Allow for certain values to always be properly initialized on construction -- namely the maps for now. I'm currently working on a change that requires a baseline constructor; this change would make the use of `chansync.BroadcastCond` and `chansync.SetOnce` obsolete -- i.e. one can have channel members without worrying about proper initialization/destruction of a `chan struct{}`.
As for why `makeClient` returns a value instead of a pointer: returning a value gives us more options -- you can always take a pointer from a value later on cheaply, and have things moved to the heap if they weren't already. The same can't be said about getting a value back from a pointer. GC also could potentially have less work to do. Plus I personally find ownership to be an important concept (semi-borrowed from rust) -- use of values make ownership clear.
* Drop bradfitz/iter dependency
`range iter.N` looks nice and doesn't allocate, but unfortunately using a `range` expression blocks a function from being inlined wherever it's used (for now). It's not that we need inlining in all cases, but I do think a C-style for loop looks just as nice and is probably clearer to the majority. There also aren't any clear disadvantages to changing (unless you just happen to dislike the look of C)
* Update misc_test.go
* Update rlreader_test.go
* Update torrent_test.go
* Update bench_test.go
* Update client_test.go
* Update iplist_test.go
* Update mse_test.go
* Update peerconn_test.go
* Update peerconn.go
* Update order_test.go
* Update decoder_test.go
* Update main.go
* Update bench-piece-mark-complete.go
* Update main.go
* Update torrent.go
* Update iplist_test.go
* Update main.go
* Inlineable reader.Seek and no-lock return for bad whence
Couldn't find an elegant way to early exit on bad whence. Using a closure could reduce code duplication slightly, but it's overkill for only 3 cases.
Note that returning 0 on an invalid whence is the behavior of `strings.Reader` and `bytes.Reader`.
* Update reader.go
Reorder some of the fields to reduce the size of `reader`. Not much of a change -- original size was 104 bytes, with changes (moving bool's to the end) the size is 96 bytes. There's also the whole cache-friendly bit -- I tried reordering fields to help here, but I doubt it makes much of a difference.
Some other notes: `sync.Locker` is a 16 byte field. I suggest changing this to something like `*sync.RWMutex` or `*lockWithDeferreds`. This would not only reduce size but would allow access to `RLock`/`RUnlock`.
`pieceRange` could also be shrunk to 8 bytes, but I feel like I'm probably asking for too much lol.
