Use IsAlNum instead (gtest-port.h), which deals with char signedness correctly. This was the only spot in gtest where a cctype function was called instead of its gtest-port.h equivalent.
Addresses https://github.com/assimp/assimp/issues/3867 and then some.
The OpenGEX importer defined a few global std::string constants, only to convert them back to C strings on use. This commit defines them as C strings from the beginning.
strncmp() was used to compare these strings to other strings, but the length limit was set to string length, which made it equivalent to strcmp(), just slower. Fixed that as well.
The DXF importer defined a global std::string constant, only to convert it back to a C string on use.
This commit defines the constant as a C string right away, thus saving 340 B of code and data.
"For a general character pointer that may also point to binary data,
POINTER(c_char) must be used." c_char_p is for a zero-terminated string.
Reference: https://docs.python.org/3/library/ctypes.html#ctypes.c_char_p
Applying this change to the 4.1.4 released python module fixes#2339 for
me in Ubuntu.
The SIB importer, upon needing an empty aiString, did not create a new one but rather copied a predefined global empty string.
Since aiStrings contain large buffers, Assimp copied 1028 B of zeros instead of setting five bytes (at least when compiled with Visual C++). Since aiString is a user-defined type without a constexpr constructor, Visual C++ had to generate a thread-safe run-time initializer as well.
Now it’s just two instructions.
Maxon’s Melange SDK has been renamed Cineware SDK as of 21.004, and with it all namespaces and types. This commit
- makes CMake use contrib/Cineware instead of contrib/Melange;
- renames Assimp’s namespace melange to namespace cineware;
- removes useless functions and formatter references from class C4DImporter;
- removes duplicate conversion of cineware::String to aiString in the importer;
- updates comments accordingly;
- updates copyright info.
the purpose is use of this loader with idtech3 FS
this requires full original material name, which euqals to Q3 shader path
result of deduction is not usable inside Q3 FS at all
option in general is "do not tinker with the path"
Jason C
Jason C
Kim Kulling
Krishty
Eric Olson
vfxgordon
kimkulling
kkulling
Tom spot Callaway
Garux
contriteobserver