www.digitalmars.com         C & C++   DMDScript  

D - Bug: Properties of pointers

reply Russell Lewis <spamhole-2001-07-16 deming-os.org> writes:
I tried this:
     struct asdf { uint thing1; uint thing2; uint thing3; uint thing4; };
     asdf foo;
     asdf* bar;

     assert(foo.size == 16);  // works
     assert((*bar).size == 16); // works
     assert(bar.size == 4); // fails!!!

It says it fails with "16 == 4", which tells me that when I do bar.size, 
  I'm getting the size of the structure underlying the pointer, NOT the 
pointer itself.  However, if I declare a pointer-to-pointer, I get a 
size of 4:
     asdf** fred;
     assert(fred.size == 4); // works



So first of all, does the property of a pointer refer to the pointer, or 
to the thing it points to?  If the former, then how do we get info about 
the pointer?  If the latter, then shouldn't it recurse down through all 
pointers?
Sep 11 2002
next sibling parent Russell Lewis <spamhole-2001-07-16 deming-os.org> writes:
Russell Lewis wrote:
 I tried this:
     struct asdf { uint thing1; uint thing2; uint thing3; uint thing4; };
     asdf foo;
     asdf* bar;
 
     assert(foo.size == 16);  // works
     assert((*bar).size == 16); // works
     assert(bar.size == 4); // fails!!!
 
 It says it fails with "16 == 4", which tells me that when I do bar.size, 
  I'm getting the size of the structure underlying the pointer, NOT the 
 pointer itself.  However, if I declare a pointer-to-pointer, I get a 
 size of 4:
     asdf** fred;
     assert(fred.size == 4); // works
 
 
 
 So first of all, does the property of a pointer refer to the pointer, or 
 to the thing it points to?  If the former, then how do we get info about 
 the pointer?  If the latter, then shouldn't it recurse down through all 
 pointers?
 
FYI: I discovered this with Burt's Linux port.
Sep 11 2002
prev sibling parent reply "Walter" <walter digitalmars.com> writes:
"Russell Lewis" <spamhole-2001-07-16 deming-os.org> wrote in message
news:3D7F9ECF.4090406 deming-os.org...
 I tried this:
      struct asdf { uint thing1; uint thing2; uint thing3; uint thing4; };
      asdf foo;
      asdf* bar;

      assert(foo.size == 16);  // works
      assert((*bar).size == 16); // works
      assert(bar.size == 4); // fails!!!

 It says it fails with "16 == 4", which tells me that when I do bar.size,
   I'm getting the size of the structure underlying the pointer, NOT the
 pointer itself.  However, if I declare a pointer-to-pointer, I get a
 size of 4:
      asdf** fred;
      assert(fred.size == 4); // works
 So first of all, does the property of a pointer refer to the pointer, or
 to the thing it points to?  If the former, then how do we get info about
 the pointer?  If the latter, then shouldn't it recurse down through all
 pointers?
What's happening here is that when faced with a pointer followed by a dot, the semantic analysis inserts one pointer derefence 'for free', so to speak. Clearly, in this case it should not. I'll see about fixing it.
Sep 12 2002
parent reply Pavel Minayev <evilone omen.ru> writes:
Walter wrote:

 What's happening here is that when faced with a pointer followed by a dot,
 the semantic analysis inserts one pointer derefence 'for free', so to speak.
Why doesn't the compiler dereference it several times if necessary? It would be logical...
Sep 12 2002
parent reply "Walter" <walter digitalmars.com> writes:
"Pavel Minayev" <evilone omen.ru> wrote in message
news:alqha5$1ii8$1 digitaldaemon.com...
 Walter wrote:
 What's happening here is that when faced with a pointer followed by a
dot,
 the semantic analysis inserts one pointer derefence 'for free', so to
speak.
 Why doesn't the compiler dereference it several times if necessary?
 It would be logical...
I'm a little unsure about that. Let's just see how far we get with one level for the moment.
Sep 12 2002
parent reply "Sean L. Palmer" <seanpalmer earthlink.net> writes:
How else will you dereference more than one level?  prefix operator * and
parens?

Sean

"Walter" <walter digitalmars.com> wrote in message
news:alqmae$2702$3 digitaldaemon.com...
 "Pavel Minayev" <evilone omen.ru> wrote in message
 news:alqha5$1ii8$1 digitaldaemon.com...
 Walter wrote:
 What's happening here is that when faced with a pointer followed by a
dot,
 the semantic analysis inserts one pointer derefence 'for free', so to
speak.
 Why doesn't the compiler dereference it several times if necessary?
 It would be logical...
I'm a little unsure about that. Let's just see how far we get with one
level
 for the moment.
Sep 13 2002
parent "Walter" <walter digitalmars.com> writes:
Yes.

"Sean L. Palmer" <seanpalmer earthlink.net> wrote in message
news:als2js$p9o$1 digitaldaemon.com...
 How else will you dereference more than one level?  prefix operator * and
 parens?

 Sean

 "Walter" <walter digitalmars.com> wrote in message
 news:alqmae$2702$3 digitaldaemon.com...
 "Pavel Minayev" <evilone omen.ru> wrote in message
 news:alqha5$1ii8$1 digitaldaemon.com...
 Walter wrote:
 What's happening here is that when faced with a pointer followed by
a
 dot,
 the semantic analysis inserts one pointer derefence 'for free', so
to
 speak.
 Why doesn't the compiler dereference it several times if necessary?
 It would be logical...
I'm a little unsure about that. Let's just see how far we get with one
level
 for the moment.
Sep 13 2002