Union member access

[Default User]

After a discussion on c.l.c++ with Ron Natalie, I'm now unclear as to
the status of reading data stored as one union member, then accessed
with a different one, in C99.

I don't have the latest standard, as I don't work in it currently, so I
use the draft standard.

My reading was that this is implementation-defined, based on this:

6.5.2.2
[#5] With one exception, if the value of a member of a union
object is used when the most recent store to the object was
to a different member, the behavior is
implementation-defined.70)
Ron Natalie states that it's usually undefined behaviro, according to
the following:

6.5
[#7] An object shall have its stored value accessed only by
an lvalue expression that has one of the following types:63)

-- a type compatible with the effective type of the
object,

-- a qualified version of a type compatible with the
effective type of the object,

-- a type that is the signed or unsigned type
corresponding to the effective type of the object,

-- a type that is the signed or unsigned type
corresponding to a qualified version of the effective
type of the object,

-- an aggregate or union type that includes one of the
aforementioned types among its members (including,
recursively, a member of a subaggregate or contained
union), or

-- a character type.

So that unless the other union member meets these qualifications, it's
UB. Is this correct? Was 6.5.2.2 [#5] modified, perhaps?

Brian Rodenborn

[Peter Nilsson]

"Default User" <fi********@boeing.com.invalid> wrote in message
news:40***************@boeing.com.invalid...

After a discussion on c.l.c++ with Ron Natalie, I'm now unclear as to
the status of reading data stored as one union member, then accessed
with a different one, in C99.

I don't have the latest standard, as I don't work in it currently, so I
use the draft standard.

My reading was that this is implementation-defined, based on this:

6.5.2.2
ITYM 6.5.2.3
[#5] With one exception, if the value of a member of a union
object is used when the most recent store to the object was
to a different member, the behavior is
implementation-defined.70)

That sentence was removed from the published C99 standard.

--
Peter

[Mathew Hendry]

On Sat, 17 Jul 2004 12:12:18 +1000, "Peter Nilsson" <ai***@acay.com.au>
wrote:

"Default User" <fi********@boeing.com.invalid> wrote in message
news:40***************@boeing.com.invalid...

ITYM 6.5.2.3

[#5] With one exception, if the value of a member of a union
object is used when the most recent store to the object was
to a different member, the behavior is
implementation-defined.70)

That sentence was removed from the published C99 standard.

It does say (6.2.6.1)

|7 When a value is stored in a member of an object of union type, the
|bytes of the object representation that do not correspond to that
|member but do correspond to other members take unspecified values,
|but the value of the union object shall not thereby become a trap
|representation.

which somehow becomes in Annex J

|J.1 Unspecified behavior
|1 The following are unspecified:
|...
|— The value of a union member other than the last one stored into
|(6.2.6.1)

That may be the intent, but Annex J is non-normative, and I don't see how
one gets from one to the other...

-- Mat.

[Default User]

Peter Nilsson wrote:

"Default User" <fi********@boeing.com.invalid> wrote in message
news:40***************@boeing.com.invalid...

[#5] With one exception, if the value of a member of a union
object is used when the most recent store to the object was
to a different member, the behavior is
implementation-defined.70)

That sentence was removed from the published C99 standard.

Ok, thanks.

Brian Rodenborn