Hi, working with async stuff decided to look at what async way of getting file
size and compare it to sync way.
Got no difference in principle, but there is 2 different implementations:
The "sync" version in io.nim:
proc getFileSize*(f: File): int64 {.tags: [ReadIOEffect], benign.} =
## retrieves the file size (in bytes) of `f`.
let oldPos = getFilePos(f)
discard c_fseek(f, 0, 2) # seek the end of the file
result = getFilePos(f)
setFilePos(f, oldPos)
Run
The "async" version in asyncfile.nim:
proc getFileSize*(f: AsyncFile): int64 =
## Retrieves the specified file's size.
when defined(windows) or defined(nimdoc):
var high: DWORD
let low = getFileSize(f.fd.Handle, addr high)
if low == INVALID_FILE_SIZE:
raiseOSError(osLastError())
result = (high shl 32) or low
else:
let curPos = lseek(f.fd.cint, 0, SEEK_CUR)
result = lseek(f.fd.cint, 0, SEEK_END)
f.offset = lseek(f.fd.cint, curPos, SEEK_SET)
assert(f.offset == curPos)
Run
There also, the 3rd variant exists, but it is a bit different variant, where
the argument not file, but string path. But in that variant, we may find
another way of getting info for windows platform.
So, the question is: is this organization, intended? Or it is a
legacy/historical moment? It is too early for me to judge but willing to better
understand std lib design ideas and organization.
