Make V8 compile with MSVS 2015

BUG=v8:4326
R=machenbach@chromium.org
LOG=y

Review URL: https://codereview.chromium.org/1259693003

Cr-Commit-Position: refs/heads/master@{#29868}
This commit is contained in:
jochen 2015-07-27 04:51:16 -07:00 committed by Commit bot
parent 4ad22295d4
commit 2714ee5539

View File

@ -671,7 +671,85 @@
}],
],
'msvs_cygwin_dirs': ['<(DEPTH)/third_party/cygwin'],
'msvs_disabled_warnings': [4355, 4800],
'msvs_disabled_warnings': [
# C4091: 'typedef ': ignored on left of 'X' when no variable is
# declared.
# This happens in a number of Windows headers. Dumb.
4091,
# C4127: conditional expression is constant
# This warning can in theory catch dead code and other problems, but
# triggers in far too many desirable cases where the conditional
# expression is either set by macros or corresponds some legitimate
# compile-time constant expression (due to constant template args,
# conditionals comparing the sizes of different types, etc.). Some of
# these can be worked around, but it's not worth it.
4127,
# C4351: new behavior: elements of array 'array' will be default
# initialized
# This is a silly "warning" that basically just alerts you that the
# compiler is going to actually follow the language spec like it's
# supposed to, instead of not following it like old buggy versions
# did. There's absolutely no reason to turn this on.
4351,
# C4355: 'this': used in base member initializer list
# It's commonly useful to pass |this| to objects in a class'
# initializer list. While this warning can catch real bugs, most of
# the time the constructors in question don't attempt to call methods
# on the passed-in pointer (until later), and annotating every legit
# usage of this is simply more hassle than the warning is worth.
4355,
# C4503: 'identifier': decorated name length exceeded, name was
# truncated
# This only means that some long error messages might have truncated
# identifiers in the presence of lots of templates. It has no effect
# on program correctness and there's no real reason to waste time
# trying to prevent it.
4503,
# Warning C4589 says: "Constructor of abstract class ignores
# initializer for virtual base class." Disable this warning because it
# is flaky in VS 2015 RTM. It triggers on compiler generated
# copy-constructors in some cases.
4589,
# C4611: interaction between 'function' and C++ object destruction is
# non-portable
# This warning is unavoidable when using e.g. setjmp/longjmp. MSDN
# suggests using exceptions instead of setjmp/longjmp for C++, but
# Chromium code compiles without exception support. We therefore have
# to use setjmp/longjmp for e.g. JPEG decode error handling, which
# means we have to turn off this warning (and be careful about how
# object destruction happens in such cases).
4611,
# TODO(jochen): These warnings are level 4. They will be slowly
# removed as code is fixed.
4100, # Unreferenced formal parameter
4121, # Alignment of a member was sensitive to packing
4244, # Conversion from 'type1' to 'type2', possible loss of data
4302, # Truncation from 'type 1' to 'type 2'
4309, # Truncation of constant value
4311, # Pointer truncation from 'type' to 'type'
4312, # Conversion from 'type1' to 'type2' of greater size
4481, # Nonstandard extension used: override specifier 'keyword'
4505, # Unreferenced local function has been removed
4510, # Default constructor could not be generated
4512, # Assignment operator could not be generated
4610, # Object can never be instantiated
4800, # Forcing value to bool.
4838, # Narrowing conversion. Doesn't seem to be very useful.
4995, # 'X': name was marked as #pragma deprecated
4996, # 'X': was declared deprecated (for GetVersionEx).
# These are variable shadowing warnings that are new in VS2015. We
# should work through these at some point -- they may be removed from
# the RTM release in the /W4 set.
4456, 4457, 4458, 4459,
],
'msvs_settings': {
'VCCLCompilerTool': {
'MinimalRebuild': 'false',