Add BoringSSL GYP files.

Add BoringSSL GYP files.

BUG=none

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=282169

[agl, 2014-07-08 23:44:29]

This only causes BoringSSL to be built on platforms where use_openssl=1 is
already defined.

[Ryan Sleevi, 2014-07-09 00:03:35]

maruel is my go-to guy for Python build script reviews. He should peep on
update_gypi_and_asm.py for Chromium Python style.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/bo...
File third_party/boringssl/boringssl.gyp (right):

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/bo...
third_party/boringssl/boringssl.gyp:12: 'type': '<(component)',
I don't think this can be built as a <(component) for the WIN case, because it
would lack the __declspec(export) and would lack a .def file that sets up the
exports.

For Mac/Linux, you're having to turn off -fvisibility=hidden, so I suspect this
is a real issue?

Building this as a static library seems correct?

Alternatively, it would mean no Windows compile, which may be intentional, but
not long-term.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/bo...
third_party/boringssl/boringssl.gyp:96: # vim: set expandtab tabstop=2
shiftwidth=2:
Our GYP files removed mode-lines. There are utilities in tools to setup
file-type rules, but we don't put them in the GYP bits

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/bo...
File third_party/boringssl/boringssl_unittest.cc (right):

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/bo...
third_party/boringssl/boringssl_unittest.cc:36: (output.size() == 5 ||
output[output.size() - 6] == '\n');
Historically, using output parsing has caused issues for situations when tests
crash, hence a preference to use return codes instead.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
File third_party/boringssl/update_gypi_and_asm.py (right):

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:192: crypto_c_files =
FindCFiles(os.path.join('src', 'crypto'), NoTests)
Chromium Python style is to do all this in a main() function

[M-A Ruel, 2014-07-09 00:24:22]

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
File third_party/boringssl/update_gypi_and_asm.py (right):

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:1: # This utility enumerates the
BoringSSL source in src/ and generates two gypi
Add chromium copyright

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:5: import os.path
not necessary

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:11: '''FindCMakeFiles recurses
through directory and returns a list of the paths
""" and verb for docstring, e.g.

"""Returns list of all CMakeLists.txt files recursively in directory."""

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:16: for dent in dents:
What you want is:
for root, dirs, files in os.walk(directory):
  if 'CMakeLists.txt' in files:
    cmakefiles.append(os.path.join(root, 'CMakeLists.txt'))

or something like that.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:53: for dent in dents:
os.walk

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:78: for line in f.readlines():
with open(cmakefile) as f:
  for line in f:
    ...

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:84: # Remove "perlasm(" from start
and ")" from end
Align comment with rest of script.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:86: params = line.split()
params = line[8:-1].split()

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:90: 'output':
os.path.join(os.path.dirname(cmakefile), params[0]),
sort keys

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:100: '''ReadPerlAsmOperations
returns a list of all perlasm() directives found in
"""Returns ...

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:112: kOSArchCombos = [
ALL_CAPS const at top of file.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:140: out_file =
file(output_filename, 'w+')
use with statement

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:171: elif (arch == 'arm' and
this seems to fit 80 cols

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:179: for (key, non_perl_asm_files)
in kNonPerlFiles.items():
s/items/iteritems/

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:192: crypto_c_files =
FindCFiles(os.path.join('src', 'crypto'), NoTests)
On 2014/07/09 00:03:35, Ryan Sleevi wrote:
> Chromium Python style is to do all this in a main() function

Yes,

def main():
  foo ...
  bar ...
  return 0


if __name__ == '__main__':
  sys.exit(main())

[agl, 2014-07-09 16:42:01]

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/bo...
File third_party/boringssl/boringssl.gyp (right):

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/bo...
third_party/boringssl/boringssl.gyp:12: 'type': '<(component)',
On 2014/07/09 00:03:35, Ryan Sleevi wrote:
> I don't think this can be built as a <(component) for the WIN case, because it
> would lack the __declspec(export) and would lack a .def file that sets up the
> exports.

I think you're correct. I took this from the existing openssl.gyp. Have switched
to a static_library and removed the -fvisibility=hidden hack. (I would like to
have proper export controls in BoringSSL, but we don't have it yet.)

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/bo...
third_party/boringssl/boringssl.gyp:96: # vim: set expandtab tabstop=2
shiftwidth=2:
On 2014/07/09 00:03:35, Ryan Sleevi wrote:
> Our GYP files removed mode-lines. There are utilities in tools to setup
> file-type rules, but we don't put them in the GYP bits

Done (again, taken from openssl.gyp).

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/bo...
File third_party/boringssl/boringssl_unittest.cc (right):

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/bo...
third_party/boringssl/boringssl_unittest.cc:36: (output.size() == 5 ||
output[output.size() - 6] == '\n');
On 2014/07/09 00:03:35, Ryan Sleevi wrote:
> Historically, using output parsing has caused issues for situations when tests
> crash, hence a preference to use return codes instead.

Changed to check both a 0 exit code and that the suffix of the output is
correct.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
File third_party/boringssl/update_gypi_and_asm.py (right):

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:1: # This utility enumerates the
BoringSSL source in src/ and generates two gypi
On 2014/07/09 00:24:21, M-A Ruel wrote:
> Add chromium copyright

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:5: import os.path
On 2014/07/09 00:24:21, M-A Ruel wrote:
> not necessary

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:11: '''FindCMakeFiles recurses
through directory and returns a list of the paths
On 2014/07/09 00:24:21, M-A Ruel wrote:
> """ and verb for docstring, e.g.
> 
> """Returns list of all CMakeLists.txt files recursively in directory."""

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:16: for dent in dents:
On 2014/07/09 00:24:21, M-A Ruel wrote:
> What you want is:
> for root, dirs, files in os.walk(directory):
>   if 'CMakeLists.txt' in files:
>     cmakefiles.append(os.path.join(root, 'CMakeLists.txt'))
> 
> or something like that.

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:53: for dent in dents:
On 2014/07/09 00:24:21, M-A Ruel wrote:
> os.walk

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:78: for line in f.readlines():
On 2014/07/09 00:24:21, M-A Ruel wrote:
> with open(cmakefile) as f:
>   for line in f:
>     ...

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:84: # Remove "perlasm(" from start
and ")" from end
On 2014/07/09 00:24:21, M-A Ruel wrote:
> Align comment with rest of script.

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:86: params = line.split()
On 2014/07/09 00:24:21, M-A Ruel wrote:
> params = line[8:-1].split()

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:90: 'output':
os.path.join(os.path.dirname(cmakefile), params[0]),
On 2014/07/09 00:24:21, M-A Ruel wrote:
> sort keys

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:100: '''ReadPerlAsmOperations
returns a list of all perlasm() directives found in
On 2014/07/09 00:24:21, M-A Ruel wrote:
> """Returns ...

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:112: kOSArchCombos = [
On 2014/07/09 00:24:21, M-A Ruel wrote:
> ALL_CAPS const at top of file.

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:140: out_file =
file(output_filename, 'w+')
On 2014/07/09 00:24:21, M-A Ruel wrote:
> use with statement

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:171: elif (arch == 'arm' and
On 2014/07/09 00:24:21, M-A Ruel wrote:
> this seems to fit 80 cols

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:179: for (key, non_perl_asm_files)
in kNonPerlFiles.items():
On 2014/07/09 00:24:21, M-A Ruel wrote:
> s/items/iteritems/

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:192: crypto_c_files =
FindCFiles(os.path.join('src', 'crypto'), NoTests)
On 2014/07/09 00:03:35, Ryan Sleevi wrote:
> Chromium Python style is to do all this in a main() function

Done.

https://codereview.chromium.org/377783004/diff/30001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:192: crypto_c_files =
FindCFiles(os.path.join('src', 'crypto'), NoTests)
On 2014/07/09 00:24:21, M-A Ruel wrote:
> On 2014/07/09 00:03:35, Ryan Sleevi wrote:
> > Chromium Python style is to do all this in a main() function
> 
> Yes,
> 
> def main():
>   foo ...
>   bar ...
>   return 0
> 
> 
> if __name__ == '__main__':
>   sys.exit(main())

Done.

[M-A Ruel, 2014-07-09 16:56:44]

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
File third_party/boringssl/update_gypi_and_asm.py (right):

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:5: # This utility enumerates the
BoringSSL source in src/ and generates two gypi
Convert to docstring, e.g.

"""Enumerates the BoringSSL ..."""

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:17: ('linux', 'arm', 'elf', 'S'),
You are free to align at +2 if you prefer.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:48: '''NoTests is a filter function
that can be passed to FindCFiles in order to
""" for docstrings, everywhere for consistency.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:109: cmakefiles =
FindCMakeFiles('src')
import itertools

...

return itertools.chain.from_iterable(
    ExtractPerlAsmFromCMakeFile(i) for i in FindCMakeFiles('src'))

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:122: except OSError:
I prefer for you to use:
base_dir = os.path.dirname(output_filename)
if not os.path.isdir(base_dir):
  os.makedirs(base_dir)

so that if there is an access control error, it won't be swallowed.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:149: if (arch == 'x86_64' and
if arch == 'x86_64' and ('x86_64' in filename or 'avx2' in filename):

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:152: elif (arch == 'x86' and 'x86'
in filename and 'x86_64' not in filename):
elif arch == 'x86' and 'x86' in filename and 'x86_64' not in filename:

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:154: elif (arch == 'arm' and 'arm'
in filename):
elif arch == 'arm' and 'arm' in filename:

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:159: asmfiles[key] =
asmfiles.get(key, []) + [output]
I prefer:
asmfiles.setdefault(key, []).append(output)

it's a tad shorter and easier to glance at for me.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:162: asmfiles[key] =
asmfiles.get(key, []) + non_perl_asm_files
asmfiles.setdefault(key, []).extend(non_perl_asm_files)

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:170: for f in files:
for f in sorted(files):

would prevent from modifying input parameters.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:179: gypi = file('boringssl.gypi',
'w+')
with open('boringssl.gypi', 'w+') as gypi:

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:181: header = '''# Copyright (c)
2014 The Chromium Authors. All rights reserved.
Could you move this out of the function? I find multiline strings significantly
break the reading flow (and confuse the hell out of my editor).

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:198: for ((osname, arch),
asm_files) in asmfiles.items():
for ((osname, arch), asm_files) in sorted(WriteAsmFiles(perlasms).iteritems()):

I recommend to sort to that the written file is deterministic. As a matter of
fact, I highly recommend to sort everything even if not strictly needed, so that
the output is deterministic. Otherwise I'll have to fix it up later myself due
to
http://www.chromium.org/developers/testing/isolated-testing/deterministic-builds
:)

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:208: test_gypi =
file('boringssl_tests.gypi', 'w+')
same

[agl, 2014-07-09 17:49:24]

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
File third_party/boringssl/update_gypi_and_asm.py (right):

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:5: # This utility enumerates the
BoringSSL source in src/ and generates two gypi
On 2014/07/09 16:56:43, M-A Ruel wrote:
> Convert to docstring, e.g.
> 
> """Enumerates the BoringSSL ..."""

Done.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:17: ('linux', 'arm', 'elf', 'S'),
On 2014/07/09 16:56:43, M-A Ruel wrote:
> You are free to align at +2 if you prefer.

Am just running pyformat over it and letting it decide.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:48: '''NoTests is a filter function
that can be passed to FindCFiles in order to
On 2014/07/09 16:56:44, M-A Ruel wrote:
> """ for docstrings, everywhere for consistency.

Done.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:109: cmakefiles =
FindCMakeFiles('src')
On 2014/07/09 16:56:43, M-A Ruel wrote:
> import itertools
> 
> ...
> 
> return itertools.chain.from_iterable(
>     ExtractPerlAsmFromCMakeFile(i) for i in FindCMakeFiles('src'))

I'm sure that this is far more efficient, but I think the existing code is
clearer and it's only dealing with lists of a couple of dozen entries.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:122: except OSError:
On 2014/07/09 16:56:43, M-A Ruel wrote:
> I prefer for you to use:
> base_dir = os.path.dirname(output_filename)
> if not os.path.isdir(base_dir):
>   os.makedirs(base_dir)
> 
> so that if there is an access control error, it won't be swallowed.

Done.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:149: if (arch == 'x86_64' and
On 2014/07/09 16:56:44, M-A Ruel wrote:
> if arch == 'x86_64' and ('x86_64' in filename or 'avx2' in filename):

Done.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:152: elif (arch == 'x86' and 'x86'
in filename and 'x86_64' not in filename):
On 2014/07/09 16:56:44, M-A Ruel wrote:
> elif arch == 'x86' and 'x86' in filename and 'x86_64' not in filename:

Done.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:154: elif (arch == 'arm' and 'arm'
in filename):
On 2014/07/09 16:56:44, M-A Ruel wrote:
> elif arch == 'arm' and 'arm' in filename:

Done.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:159: asmfiles[key] =
asmfiles.get(key, []) + [output]
On 2014/07/09 16:56:44, M-A Ruel wrote:
> I prefer:
> asmfiles.setdefault(key, []).append(output)
> 
> it's a tad shorter and easier to glance at for me.

Done.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:162: asmfiles[key] =
asmfiles.get(key, []) + non_perl_asm_files
On 2014/07/09 16:56:43, M-A Ruel wrote:
> asmfiles.setdefault(key, []).extend(non_perl_asm_files)

Done.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:170: for f in files:
On 2014/07/09 16:56:44, M-A Ruel wrote:
> for f in sorted(files):
> 
> would prevent from modifying input parameters.

Done.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:179: gypi = file('boringssl.gypi',
'w+')
On 2014/07/09 16:56:44, M-A Ruel wrote:
> with open('boringssl.gypi', 'w+') as gypi:

Done.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:181: header = '''# Copyright (c)
2014 The Chromium Authors. All rights reserved.
On 2014/07/09 16:56:44, M-A Ruel wrote:
> Could you move this out of the function? I find multiline strings
significantly
> break the reading flow (and confuse the hell out of my editor).

Done.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:198: for ((osname, arch),
asm_files) in asmfiles.items():
On 2014/07/09 16:56:44, M-A Ruel wrote:
> for ((osname, arch), asm_files) in
sorted(WriteAsmFiles(perlasms).iteritems()):

Done. Although note that this script isn't running at build time, it'll only be
rerun by hand when the set of sources needs to be updated.

https://codereview.chromium.org/377783004/diff/50001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:208: test_gypi =
file('boringssl_tests.gypi', 'w+')
On 2014/07/09 16:56:44, M-A Ruel wrote:
> same

Done.

[M-A Ruel, 2014-07-09 17:57:58]

python file lgtm with a few additional nits

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
File third_party/boringssl/update_gypi_and_asm.py (right):

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:33: FILE_HEADER = '''# Copyright
(c) 2014 The Chromium Authors. All rights reserved.
""" for multiline strings.

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:55: """NoTests is a filter function
that can be passed to FindCFiles in order to
Filter function that ..

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:63: """OnlyTests is a filter
function that can be passed to FindCFiles in order to
Filter function that ..

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:125: """PerlAsm runs the a perlasm
script and puts the output into
Runs ..

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:137: """WriteAsmFiles generates asm
files from perlasm directives for each
Generates ..

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:176: out.write('''    ],\n''')
out.write('    ],\n')

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:184: gypi.write(FILE_HEADER + '''{
gypi.write(FILE_HEADER + '{\n  \'variables\': {\n')

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:197: gypi.write('''  }\n}\n''')
gypi.write('  }\n}\n')

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:202: test_gypi.write(FILE_HEADER +
'''{
test_gypi.write(FILE_HEADER + '{  \'targets\': [\n')

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:208: test_gypi.write('''    {
"""

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:222: test_gypi.write('''  ],
"""

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:229: test_gypi.write('''    ],
test_gypi.write('    ],  }\n}\n')

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:233: 
2 lines

[agl, 2014-07-09 18:41:04]

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
File third_party/boringssl/update_gypi_and_asm.py (right):

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:33: FILE_HEADER = '''# Copyright
(c) 2014 The Chromium Authors. All rights reserved.
On 2014/07/09 17:57:58, M-A Ruel wrote:
> """ for multiline strings.

Done.

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:55: """NoTests is a filter function
that can be passed to FindCFiles in order to
On 2014/07/09 17:57:58, M-A Ruel wrote:
> Filter function that ..

Done.

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:63: """OnlyTests is a filter
function that can be passed to FindCFiles in order to
On 2014/07/09 17:57:58, M-A Ruel wrote:
> Filter function that ..

Done.

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:125: """PerlAsm runs the a perlasm
script and puts the output into
On 2014/07/09 17:57:58, M-A Ruel wrote:
> Runs ..

Done.

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:137: """WriteAsmFiles generates asm
files from perlasm directives for each
On 2014/07/09 17:57:58, M-A Ruel wrote:
> Generates ..

Done.

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:176: out.write('''    ],\n''')
On 2014/07/09 17:57:57, M-A Ruel wrote:
> out.write('    ],\n')

Done.

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:184: gypi.write(FILE_HEADER + '''{
On 2014/07/09 17:57:58, M-A Ruel wrote:
> gypi.write(FILE_HEADER + '{\n  \'variables\': {\n')

Done.

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:197: gypi.write('''  }\n}\n''')
On 2014/07/09 17:57:58, M-A Ruel wrote:
> gypi.write('  }\n}\n')

Done.

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:202: test_gypi.write(FILE_HEADER +
'''{
On 2014/07/09 17:57:58, M-A Ruel wrote:
> test_gypi.write(FILE_HEADER + '{  \'targets\': [\n')

Done.

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:208: test_gypi.write('''    {
On 2014/07/09 17:57:58, M-A Ruel wrote:
> """

Done.

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:222: test_gypi.write('''  ],
On 2014/07/09 17:57:58, M-A Ruel wrote:
> """

Done.

https://codereview.chromium.org/377783004/diff/60001/third_party/boringssl/up...
third_party/boringssl/update_gypi_and_asm.py:229: test_gypi.write('''    ],
On 2014/07/09 17:57:57, M-A Ruel wrote:
> test_gypi.write('    ],  }\n}\n')

Done.

[Ryan Sleevi, 2014-07-09 18:57:27]

GYP LGTM

Note that GN is coming along rather quickly (on Linux and Mac particularly).
That said, with the goal of adding these tests / targets to the build waterfall,
as long as you exclude the GN bots (which should be the default, no special
action), you should be fine.

Just FYI

[agl, 2014-07-09 19:55:48]

The CQ bit was checked by agl@chromium.org

[commit-bot: I haz the power, 2014-07-09 19:58:14]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/agl@chromium.org/377783004/80001

[commit-bot: I haz the power, 2014-07-09 23:20:27]

Change committed as 282169

[aurimas (slooooooooow), 2014-07-10 05:36:30]

On 2014/07/09 23:20:27, I haz the power (commit-bot) wrote:
> Change committed as 282169

This change broke Android MIPS compilation.
http://build.chromium.org/p/chromium.fyi/builders/Android%20MIPS%20Builder%20...

Please land a fix or revert this change.

[Ryan Sleevi, 2014-07-10 07:01:27]

As an FYI builder, a request to revert is not generally appropriate.

The FYI builders bring up the FYI bots until they are on the main
waterfall. I agree we should find a solution, but a revert is not typically
an appropriate request for such bots.
On Jul 9, 2014 10:36 PM, <aurimas@chromium.org> wrote:

> On 2014/07/09 23:20:27, I haz the power (commit-bot) wrote:
>
>> Change committed as 282169
>>
>
> This change broke Android MIPS compilation.
> http://build.chromium.org/p/chromium.fyi/builders/Android%
> 20MIPS%20Builder%20%28dbg%29?numbuilds=200
>
> Please land a fix or revert this change.
>
> https://codereview.chromium.org/377783004/
>

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[pasko, 2014-07-10 16:16:29]

On 2014/07/10 07:01:27, Ryan Sleevi wrote:
> As an FYI builder, a request to revert is not generally appropriate.
> 
> The FYI builders bring up the FYI bots until they are on the main
> waterfall. I agree we should find a solution, but a revert is not typically
> an appropriate request for such bots.

This also breaks the downstream mips builder :(

[aurimas (slooooooooow), 2014-07-10 17:02:48]

On 2014/07/10 16:16:29, pasko wrote:
> On 2014/07/10 07:01:27, Ryan Sleevi wrote:
> > As an FYI builder, a request to revert is not generally appropriate.
> > 
> > The FYI builders bring up the FYI bots until they are on the main
> > waterfall. I agree we should find a solution, but a revert is not typically
> > an appropriate request for such bots.
> 
> This also breaks the downstream mips builder :(

I filled a bug https://code.google.com/p/chromium/issues/detail?id=392745

[Ryan Sleevi, 2014-07-10 21:52:12]

To circle back on this:

After a long discussion with pasko@, a few action items
- pasko@/aurimas@ to work with chrome-infrastructure to get those bots from FYI
onto the main waterfall, running content_shell or appropriate
- Documentation for how to build for Android x86/MIPS is at
https://code.google.com/p/chromium/wiki/AndroidBuildInstructions - so if we need
to reproduce/investigate tests, using Chromium checkouts, those steps are what
we should use.