Thread overview
[Issue 5002] New: possible std.stream improvements
Oct 06, 2010
Daniel Gibson
Oct 07, 2010
nfxjfg@gmail.com
Oct 11, 2010
Daniel Gibson
October 06, 2010
http://d.puremagic.com/issues/show_bug.cgi?id=5002

           Summary: possible std.stream improvements
           Product: D
           Version: D1 & D2
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Phobos
        AssignedTo: nobody@puremagic.com
        ReportedBy: metalcaedes@gmail.com


--- Comment #0 from Daniel Gibson <metalcaedes@gmail.com> 2010-10-06 09:21:40 PDT ---
I've got a few suggestions to make std.stream more convenient (and in one case just the documentation more clear). See my NG post [1] for details, I'll just summarize it here:

* The documentation claims write( <basic type> ) is implementation specific - this is not true for most types - it's actually platform specific and with the EndianStream most read/write operations can safely be used between different platforms. Especially for the SocketStream this is good to know.

* InputStream's read( <type> val ) sometimes is inconvenient, please add
something like
T read(T)() { T ret; readExact(&ret, ret.sizeof); return ret; }
So one can use "myFun( s.read!float );" instead of "float x; s.read(x);
myFun(x);"

* Minor inconsistencies: add readExact(ubyte[] buf) to Inputstream, because
there is a read(ubyte[] buf) and add writeExact(ubyte[] buf) to OutputStream
because there is write(ubyte[] buf).

* For convenience ubyte[] readExact(size_t len) might be nice.

Thanks,
- Daniel

[1] http://www.digitalmars.com/webnews/newsgroups.php?art_group=digitalmars.D&article_id=118413 - my NG post on these matters (and also on bug 5001)

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
October 07, 2010
http://d.puremagic.com/issues/show_bug.cgi?id=5002


nfxjfg@gmail.com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |nfxjfg@gmail.com


--- Comment #1 from nfxjfg@gmail.com 2010-10-06 20:54:14 PDT ---
I thought std.stream is deprecated.
Maybe "someone" could confirm and close this as invalid.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
October 11, 2010
http://d.puremagic.com/issues/show_bug.cgi?id=5002



--- Comment #2 from Daniel Gibson <metalcaedes@gmail.com> 2010-10-11 13:20:40 PDT ---
It will probably be deprecated, but at this time it isn't (there is no
alternative yet, anyway).

However I just realized that my idea for read() can't be implemented *that*
easy because it wouldn't work properly with EndianStream.
For that to work one would have to check whether T is a simple type (int,
float, double, long, etc) and in that case call fixBO() in EndianStream (like
the available, inconvenient read(<basic type>) methods in EndianStream already
do).

Or do it old school and provide readInt(), readUInt(), readLong(), readUlong(),
readFloat() etc instead of using a template.

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