⭐ 欢迎来到虫虫下载站! | 📦 资源下载 📁 资源专辑 ℹ️ 关于我们
⭐ 虫虫下载站

📄 todo

📁 编译工具
💻
字号:
Combined list of things to-do and things to think about-------------------------------------------------------* It's possible to contrive IDL examples which clash with internal names:    exception e{      long pd_insertToAnyFn;     // names inherited from CORBA::Exception      long pd_insertToAnyFnNCP;  // and not private      long insertExceptionToAny;    };  A suitable fix is to _ prefix the internal names, but the changes will  ripple outwards from the IDL compiler through include/omniORB3/CORBA.h  Of course it is still possible to contrive IDL examples which don't map  properly onto C++ at all:    interface T {};    interface T_var {};* Slight assymmetry in call desc unmarshalling code (positions where temp  variables are declared seems to vary)* Unify string marshalling/ unmarshalling- why is it sometimes done through  a String_member and sometimes done at the char level? -- Apparently this   was a changeover which was never 100% completed.* Surely a return value is semantically identical to an out argument value?  Why does it become a special case in the marshalling routines?* Perhaps we could simplify handling of anonymous sequences etc by introducing  (internal) typedefs?* Would it be possible/ easier to remove need to consider declarators when  looking at an object's type? (Maybe this is a just personal dislike of   C/ C++ / IDL declarator syntax) Shouldn't a type be a simple structure    t = Sequence of (Array of (Array of (Alias of (Struct foo))))  and be independent of it's actual _name(s)_. Perhaps make functions    marshal : fn Type -> Function block  and store a hashtable of functions available to promote reuse? Or could  expand each inline for speed?  Perhaps create a type class with appropriate constructors (taking types and  optional declarators) which contains the marshalling, unmarshalling and  alignment code?

⌨️ 快捷键说明

复制代码 Ctrl + C
搜索代码 Ctrl + F
全屏模式 F11
切换主题 Ctrl + Shift + D
显示快捷键 ?
增大字号 Ctrl + =
减小字号 Ctrl + -