Fix size regression due to missing strip.

Fix size regression due to missing strip.

Add a strip script based on the old build_nacl_irt.py (just keep strip piece)
Rename nacl_irt target to nacl_irt_unstripped and add a new target called
nacl_irt which strips the previous target.
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=109764

[Ryan Sleevi, 2011-11-12 02:38:14]

Quick drive by

http://codereview.chromium.org/8539039/diff/1/ppapi/native_client/native_clie...
File ppapi/native_client/native_client.gyp (right):

http://codereview.chromium.org/8539039/diff/1/ppapi/native_client/native_clie...
ppapi/native_client/native_client.gyp:101:
'<(SHARED_INTERMEDIATE_DIR)/tc_newlib/nacl_irt_x86_32.nexe',
Should you also add strip_nacl_irt.py as an 'input'?

I believe it needs to be the second dependency, not the first, due to how gyp
coalesces inputs for action de-duplication.

Same for line 125

[noelallen1, 2011-11-12 02:46:11]

http://codereview.chromium.org/8539039/diff/1/ppapi/native_client/native_clie...
File ppapi/native_client/native_client.gyp (right):

http://codereview.chromium.org/8539039/diff/1/ppapi/native_client/native_clie...
ppapi/native_client/native_client.gyp:101:
'<(SHARED_INTERMEDIATE_DIR)/tc_newlib/nacl_irt_x86_32.nexe',
On 2011/11/12 02:38:14, Ryan Sleevi wrote:
> Should you also add strip_nacl_irt.py as an 'input'?
> 
> I believe it needs to be the second dependency, not the first, due to how gyp
> coalesces inputs for action de-duplication.
> 
> Same for line 125

Done.

[Ryan Sleevi, 2011-11-12 02:52:51]

LGTM

http://codereview.chromium.org/8539039/diff/3001/chrome/strip_nacl_irt.py
File chrome/strip_nacl_irt.py (right):

http://codereview.chromium.org/8539039/diff/3001/chrome/strip_nacl_irt.py#new...
chrome/strip_nacl_irt.py:1: #!/usr/bin/python
In my msys checkout, this has historically caused problems.

#!/usr/bin/env python

Although from gyp, at least, you're invoking directly with python_exe, so this
is not critical.

http://codereview.chromium.org/8539039/diff/3001/chrome/strip_nacl_irt.py#new...
chrome/strip_nacl_irt.py:27: uplatform = platform.replace('-', '_')
This variable seems is never used.

Did you mean to use uplatform as the default for missing platforms on line 34?

http://codereview.chromium.org/8539039/diff/3001/chrome/strip_nacl_irt.py#new...
chrome/strip_nacl_irt.py:39: }.get(sys.platform, 'linux')
I suspect it's not an issue, but any concern for the freebsd builds? TODO patch
later? Or is 'linux' ok here?

[noelallen1, 2011-11-12 03:00:53]

http://codereview.chromium.org/8539039/diff/3001/chrome/strip_nacl_irt.py
File chrome/strip_nacl_irt.py (right):

http://codereview.chromium.org/8539039/diff/3001/chrome/strip_nacl_irt.py#new...
chrome/strip_nacl_irt.py:1: #!/usr/bin/python
On 2011/11/12 02:52:51, Ryan Sleevi wrote:
> In my msys checkout, this has historically caused problems.
> 
> #!/usr/bin/env python
> 
> Although from gyp, at least, you're invoking directly with python_exe, so this
> is not critical.

We allow cygwin for Windows in NaCl land.  But yeah we always run python just in
case.

This is sta

http://codereview.chromium.org/8539039/diff/3001/chrome/strip_nacl_irt.py#new...
chrome/strip_nacl_irt.py:27: uplatform = platform.replace('-', '_')
On 2011/11/12 02:52:51, Ryan Sleevi wrote:
> This variable seems is never used.
> 
> Did you mean to use uplatform as the default for missing platforms on line 34?

Lol, looks like a bug in the original I pulled this from.
Done.

http://codereview.chromium.org/8539039/diff/3001/chrome/strip_nacl_irt.py#new...
chrome/strip_nacl_irt.py:39: }.get(sys.platform, 'linux')
On 2011/11/12 02:52:51, Ryan Sleevi wrote:
> I suspect it's not an issue, but any concern for the freebsd builds? TODO
patch
> later? Or is 'linux' ok here?

We don't support it yet so...

[Ryan Sleevi, 2011-11-12 03:07:14]

http://codereview.chromium.org/8539039/diff/3001/chrome/strip_nacl_irt.py
File chrome/strip_nacl_irt.py (right):

http://codereview.chromium.org/8539039/diff/3001/chrome/strip_nacl_irt.py#new...
chrome/strip_nacl_irt.py:1: #!/usr/bin/python
On 2011/11/12 03:00:54, noelallen1 wrote:
> On 2011/11/12 02:52:51, Ryan Sleevi wrote:
> > In my msys checkout, this has historically caused problems.
> > 
> > #!/usr/bin/env python
> > 
> > Although from gyp, at least, you're invoking directly with python_exe, so
this
> > is not critical.
> 
> We allow cygwin for Windows in NaCl land.  But yeah we always run python just
in
> case.
> 
> This is sta

Right, I know the Google Open-Source Python guide says to use #!/usr/bin/python
(or #!/usr/bin/python{version}), but within the Chromium tree, scripts are more
evenly split between #!/usr/bin/python and #!/usr/bin/env python

The latter is so that the hermetic python (from depot_tools) can be located,
rather than the one bundled (which, in the case of msys, there is none)