Following functions are implemented for further performance
optimization.
* JoinStringPiece
* SplitStringPiece
* EqualsCaseInsensitiveASCII
* ToLowerASCII
To improve performance of CLParser, I will introduce above functions
into include_normalize-win32.cc.
Previously the browse server would bind to "", which is translated to
0.0.0.0 (all interfaces), and then the hostname as retrieved by
socket.gethostname() was presented to the user.
The hostname is now "localhost" by default and is configurable, so the
original behavior is achieved with `ninja -t browse -a ""`.
As a fix for #874, we started reloading the entire manifest even if the
manifest was never rebuilt due to a restat rule. But this can be slow,
so call State::Reset instead, which also fixes the original crash.
Fixes#987
.\src\clparser.cc(56): note: see reference to function template instantiation '_OutIt std::transform<std::_String_iterator<std::_String_val<std::_Simple_types<char>>>,std::_String_iterator<std::_String_val<std::_Simple_types<char>>>,int(__cdecl *)(int)>(_InIt,_InIt,_OutIt,_Fn1)' being compiled
with
[
_OutIt=std::_String_iterator<std::_String_val<std::_Simple_types<char>>>,
_InIt=std::_String_iterator<std::_String_val<std::_Simple_types<char>>>,
_Fn1=int (__cdecl *)(int)
]
D:\bin\dev\VS\2017\BuildTools\VC\Tools\MSVC\14.10.24911\include\algorithm(946): warning C4244: '=': conversion from 'int' to 'char', possible loss of data
Non-console subprocesses have the write end of a pipe connected to fds
1 and 2 for stdout and stderr, but they also have the it connected to
whatever fd was assigned in the ninja process when the pipe was
created. Add a call to posix_spawn_file_actions_addclose after
the posix_spawn_file_actions_adddup2 calls to close the original fd
once it has been dup'd to stdout and stderr.
This fixes an issue seen in the Android build, where one of the
subprocesses is used to start a background helper process. The
background process attempts to close any inherited fds, but if ninja
used a very large fd number due to a high parallelism count the
background process would not close the fd and ninja would never
consider the subprocess finished.
The recommended replacement VerifyVersionInfo should work with the same
SDKs that GetVersionEx worked with (while the wrappers in
VersionHelpers.h require a recent SDK). This patch should not change
behavior, and it's not supposed to increase build requirements. If
this makes things harder to build, please let me know.
