http://d.puremagic.com/issues/show_bug.cgi?id=9501

           Summary: Struct std.mmfile.MmFile?
           Product: D
           Version: D2
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: Phobos
        AssignedTo: nobody@puremagic.com
        ReportedBy: bearophile_hugs@eml.cc


--- Comment #0 from bearophile_hugs@eml.cc 2013-02-12 17:45:01 PST ---
Maybe it's a good idea for std.mmfile.MmFile to become a struct like std.stdio.File, so the file is closed when this struct goes out of scope and the reference count it keeps is zero.

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

http://d.puremagic.com/issues/show_bug.cgi?id=9501


Jonathan M Davis <jmdavisProg@gmx.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jmdavisProg@gmx.com


--- Comment #1 from Jonathan M Davis <jmdavisProg@gmx.com> 2013-02-12 17:54:43 PST ---
Yes, std.mmfile.MmFile should be a struct. It doesn't use polymorphism and has no need to be a class. But making it a struct would break all code that uses it, so we're arguably stuck with it being the way that it is.

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

http://d.puremagic.com/issues/show_bug.cgi?id=9501



--- Comment #2 from bearophile_hugs@eml.cc 2013-02-12 18:19:08 PST ---
(In reply to comment #1)
> Yes, std.mmfile.MmFile should be a struct. It doesn't use polymorphism and has no need to be a class.

And maybe it doesn't have an explicit close() method, I don't know why, maybe for safety in accessing the memory mapped bytes.


> But making it a struct would break all code that uses
> it, so we're arguably stuck with it being the way that it is.

There are always acceptable deprecation paths, if we want them. Like adding the struct to std.mmfile with another name ("RcMmFile"?) and deprecating std.mmfile.MmFile for two years.

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

http://d.puremagic.com/issues/show_bug.cgi?id=9501



--- Comment #3 from bearophile_hugs@eml.cc 2013-02-12 20:05:43 PST ---
Mostly unrelated, about std.mmfile: I think in the std.mmfile module this line (line 270) should offer a string message too when the file is not found:

errnoEnforce(false);


And I think switches like this one should become "final":

            switch (mode)
            {
            case Mode.read:
...
            default:
                assert(0);
            }

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