git://developer.intra2net.com / libasyncio / blame
| Commit | Line | Data |
|---|---|---|
| 4f6228c3 | 1 | ideas/todo libsimpleio |
| 21ffe77a GE |
2 | ------------------------ |
| 3 | ||
| 4 | name | |
| 5 | ----- | |
| 4f6228c3 RP |
6 | don't use namespace i2n but namespace according to new project name |
| 7 | ||
| 8 | ideas: | |
| 9 | libsimpleio | |
| 10 | libastiop (LIBrary for ASynchronous Timer, I/O and Process) | |
| 11 | libsimpleasio | |
| 12 | libasiocontrol | |
| 13 | libasiohandle | |
| 14 | libasioflow | |
| 15 | libasiomaster | |
| 16 | libiocontrol | |
| 17 | libasyncrap | |
| 18 | libasyncwrap | |
| 19 | libasyncio | |
| 21ffe77a GE |
20 | |
| 21 | ||
| 22 | smaller refactoring | |
| 23 | ----------- | |
| 24 | backend in separate file | |
| 25 | break backend::doOneStep in 4 functions | |
| 26 | document reason for ptrlist: keeps iterators valid during loops which do inserts/deletes | |
| 27 | separate ptrlist | |
| 28 | protect ptrlist against evil usage (e.g. forbid or overload erase, push_back, insert,...) | |
| 29 | separate MilliTime | |
| 30 | separate pollfd, maybe encapsulate the pollfd-struct instead of deriving from it | |
| 31 | separate polldatacluster | |
| 32 | document that timerbase is always based on monotonic clock, getrealwhentime converts the time live | |
| 33 | document handling of error-cases, m_errno (e.g. mask ENOTSOCK for non-socket fds) | |
| f0804030 | 34 | provide simple replacements for system() and pipestream using simpleio and timeouts |
| 21ffe77a GE |
35 | |
| 36 | bigger refactoring | |
| 37 | ------------------ | |
| 4f6228c3 | 38 | put iolist & timerlist into backend to reduce the usage global objects |
| 21ffe77a GE |
39 | make IOImplementations require a link to the backend they are used with |
| 40 | ||
| 41 | ideas | |
| 42 | ----- | |
| 43 | offer a common io-client or io-server, abstracting out the real communication channel used. | |
| 44 | makes it possible to switch between ways of communication at runtime | |
| f0804030 GE |
45 | |
| 46 | glue_t2n | |
| 47 | -------- | |
| 48 | don't call doOneStep from within the fill_buffer()-variants | |
| 49 | - for the server-variants not that hard if you require that command_server::handle will be called on every socket-event | |
| 50 | - for the client: needs callback-interface on t2n-side, need to think about it |