium.org/1142323006/diff/20001/pylib/gyp/generator/ana...
pylib/gyp/generator/analyzer.py:436: not target.is_or_has_executable_ancestor)):
On 2015/06/08 17:37:56, Nico wrote:
> you care about "is_or_has_linked_ancestor", right? (in particular, dlls /
> shared_library targets are linked but not executable)

Done.

[Nico, 2015-06-08 17:53:31]

lgtm

https://codereview.chromium.org/1142323006/diff/40001/pylib/gyp/generator/ana...
File pylib/gyp/generator/analyzer.py (right):

https://codereview.chromium.org/1142323006/diff/40001/pylib/gyp/generator/ana...
pylib/gyp/generator/analyzer.py:430: # executables. This is necessary as the
other dependencies on them may be
s/executables/linkables/ in comment too

https://codereview.chromium.org/1142323006/diff/40001/test/analyzer/gyptest-a...
File test/analyzer/gyptest-analyzer.py (right):

https://codereview.chromium.org/1142323006/diff/40001/test/analyzer/gyptest-a...
test/analyzer/gyptest-analyzer.py:154: #test.pass_test()
is this intentionally commented out? probably not?

[sky, 2015-06-08 18:01:17]

https://codereview.chromium.org/1142323006/diff/40001/pylib/gyp/generator/ana...
File pylib/gyp/generator/analyzer.py (right):

https://codereview.chromium.org/1142323006/diff/40001/pylib/gyp/generator/ana...
pylib/gyp/generator/analyzer.py:430: # executables. This is necessary as the
other dependencies on them may be
On 2015/06/08 17:53:31, Nico wrote:
> s/executables/linkables/ in comment too

Done.

https://codereview.chromium.org/1142323006/diff/40001/test/analyzer/gyptest-a...
File test/analyzer/gyptest-analyzer.py (right):

https://codereview.chromium.org/1142323006/diff/40001/test/analyzer/gyptest-a...
test/analyzer/gyptest-analyzer.py:154: #test.pass_test()
On 2015/06/08 17:53:31, Nico wrote:
> is this intentionally commented out? probably not?

Ya, sorry, I removed this line.

test.pass_test() is effectively an early out. I do that some times when I only
care about this test and not any other future failures. I forgot to remove it.

[sky, 2015-06-08 18:25:35]

Committed patchset #5 (id:80001) manually as
fdc7b812f99e48c00e9a487bd56751bbeae07043 (presubmit successful).