[Issue 4025] Making network with the std.stdio.File interface

2018-05-17 Thread d-bugmail--- via Digitalmars-d-bugs
https://issues.dlang.org/show_bug.cgi?id=4025

Dmitry Olshansky  changed:

   What|Removed |Added

 Status|ASSIGNED|RESOLVED
 CC||dmitry.o...@gmail.com
 Resolution|--- |WONTFIX

--- Comment #2 from Dmitry Olshansky  ---
Hell no. Network can be done in a multitude of ways and putting the awkward
LibC buffering on top of it is the last thing we'd want. 

Especially in std. There is now quite a few better things on Dub that does
network and I/O.

All that aside - in a Dub package this addition might be simple thing that will
find its users.

--


[Issue 4025] Making network with the std.stdio.File interface

2011-08-23 Thread d-bugmail
http://d.puremagic.com/issues/show_bug.cgi?id=4025


Vladimir Panteleev thecybersha...@gmail.com changed:

   What|Removed |Added

 CC||thecybersha...@gmail.com


--- Comment #1 from Vladimir Panteleev thecybersha...@gmail.com 2011-08-23 
19:31:06 PDT ---
Some thoughts:

* The concept is POSIX-only. Windows uses different APIs and handles for files
and network connections.
* I'm not sure why you used the C socket functions when you could have used
std.socket for the setup and obtained the handle. This could have shortened the
code considerably. (One drawback against using std.socket I can see is that its
destructor closes the socket, so you must maintain a reference to the Socket
object.)
* Perhaps std.stdio.File should just have public constructors which accept a
FILE* or (on POSIX) a file descriptor. This would eliminate the need for such
dedicated code, and instead provide better general support (for pipes and
whatever else file descriptors are used for).
* Alternatively, there's always std.socketstream...

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
--- You are receiving this mail because: ---