- sprintf, snprintf -> sprintf_s (ensures 0-termination of result)
- use %s instead of %hs with sprintf_s to avoid GCC warnings
- convert h_mgr's to_string to unicode
This was SVN commit r7170.
convert remaining %s to %hs or %ls
use WPRINTF_ARGS on unicode printf functions
fix: __func__ can't be widened via preprocessor (-> revert to char*)
convert remaining external_[directory|file_string to string()
Util: don't mix fwprintf and fprintf
Formation: fix dangling pointer (storing result of c_str())
This was SVN commit r7164.
also took the opportunity to move file_system_posix to file_system since
a further FS abstraction layer = YAGNI. also namespaced
file_system_util.
This was SVN commit r7074.
Use noncopyable instead of boost::noncopyable. (But maybe this should be
changed to the NONCOPYABLE macro instead?)
Use boost::filesystem::wpath::file_string instead of
external_file_string, since the latter varies between std::string on
Linux and std::wstring on Windows.
Use wcstombs instead of wcstombs_s.
Use rtl_AllocateAligned instead of _mm_malloc.
This was SVN commit r6574.
a) Run commands were not being sent accross properly because the JS just
directly set flags on the entity. Note: The current fix has the
side-effect of making it no longer possible to double-rightclick a
target to charge, run-gather, etc. That will be fixed in a later commit.
b) EntityLists were serialized in a format that made it impossible to
deserialize a zero-length list.
This was SVN commit r6221.
class static const member apparently sometimes requires external linkage
and enum causes conversion warnings, so go with file-scope constants.
This was SVN commit r6161.
snd_mgr: remove support for sound cones (will not be used); only set
VSrc rolloff factor if relative flag is set
main: hack: flip listener up vector to 'fix' L/R flip (cause is unknown,
coord systems appear to be correct)
This was SVN commit r6092.
also includes sound support.
GameSetup: ensure all ScriptingInit are called; moved simulation
ScriptingInits and init/shutdown to SimulationScriptInit.cpp (reduces
#includes)
This was SVN commit r6028.
this snowballed into a massive search+destroy of the hodgepodge of
mostly equivalent types we had in use (int, uint, unsigned, unsigned
int, i32, u32, ulong, uintN).
it is more efficient to use 64-bit types in 64-bit mode, so the
preferred default is size_t (for anything remotely resembling a size or
index). tile coordinates are ssize_t to allow more efficient conversion
to/from floating point. flags are int because we almost never need more
than 15 distinct bits, bit test/set is not slower and int is fastest to
type. finally, some data that is pretty much directly passed to OpenGL
is now typed accordingly.
after several hours, the code now requires fewer casts and less
guesswork.
other changes:
- unit and player IDs now have an "invalid id" constant in the
respective class to avoid casting and -1
- fix some endian/64-bit bugs in the map (un)packing. added a
convenience function to write/read a size_t.
- ia32: change CPUID interface to allow passing in ecx (required for
cache topology detection, which I need at work). remove some unneeded
functions from asm, replace with intrinsics where possible.
This was SVN commit r5942.
replace ONCE with direct calls to ScriptingInit in simulation/network
code as well.
note: GameSetup is now responsible for calling ServerSession's init
instead of Server doing it
added simulation/scripting with a module that takes care of all
scripting init (reduces CCD)
This was SVN commit r5883.
remove ONCE(ScriptingInit) constructs, replace with direct call from
GameSetup
revised GameSetup's InitScripting (organize into groups)
JSCollection: remove #define of two concrete collection types (made
typedef, moved to their respective entity.h and player.h)
This was SVN commit r5871.
in 2008-02-09 meeting, decided to stick with a fixed-size array for
simplicity. (changing to vector would require modifying the m_refd
bitset optimization. there is a 32k handle limit due to serialization
code anyway, and no pressing need to justify changing either)
This was SVN commit r5612.
The problem was incorrect logic in the start-next-segment-of-the-path
code that was added to allow multiple segments to be handled within the
same update. I'm not sure how this worked before but it seemed to give
continuous paths! Likely that's because the pathing for contact actions
is different.
Also added more sanity checks in Aura.cpp for deleted entities in the
m_Influenced queue. There was a crash caused by something here though
it's hard to reproduce.
This was SVN commit r5597.
