Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
typedef struct
170 views
Skip to first unread message
John McDermick
Oct 15, 2011, 2:38:11 PM10/15/11
to
Hello,
If I have this definition in a header file:
typedef struct A
{
int test;
} B;
What exactly is A and what is B ?
What does it mean when I declare like this:
typedef struct A
{
int test;
};
And what does it mean when I declare like this:
typedef struct
{
int test;
} B;
If I have this definition in a header file:
typedef struct A
{
int test;
} B;
What exactly is A and what is B ?
What does it mean when I declare like this:
typedef struct A
{
int test;
};
And what does it mean when I declare like this:
typedef struct
{
int test;
} B;
Eric Sosman
Oct 15, 2011, 3:40:07 PM10/15/11
to
On 10/15/2011 2:38 PM, John McDermick wrote:
> Hello,
>
> If I have this definition in a header file:
>
> typedef struct A
> {
> int test;
> } B;
>
> What exactly is A and what is B ?
`A' is a "struct tag," something that distinguishes one particular
> Hello,
>
> If I have this definition in a header file:
>
> typedef struct A
> {
> int test;
> } B;
>
> What exactly is A and what is B ?
kind of struct from other kinds. The combination `struct A' names the
type of that particular kind of struct, so you can use that name in
other contexts like `struct A instance;' to declare a variable of that
struct type or `struct A *pointer;' to declare a variable that points
to objects of that type. `A' by itself means nothing in C; the rules
in C++ are different.
`B' is an alias for the type `struct A', and can be used wherever
`struct A' could. So `B instance;' declares a variable of the struct
type, and `B *pointer;' declares a variable that can point to it. `B'
and `struct A' are interchangeable.
> What does it mean when I declare like this:
>
> typedef struct A
> {
> int test;
> };
Then you start to declare an alias for `struct A' but never provide the
alias' name. I'm not sure the Standard requires a diagnostic for this
vacuous attempt at an alias, but compilers are likely to warn about it
in any case. It's a lot like writing `typedef float;' and never giving
the name that you want to use as an alias for `float'.
> And what does it mean when I declare like this:
>
> typedef struct
> {
> int test;
> } B;
own: There is no way to complete `struct ???' to name the type of
this struct. However, you also declare `B' as an alias for that
anonymous struct type, so you can still write `B instance;' and
`B *pointer;'. So, `B' is an alias for a type that has no true name
of its own, but `B' will serve as a perfectly good substitute.
--
Eric Sosman
eso...@ieee-dot-org.invalid
BartC
Oct 15, 2011, 3:42:01 PM10/15/11
to
"John McDermick" <johnth...@gmail.com> wrote in message
news:ff89df4d-151e-4537...@n13g2000vbv.googlegroups.com...
> Hello,
>
> If I have this definition in a header file:
>
> typedef struct A
> {
> int test;
> } B;
>
> What exactly is A and what is B ?
It gets more confusing when you write it like this:
>
> If I have this definition in a header file:
>
> typedef struct A
> {
> int test;
> } B;
>
> What exactly is A and what is B ?
typedef struct A
{
int test;
Both A and A can co-exist, because the first A (A in your example) needs to
be used as:
struct A x;
while the second A (B in your example) can be written like this:
A x;
> What does it mean when I declare like this:
>
> typedef struct A
> {
> int test;
> };
just writing struct A{int test;}. In fact, gcc will complain about it so it
agrees with me!
> And what does it mean when I declare like this:
>
> typedef struct
> {
> int test;
> } B;
--
Bartc
Keith Thompson
Oct 15, 2011, 7:26:51 PM10/15/11
to
Opinions differ on this, but personally I don't feel the need for a
typedef at all. I'd just declare the type as
struct A {
int test;
};
and then refer to it as "struct A". There's no real need to use a
typedef to give the type another name; it already has a perfectly good
one.
(The counterargument is that many people like to have a one-word name
for the type.)
You can also use the same identifier for both the tag and the typedef:
typedef struct A {
int test;
You can then refer to the type either as "struct A" or as "A". Since
struct tags are in a different namespace than typedef names (since the
tag can only appear immediately after the keyword "struct', and the
typedef name never can), there's no conflict.
(Note that C++ makes the struct tag usable directly as a type name; even
without the typedef, you can refer to the type as just "A". C doesn't
do this.)
--
Keith Thompson (The_Other_Keith) ks...@mib.org <http://www.ghoti.net/~kst>
"We must do something. This is something. Therefore, we must do this."
-- Antony Jay and Jonathan Lynn, "Yes Minister"
John McDermick
Oct 16, 2011, 12:48:10 PM10/16/11
to
Awesome and easy to understand explanations!
Thank you to all of you :o)
Thank you to all of you :o)
Kulin
May 31, 2012, 3:10:16 PM5/31/12
to
Keith Thompson <ks...@mib.org> wrote:
> John McDermick <johnth...@gmail.com> writes:
> > If I have this definition in a header file:
> >
> > typedef struct A
> > {
> > int test;
> > } B;
> >
> > What exactly is A and what is B ?
What if I want to create a pointer to the struct inside the struct?
> John McDermick <johnth...@gmail.com> writes:
> > If I have this definition in a header file:
> >
> > typedef struct A
> > {
> > int test;
> > } B;
> >
> > What exactly is A and what is B ?
typedef struct A {
A *p1; /* not allowed */
B *p2; /* not allowed */
} B;
This gives compilation errors. Is it possible to create a pointer to the
struct being defined, within the struct, using the original typedef and
thereby not having to code 'struct' to define the pointer?
> struct A {
> struct A *p;
> };
> and then refer to it as "struct A". There's no real need to use a
> typedef to give the type another name; it already has a perfectly good
> one.
>
> (The counterargument is that many people like to have a one-word name
> for the type.)
My desire is simply to avoid having to create pointers as struct * inside
> and then refer to it as "struct A". There's no real need to use a
> typedef to give the type another name; it already has a perfectly good
> one.
>
> (The counterargument is that many people like to have a one-word name
> for the type.)
the structure. Other than that I would tend to agree.
Eric Sosman
May 31, 2012, 3:14:00 PM5/31/12
to
struct A { // "completes" the type `struct A'
int test;
B *p1;
};
--
Eric Sosman
eso...@ieee-dot-org.invalid
Xavier Roche
May 31, 2012, 3:14:32 PM5/31/12
to
Le 31/05/2012 21:10, Kulin a écrit :
> What if I want to create a pointer to the struct inside the struct?
> typedef struct A {
> A *p1; /* not allowed */
> B *p2; /* not allowed */
> } B;
Hum, why don't you first define the type incompletely ?
> What if I want to create a pointer to the struct inside the struct?
> typedef struct A {
> A *p1; /* not allowed */
> B *p2; /* not allowed */
> } B;
typedef struct A A;
struct A {
A *p1;
A *p2;
};
Keith Thompson
May 31, 2012, 5:24:30 PM5/31/12
to
Kulin <rema...@reece.net.au> writes:
[...]
But if you want to use a typedef, there's no real need to use different
identifiers for the struct tag and the typedef name. This:
typedef struct A A;
is perfectly valid and doesn't create any potential ambiguity. Or at
least use a consistent convention, like:
typedef struct A_ A;
Will write code for food.
[...]
> What if I want to create a pointer to the struct inside the struct?
>
> typedef struct A {
> A *p1; /* not allowed */
> B *p2; /* not allowed */
> } B;
>
> This gives compilation errors. Is it possible to create a pointer to the
> struct being defined, within the struct, using the original typedef and
> thereby not having to code 'struct' to define the pointer?
[snip]
>
> typedef struct A {
> A *p1; /* not allowed */
> B *p2; /* not allowed */
> } B;
>
> This gives compilation errors. Is it possible to create a pointer to the
> struct being defined, within the struct, using the original typedef and
> thereby not having to code 'struct' to define the pointer?
> My desire is simply to avoid having to create pointers as struct * inside
> the structure. Other than that I would tend to agree.
Why do you want to avoid using the "struct" keyword?
> the structure. Other than that I would tend to agree.
But if you want to use a typedef, there's no real need to use different
identifiers for the struct tag and the typedef name. This:
typedef struct A A;
is perfectly valid and doesn't create any potential ambiguity. Or at
least use a consistent convention, like:
typedef struct A_ A;
Will write code for food.
Nomen Nescio
May 31, 2012, 8:50:38 PM5/31/12
to
Thank you very much!
Carl
Carl
Nomen Nescio
Jun 1, 2012, 4:20:05 AM6/1/12
to
Keith Thompson <ks...@mib.org> wrote:
> Kulin <rema...@reece.net.au> writes:
> [...]
> > What if I want to create a pointer to the struct inside the struct?
> >
> > typedef struct A {
> > A *p1; /* not allowed */
> > B *p2; /* not allowed */
> > } B;
> >
> > This gives compilation errors. Is it possible to create a pointer to the
> > struct being defined, within the struct, using the original typedef and
> > thereby not having to code 'struct' to define the pointer?
> [snip]
> > My desire is simply to avoid having to create pointers as struct * inside
> > the structure. Other than that I would tend to agree.
>
> Why do you want to avoid using the "struct" keyword?
If you have a struct with several self-referential pointers it looks
> Kulin <rema...@reece.net.au> writes:
> [...]
> > What if I want to create a pointer to the struct inside the struct?
> >
> > typedef struct A {
> > A *p1; /* not allowed */
> > B *p2; /* not allowed */
> > } B;
> >
> > This gives compilation errors. Is it possible to create a pointer to the
> > struct being defined, within the struct, using the original typedef and
> > thereby not having to code 'struct' to define the pointer?
> [snip]
> > My desire is simply to avoid having to create pointers as struct * inside
> > the structure. Other than that I would tend to agree.
>
> Why do you want to avoid using the "struct" keyword?
cluttered to have to say struct struct struct. Really no more than that.
> But if you want to use a typedef, there's no real need to use different
> identifiers for the struct tag and the typedef name. This:
>
> typedef struct A A;
>
> is perfectly valid and doesn't create any potential ambiguity. Or at
> least use a consistent convention, like:
>
> typedef struct A_ A;
Carl
Keith Thompson
Jun 1, 2012, 2:45:33 PM6/1/12
to
Nomen Nescio <nob...@dizum.com> writes:
> Keith Thompson <ks...@mib.org> wrote:
>> Kulin <rema...@reece.net.au> writes:
[...]
> Keith Thompson <ks...@mib.org> wrote:
>> Kulin <rema...@reece.net.au> writes:
[...]
>> Why do you want to avoid using the "struct" keyword?
>
> If you have a struct with several self-referential pointers it looks
> cluttered to have to say struct struct struct. Really no more than that.
[...]
>
> If you have a struct with several self-referential pointers it looks
> cluttered to have to say struct struct struct. Really no more than that.
> Carl
My question was directed to the previous poster, Kulin. Are you posting
under multiple names?
Ian Collins
Jun 1, 2012, 3:52:21 PM6/1/12
to
On 06/ 2/12 06:45 AM, Keith Thompson wrote:
> Nomen Nescio<nob...@dizum.com> writes:
>> Keith Thompson<ks...@mib.org> wrote:
>>> Kulin<rema...@reece.net.au> writes:
> [...]
>>> Why do you want to avoid using the "struct" keyword?
>>
>> If you have a struct with several self-referential pointers it looks
>> cluttered to have to say struct struct struct. Really no more than that.
> [...]
>> Carl
>
> My question was directed to the previous poster, Kulin. Are you posting
> under multiple names?
He or she certainly does on other groups and mail lists.
> Nomen Nescio<nob...@dizum.com> writes:
>> Keith Thompson<ks...@mib.org> wrote:
>>> Kulin<rema...@reece.net.au> writes:
> [...]
>>> Why do you want to avoid using the "struct" keyword?
>>
>> If you have a struct with several self-referential pointers it looks
>> cluttered to have to say struct struct struct. Really no more than that.
> [...]
>> Carl
>
> My question was directed to the previous poster, Kulin. Are you posting
> under multiple names?
--
Ian Collins
Ike Naar
Jun 1, 2012, 5:54:00 PM6/1/12
to
On 2012-06-01, Keith Thompson <ks...@mib.org> wrote:
> Nomen Nescio <nob...@dizum.com> writes:
>> Keith Thompson <ks...@mib.org> wrote:
>>> Kulin <rema...@reece.net.au> writes:
> [...]
>>> Why do you want to avoid using the "struct" keyword?
>>
>> If you have a struct with several self-referential pointers it looks
>> cluttered to have to say struct struct struct. Really no more than that.
> [...]
>> Carl
>
> My question was directed to the previous poster, Kulin. Are you posting
> under multiple names?
Remember "George Orwell", "Borked Pseudo Mailed", "Han from China" et al. ?
> Nomen Nescio <nob...@dizum.com> writes:
>> Keith Thompson <ks...@mib.org> wrote:
>>> Kulin <rema...@reece.net.au> writes:
> [...]
>>> Why do you want to avoid using the "struct" keyword?
>>
>> If you have a struct with several self-referential pointers it looks
>> cluttered to have to say struct struct struct. Really no more than that.
> [...]
>> Carl
>
> My question was directed to the previous poster, Kulin. Are you posting
> under multiple names?
BartC
Jun 1, 2012, 6:36:45 PM6/1/12
to
"Keith Thompson" <ks...@mib.org> wrote in message
news:ln7gvqh...@nuthaus.mib.org...
But with no warning that they will be using different names, and no
signature on the posts either, things get confusing.
--
Bartc
news:ln7gvqh...@nuthaus.mib.org...
> Nomen Nescio <nob...@dizum.com> writes:
>> Keith Thompson <ks...@mib.org> wrote:
>>> Kulin <rema...@reece.net.au> writes:
> [...]
>>> Why do you want to avoid using the "struct" keyword?
>>
>> If you have a struct with several self-referential pointers it looks
>> cluttered to have to say struct struct struct. Really no more than that.
> [...]
>> Carl
>
> My question was directed to the previous poster, Kulin. Are you posting
> under multiple names?
"Nomen Nescio" said on another group that he/she uses 'anonymous remailers'.
>> Keith Thompson <ks...@mib.org> wrote:
>>> Kulin <rema...@reece.net.au> writes:
> [...]
>>> Why do you want to avoid using the "struct" keyword?
>>
>> If you have a struct with several self-referential pointers it looks
>> cluttered to have to say struct struct struct. Really no more than that.
> [...]
>> Carl
>
> My question was directed to the previous poster, Kulin. Are you posting
> under multiple names?
But with no warning that they will be using different names, and no
signature on the posts either, things get confusing.
--
Bartc
Fritz Wuehler
Jun 2, 2012, 5:31:06 PM6/2/12
to
Keith Thompson <ks...@mib.org> wrote:
> Nomen Nescio <nob...@dizum.com> writes:
> > Keith Thompson <ks...@mib.org> wrote:
> >> Kulin <rema...@reece.net.au> writes:
> [...]
> >> Why do you want to avoid using the "struct" keyword?
> >
> > If you have a struct with several self-referential pointers it looks
> > cluttered to have to say struct struct struct. Really no more than that.
> [...]
> > Carl
>
> My question was directed to the previous poster, Kulin. Are you posting
> under multiple names?
No, I'm posting as myself but it gets posted through various remailers!
> Nomen Nescio <nob...@dizum.com> writes:
> > Keith Thompson <ks...@mib.org> wrote:
> >> Kulin <rema...@reece.net.au> writes:
> [...]
> >> Why do you want to avoid using the "struct" keyword?
> >
> > If you have a struct with several self-referential pointers it looks
> > cluttered to have to say struct struct struct. Really no more than that.
> [...]
> > Carl
>
> My question was directed to the previous poster, Kulin. Are you posting
> under multiple names?
Thanks for the help!
Carl
Keith Thompson
Jun 2, 2012, 9:30:57 PM6/2/12
to
Fritz Wuehler <fr...@spamexpire-201206.rodent.frell.theremailer.net>
so we can tell it's you?
writes:
> Keith Thompson <ks...@mib.org> wrote:
>> Nomen Nescio <nob...@dizum.com> writes:
>> > Keith Thompson <ks...@mib.org> wrote:
>> >> Kulin <rema...@reece.net.au> writes:
>> [...]
>> >> Why do you want to avoid using the "struct" keyword?
>> >
>> > If you have a struct with several self-referential pointers it looks
>> > cluttered to have to say struct struct struct. Really no more than that.
>> [...]
>> > Carl
>>
>> My question was directed to the previous poster, Kulin. Are you posting
>> under multiple names?
>
> No, I'm posting as myself but it gets posted through various remailers!
>
> Thanks for the help!
Would you consider adding some consistent header line to all your posts
> Keith Thompson <ks...@mib.org> wrote:
>> Nomen Nescio <nob...@dizum.com> writes:
>> > Keith Thompson <ks...@mib.org> wrote:
>> >> Kulin <rema...@reece.net.au> writes:
>> [...]
>> >> Why do you want to avoid using the "struct" keyword?
>> >
>> > If you have a struct with several self-referential pointers it looks
>> > cluttered to have to say struct struct struct. Really no more than that.
>> [...]
>> > Carl
>>
>> My question was directed to the previous poster, Kulin. Are you posting
>> under multiple names?
>
> No, I'm posting as myself but it gets posted through various remailers!
>
> Thanks for the help!
so we can tell it's you?
张源 西北工业大学
Jun 8, 2012, 1:02:14 PM6/8/12
to
> struct A { // "completes" the type `struct A'
> int test;
> B *p1;
> };
>
> --
> Eric Sosman
> eso...@ieee-dot-org.invalid
Eric Sosman
Jun 8, 2012, 1:20:16 PM6/8/12
to
On 6/8/2012 1:02 PM, 张源 西北工业大学 wrote:
> On Friday, June 1, 2012 3:14:00 AM UTC+8, Eric Sosman wrote:
>> On 5/31/2012 3:10 PM, Kulin wrote:
>>> [...]
> On Friday, June 1, 2012 3:14:00 AM UTC+8, Eric Sosman wrote:
>> On 5/31/2012 3:10 PM, Kulin wrote:
>>> My desire is simply to avoid having to create pointers as struct * inside
>>> the structure. Other than that I would tend to agree.
>>
>> typedef struct A B; // "incomplete type" declaration
>
> can i understand in this way:the ambiguity of struct A is B ?
I don't know what you mean by "the ambiguity of." The line
>>> the structure. Other than that I would tend to agree.
>>
>> typedef struct A B; // "incomplete type" declaration
>
> can i understand in this way:the ambiguity of struct A is B ?
actually says two things:
1) A type named `struct A' exists. The type is "incomplete,"
meaning that there is no information about the number of bytes a
`struct A' requires or what its member elements are. Perhaps the
code will provide such details later to "complete" the type, but
no details are available yet.
2) `B' is an alias for `struct A'. `B' and `struct A' can
be used interchangeably; both refer to the exact same type. Since
`struct A' is an incomplete type, `B' is also an incomplete type
(because it is, in fact, the same type).
>> struct A { // "completes" the type `struct A'
>> int test;
>> B *p1;
>> };
struct elements. Since `struct A' is now complete, its alias `B'
is also complete: As before, `struct A' and `B' mean the same thing.
--
Eric Sosman
eso...@ieee-dot-org.invalid
张源 西北工业大学
Jun 8, 2012, 1:20:03 PM6/8/12
to
such as some basic implementation on computer for this incompletely typedef?
(sorry for my awful English)
张源 西北工业大学
Jun 8, 2012, 1:27:39 PM6/8/12
to
can you exlain why the code below are forbidden ?
Eric Sosman
Jun 8, 2012, 1:32:26 PM6/8/12
to
On 6/8/2012 1:27 PM, 张源 西北工业大学 wrote:
> [...]
has no idea at all what `p1' should point to.
> B *p2; /* not allowed */
`B' means nothing (yet), because `B' has not been declared. The
compiler has no idea what `p2' should point to.
> } B;
>
> This gives compilation errors.
Yes.
--
Eric Sosman
eso...@ieee-dot-org.invalid
> [...]
> thanks a lot,sorry for my awful english,since i'm a Chinese.
> can you exlain why the code below are forbidden ?
> typedef struct A {
> A *p1; /* not allowed */
`struct A' is the name of a type, but `A' alone is not. The compiler
> can you exlain why the code below are forbidden ?
> typedef struct A {
> A *p1; /* not allowed */
has no idea at all what `p1' should point to.
> B *p2; /* not allowed */
compiler has no idea what `p2' should point to.
> } B;
>
> This gives compilation errors.
--
Eric Sosman
eso...@ieee-dot-org.invalid
张源 西北工业大学
Jun 8, 2012, 1:37:46 PM6/8/12
to
On Saturday, June 9, 2012 1:32:26 AM UTC+8, Eric Sosman wrote:
Thank you for your patience .I'm new to C.
Can you give me some advice?
blog or maillist?
Thank you,this is first time i contact with others in google group.I like the atmosphere here.
Can you give me some advice?
blog or maillist?
Thank you,this is first time i contact with others in google group.I like the atmosphere here.
Eric Sosman
Jun 8, 2012, 1:55:22 PM6/8/12
to
On 6/8/2012 1:37 PM, 张源 西北工业大学 wrote:
> [...]
> [...]
> Thank you for your patience .I'm new to C.
> Can you give me some advice?
> blog or maillist?
> Thank you,this is first time i contact with others in google group.I like the atmosphere here.
This is not "Google group;" this is Usenet. Usenet has been around since
> Can you give me some advice?
> blog or maillist?
> Thank you,this is first time i contact with others in google group.I like the atmosphere here.
Google's founders were three years old. Their company does many things well,
but the "Google Groups" interface to Usenet is not among their successes.
If you are new to C, get yourself a good book. "The C Programming
Language"
by Kernighan and Ritchie is a good one if you already have experience
with other
programming languages. (It will teach you C, but will not teach you how
to write
programs.) Other books exist, some good and some not. I do not know whether
there are good C books in Chinese.
Another useful source is the comp.lang.c Frequently Asked Questions
(FAQ)
page at <http://www.c-faq.com/>. The coverage is somewhat spotty since
it tries
to address question that arise frequently, and some answers are a little
dated.
Nonetheless, it's an excellent place to look before posting a question.
There's also a Usenet group called "alt.comp.lang.learn.c-c++". I know
nothing about its quality or usefulness, but the title looks helpful.
There might
be some Chinese-language C groups or forums; I don't know.
--
Eric Sosman
eso...@ieee-dot-org.invalid
Zhang Yuan
Jun 8, 2012, 2:22:10 PM6/8/12
to
On Saturday, June 9, 2012 1:55:22 AM UTC+8, Eric Sosman wrote:
Thank you.
the quality of this usenet is higher than those in china .
chinese students prefer to read american books rather than natives.
i'm reading c primer plus now.
thank you,
It's a pitty that cfaq is blocked in china. - -!
the quality of this usenet is higher than those in china .
chinese students prefer to read american books rather than natives.
i'm reading c primer plus now.
thank you,
It's a pitty that cfaq is blocked in china. - -!
James Kuyper
Jun 8, 2012, 2:31:00 PM6/8/12
to
to declare an incomplete struct type:
struct incomplete_type;
struct another_incomplete_type *pointer_to_incomplete_type;
The key point is that for some purposes there's no need to know the
actual type of the struct. This is feasible because the standard
requires that all pointers to structs have the same representation and
alignment requirements, so you don't need to tell the compiler the exact
type, so long as all you're doing is receiving pointers to the type, and
passing them on to other functions that do know what the exact type is.
James Kuyper
Jun 8, 2012, 2:35:04 PM6/8/12
to
On 06/08/2012 02:22 PM, Zhang Yuan wrote:
...
known that their censors are crazy, I'm just not clear on the particular
brand of craziness that produced that decision.
...
> It's a pitty that cfaq is blocked in china. - -!
What in the world? Why do they find objectionable about the cfaq? I
known that their censors are crazy, I'm just not clear on the particular
brand of craziness that produced that decision.
Zhang Yuan
Jun 8, 2012, 2:41:40 PM6/8/12
to
This is not difficult for students in china .
pitty.sometimes ,i even can't go to google group.
Message has been deleted
Message has been deleted
Zhang Yuan
Jun 8, 2012, 2:08:52 PM6/8/12
to
On Saturday, June 9, 2012 1:55:22 AM UTC+8, Eric Sosman wrote:
Thank you.This Usenet is more helpful than those in china.
Students in china usually read books translated from English .
I'm reading c primer plus in Chinese.
Thank you!
Students in china usually read books translated from English .
I'm reading c primer plus in Chinese.
Thank you!
James Kuyper
Jun 8, 2012, 3:10:59 PM6/8/12
to
> On Saturday, June 9, 2012 2:31:00 AM UTC+8, James Kuyper wrote:
>> On 06/08/2012 01:20 PM, 张源 西北工业大学 wrote:
...
>> On 06/08/2012 01:20 PM, 张源 西北工业大学 wrote:
>> Being able to typedef incompletely is just a side effect of being able
>> to declare an incomplete struct type:
>>
>> struct incomplete_type;
>> struct another_incomplete_type *pointer_to_incomplete_type;
>>
>> The key point is that for some purposes there's no need to know the
>> actual type of the struct. This is feasible because the standard
>> requires that all pointers to structs have the same representation and
>> alignment requirements, so you don't need to tell the compiler the exact
>> type, so long as all you're doing is receiving pointers to the type, and
>> passing them on to other functions that do know what the exact type is.
>
> thank you.i know a lot .
>> to declare an incomplete struct type:
>>
>> struct incomplete_type;
>> struct another_incomplete_type *pointer_to_incomplete_type;
>>
>> The key point is that for some purposes there's no need to know the
>> actual type of the struct. This is feasible because the standard
>> requires that all pointers to structs have the same representation and
>> alignment requirements, so you don't need to tell the compiler the exact
>> type, so long as all you're doing is receiving pointers to the type, and
>> passing them on to other functions that do know what the exact type is.
>
> what do the same representation and alignment requirements mean?
Every C object is stored in a series of bits; except for bit-fields,
those bits usually fill up entire bytes. Every C type has an
implementation-dependent mapping which connects bit patterns to values.
That mapping is called the representation used by that type. Unsigned
integer types have a simple binary representation. C signed integers
have one of three types of representation: 1's complement, 2's
complement, and sign-magnitude. They all store positive numbers the same
way as the corresponding unsigned type, they differ only in the way they
represent negative values. The C standard imposes no requirements on the
representation of floating point types, but does say that if
__STDC_IEC_559__ is pre-defined by the implementation, the
implementation of floating point types should conform to the
requirements of annex F, which details the ways in which they are
required to conform to the requirements of IEEE/IEC 60559.
Byte ordering (bigendian, littleendian, middleendian, etc.) is also part
of the representation. The standard imposes no requirements on byte
ordering. The C standard also says nothing about how pointers are
represented. That's entirely up to the implementation. However, it does
require that, in some cases, pointers to different types must be
represented the same way.
The alignment of a type is a requirement that objects of that type be
allocated only at particular types of addresses. For instance, if an
object may only be allocated at even addresses, it has an alignment
requirement of 2. If it can only be allocated at addresses which are
integer multiples of 4, it has an alignment requirement of 4.
Zhang Yuan
Jun 8, 2012, 3:27:57 PM6/8/12
to
your explanation is very explicit！
Keith Thompson
Jun 8, 2012, 5:22:20 PM6/8/12
to
Zhang Yuan <zhabgy...@gmail.com> writes:
> On Saturday, June 9, 2012 1:55:22 AM UTC+8, Eric Sosman wrote:
[...]
> On Saturday, June 9, 2012 1:55:22 AM UTC+8, Eric Sosman wrote:
>> Another useful source is the comp.lang.c Frequently Asked
>> Questions (FAQ) page at <http://www.c-faq.com/>. The coverage is
>> somewhat spotty since it tries to address question that arise
>> frequently, and some answers are a little dated. Nonetheless, it's
>> an excellent place to look before posting a question.
[...]
>> Questions (FAQ) page at <http://www.c-faq.com/>. The coverage is
>> somewhat spotty since it tries to address question that arise
>> frequently, and some answers are a little dated. Nonetheless, it's
>> an excellent place to look before posting a question.
> It's a pitty that cfaq is blocked in china. - -!
That's surprising.
The author of the FAQ is Steve Summit <s...@eskimo.com>. He might
be able to offer you some advice about how to obtain a copy.
(It's up to you to decide whether that's a good idea.)
Zhang Yuan
Jun 8, 2012, 9:28:49 PM6/8/12
to
I found copy in ftp://ftp.eskimo.com/u/s/scs/C-faq
thank you!
Robert Miles
Jun 29, 2012, 5:29:44 PM6/29/12
to
On 6/1/2012 1:45 PM, Keith Thompson wrote:
> Keith Thompson (The_Other_Keith)ks...@mib.org <http://www.ghoti.net/~kst>
is something. Therefore, we must do that too."
> Keith Thompson (The_Other_Keith)ks...@mib.org <http://www.ghoti.net/~kst>
> Will write code for food.
> "We must do something. This is something. Therefore, we must do this."
> -- Antony Jay and Jonathan Lynn, "Yes Minister"
Looks like someone needs to answer "Pushing you out a rather high window
> "We must do something. This is something. Therefore, we must do this."
> -- Antony Jay and Jonathan Lynn, "Yes Minister"
is something. Therefore, we must do that too."
Keith Thompson
Jun 29, 2012, 5:58:00 PM6/29/12
to
--
Keith Thompson (The_Other_Keith) ks...@mib.org <http://www.ghoti.net/~kst>
Robert Miles
Jul 1, 2012, 11:55:56 PM7/1/12
to
On 6/29/2012 4:58 PM, Keith Thompson wrote:
> Robert Miles <mil...@Usenet-News.net> writes:
>> On 6/1/2012 1:45 PM, Keith Thompson wrote:
>>> Keith Thompson (The_Other_Keith)ks...@mib.org <http://www.ghoti.net/~kst>
>>> Will write code for food.
>>> "We must do something. This is something. Therefore, we must do this."
>>> -- Antony Jay and Jonathan Lynn, "Yes Minister"
>>
>> Looks like someone needs to answer "Pushing you out a rather high window
>> is something. Therefore, we must do that too."
>
> Any particular reason?
Whoever that quote came from appears to need to think some more, and
> Robert Miles <mil...@Usenet-News.net> writes:
>> On 6/1/2012 1:45 PM, Keith Thompson wrote:
>>> Keith Thompson (The_Other_Keith)ks...@mib.org <http://www.ghoti.net/~kst>
>>> Will write code for food.
>>> "We must do something. This is something. Therefore, we must do this."
>>> -- Antony Jay and Jonathan Lynn, "Yes Minister"
>>
>> Looks like someone needs to answer "Pushing you out a rather high window
>> is something. Therefore, we must do that too."
>
> Any particular reason?
this answer should help start that thinking.
Keith Thompson
Jul 2, 2012, 2:00:26 AM7/2/12
to
--
Keith Thompson (The_Other_Keith) ks...@mib.org <http://www.ghoti.net/~kst>
James Kuyper
Jul 2, 2012, 7:12:07 AM7/2/12
to
On 07/02/2012 02:00 AM, Keith Thompson wrote:
> Robert Miles <mil...@Usenet-News.net> writes:
>> On 6/29/2012 4:58 PM, Keith Thompson wrote:
>>> Robert Miles <mil...@Usenet-News.net> writes:
>>>> On 6/1/2012 1:45 PM, Keith Thompson wrote:
>>>>> Keith Thompson (The_Other_Keith)ks...@mib.org <http://www.ghoti.net/~kst>
>>>>> Will write code for food.
>>>>> "We must do something. This is something. Therefore, we must do this."
>>>>> -- Antony Jay and Jonathan Lynn, "Yes Minister"
>>>>
>>>> Looks like someone needs to answer "Pushing you out a rather high window
>>>> is something. Therefore, we must do that too."
>>>
>>> Any particular reason?
>>
>> Whoever that quote came from appears to need to think some more, and
>> this answer should help start that thinking.
>
> Perhaps you missed the fact that the quote is intended to be humorous.
If was clearly intended to be humorous by the real-world author.
> Robert Miles <mil...@Usenet-News.net> writes:
>> On 6/29/2012 4:58 PM, Keith Thompson wrote:
>>> Robert Miles <mil...@Usenet-News.net> writes:
>>>> On 6/1/2012 1:45 PM, Keith Thompson wrote:
>>>>> Keith Thompson (The_Other_Keith)ks...@mib.org <http://www.ghoti.net/~kst>
>>>>> Will write code for food.
>>>>> "We must do something. This is something. Therefore, we must do this."
>>>>> -- Antony Jay and Jonathan Lynn, "Yes Minister"
>>>>
>>>> Looks like someone needs to answer "Pushing you out a rather high window
>>>> is something. Therefore, we must do that too."
>>>
>>> Any particular reason?
>>
>> Whoever that quote came from appears to need to think some more, and
>> this answer should help start that thinking.
>
> Perhaps you missed the fact that the quote is intended to be humorous.
However, was it intended to be humorous by the fictional character who
spoke that line? Having seen only a little of that show, I had assumed
that it was an expression of bureaucratic incompetence so extreme that
"Pushing you out a rather high window" would have led to a significant
improvement in government efficiency. There did seem to be several
characters on that show who could plausibly have spoken such a line
without humorous intent.
--
James Kuyper
0 new messages