C++Talk.NET C++ language newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  a[i] not equivalent to *(a+i)? Goto page 1, 2, 3, 4, 5  Next          C++Talk.NET Forum Index -> C++ Language (Moderated) View previous topic :: View next topic   Author Message Michal Kowalski Guest Posted: Mon Dec 08, 2003 10:58 am    Post subject: a[i] not equivalent to *(a+i)? Hello, I have a question regarding old good array and pointers. I remember reading a remark by Bjarne Stroustrup that a[i] is not equivalent to *(a+i) when it comes to reaching pass the end of an array. Say I have an array int a[100]. Then (a+100) is safe, but &a[100] is not. Master didn't explain what he meant by those two being not equivalent. I can kind of feel &a[100] is not quite right, but what does C++ standard say? Where does similarity end? Am I overlooking something obvious? Can someone comment on this issue please. Thank you, MK [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Francis Glassborow Guest Posted: Mon Dec 08, 2003 9:29 pm    Post subject: Re: a[i] not equivalent to *(a+i)? In article <M2XAb.67706$zY1.30782 (AT) newssvr25 (DOT) news.prodigy.com>, Michal Kowalski <michal_kowalski (AT) hotmail (DOT) com> writes Quote: Hello, I have a question regarding old good array and pointers. I remember reading a remark by Bjarne Stroustrup that a[i] is not equivalent to *(a+i) when it comes to reaching pass the end of an array. Say I have an array int a[100]. Then (a+100) is safe, but &a[100] is not. Master didn't explain what he meant by those two being not equivalent. I can kind of feel &a[100] is not quite right, but what does C++ standard say? Where does similarity end? Am I overlooking something obvious? Can someone comment on this issue please. I think this is strictly language law stuff. In theory &a[100] is equivalent to &*(a+100) which has the singular pointer (a+100) dereferenced before its address is taken. When dealing with raw C-style arrays I suspect this is purely of academic interest but were you to be dealing with a user defined subscript operator then it would be a very different issue. -- Francis Glassborow ACCU If you are not using up-to-date virus protection you should not be reading this. Viruses do not just hurt the infected but the whole community. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Maciej Sobczak Guest Posted: Mon Dec 08, 2003 9:36 pm    Post subject: Re: a[i] not equivalent to *(a+i)? Hi, Michal Kowalski wrote: Quote: to reaching pass the end of an array. Say I have an array int a[100]. Then (a+100) is safe, but &a[100] is not. The problem comes from the fact that Standard differentiates between a "valid pointer value" and "dereferenceable pointer". The first is a pointer which has legal value in the sense that you can compare such pointer to another. Or you can do pointer arithmetic on such values - one example of such pointer arithmetic is: int a[100]; int *p = (a + 100); // ok The second kind (dereferenceable) means that you can read the value that it points to. The difference is that the Standard requires the "one-past-the-end" pointer to be valid for all arrays, but does not require it to be dereferenceable. In other words, you can point to (a+100), but you cannot read it: int i = *p; // illegal! p -= 50; int i = *p; // but here ok, because p is dereferenceable The difference between (a+100) and &a[100] is that the first is only a pointer arithmetic which gives the "one-past-the-end", legal pointer value, whereas the second tries to get the address of *object*, which does not exist - in other words, a[100] is "executed" before & and a[100] is not a legal l-value, because it is (by definition) the same as *(a + 100) and (a + 100) is not dereferenceable. -- Maciej Sobczak : http://www.msobczak.com/ Programming : http://www.msobczak.com/prog/ [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top kanze@gabi-soft.fr Guest Posted: Mon Dec 08, 2003 10:42 pm    Post subject: Re: a[i] not equivalent to *(a+i)? "Michal Kowalski" <michal_kowalski (AT) hotmail (DOT) com> wrote Quote: I have a question regarding old good array and pointers. I remember reading a remark by Bjarne Stroustrup that a[i] is not equivalent to *(a+i) when it comes to reaching pass the end of an array. Say I have an array int a[100]. Then (a+100) is safe, but &a[100] is not. Master didn't explain what he meant by those two being not equivalent. I can kind of feel &a[100] is not quite right, but what does C++ standard say? Where does similarity end? Am I overlooking something obvious? Can someone comment on this issue please. I think that Stroustrup must have been anticipating a change which didn't take place. For built-in arrays, both in C++ and in C90, the expression is undefined behavior; in C99, there is a special rule which allows it ONLY as the operand of a unary & operator. For containers in the STL, it is also undefined behavior, and wht at least some implementations, it actually fails. -- James Kanze GABI Software mailto:kanze (AT) gabi-soft (DOT) fr Conseils en informatique orientée objet/ http://www.gabi-soft.fr Beratung in objektorientierter Datenverarbeitung 11 rue de Rambouillet, 78460 Chevreuse, France, +33 (0)1 30 23 45 16 [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Michal Kowalski Guest Posted: Tue Dec 09, 2003 10:53 am    Post subject: Re: a[i] not equivalent to *(a+i)? Thank you all, Your answers are very helpful. Just to give credit where it's due: I believe master was well aware of uncertain status of (a+100) construction James mentioned. MK [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Maciej Sobczak Guest Posted: Tue Dec 09, 2003 11:03 am    Post subject: Re: a[i] not equivalent to *(a+i)? Hi, Francis Glassborow wrote: Quote: In theory &a[100] is equivalent to &*(a+100) which has the singular pointer (a+100) Assuming that 24.1/5 applies to pointer to arrays: "Dereferenceable and past-the-end values are always nonsingular." Which brings us to the question what does it really mean that the pointer (or iterator) is "singular". This word is used in the Standard but not explained beside few words stating that iterator with singular values "are not associated with any container". I think that one-past-the-end is associated with the container. Is there any formal (or better clear ) definition what singular really means? Quote: When dealing with raw C-style arrays I suspect this is purely of academic interest Because the compiler can eliminate & and * standing one beside the other (this is exactly what we do mentally), but even outside the academy I can imagine implementations where such violations are trapped. -- Maciej Sobczak : http://www.msobczak.com/ Programming : http://www.msobczak.com/prog/ [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Francis Glassborow Guest Posted: Tue Dec 09, 2003 5:06 pm    Post subject: Re: a[i] not equivalent to *(a+i)? In article <hobBb.68066$hE3.57444 (AT) newssvr25 (DOT) news.prodigy.com>, Michal Kowalski <michal_kowalski (AT) hotmail (DOT) com> writes Quote: Your answers are very helpful. Just to give credit where it's due: I believe master was well aware of uncertain status of (a+100) construction James mentioned. I do not think that there is any uncertainty wrt to (a+100) the uncertainty is with &a[100]. Given: T a[100]; Both a[100] and *(a+100) give undefined behaviour. In practice I think most compilers will convert &a[100] to the well-defined a+100 but they are not required to do so. C99 IIRC explicitly requires that &* be treated as an identity operator but C++ does not. -- Francis Glassborow ACCU Author of 'You Can Do It!' see http://www.spellen.org/youcandoit [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top kanze@gabi-soft.fr Guest Posted: Tue Dec 09, 2003 8:04 pm    Post subject: Re: a[i] not equivalent to *(a+i)? Maciej Sobczak <no.spam (AT) no (DOT) smap.com> wrote Quote: Francis Glassborow wrote: In theory &a[100] is equivalent to &*(a+100) which has the singular pointer (a+100) Assuming that 24.1/5 applies to pointer to arrays: "Dereferenceable and past-the-end values are always nonsingular." Which brings us to the question what does it really mean that the pointer (or iterator) is "singular". This word is used in the Standard but not explained beside few words stating that iterator with singular values "are not associated with any container". I think that one-past-the-end is associated with the container. Is there any formal (or better clear ) definition what singular really means? When dealing with raw C-style arrays I suspect this is purely of academic interest Because the compiler can eliminate & and * standing one beside the other (this is exactly what we do mentally), but even outside the academy I can imagine implementations where such violations are trapped. You don't even need imagination. Centerline once sold such an implementation. -- James Kanze GABI Software mailto:kanze (AT) gabi-soft (DOT) fr Conseils en informatique orientée objet/ http://www.gabi-soft.fr Beratung in objektorientierter Datenverarbeitung 11 rue de Rambouillet, 78460 Chevreuse, France, +33 (0)1 30 23 45 16 [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Dietmar Kuehl Guest Posted: Tue Dec 09, 2003 10:03 pm    Post subject: Re: a[i] not equivalent to *(a+i)? "Michal Kowalski" <michal_kowalski (AT) hotmail (DOT) com> wrote: Quote: Your answers are very helpful. Just to give credit where it's due: I believe master was well aware of uncertain status of (a+100) construction James mentioned. Well for 'int a[100]', the expression '(a+100)' is perfectly valid and yields a non-dereferencable past the end iterator. The resulting pointer can only be used in certain expressions. It can, however, not be dereferenced. The expression '&a[100]' yields undefined behavior, because the sub-expression 'a[100]' would access a non-existing element. To answer the question from your subject for built-in arrays 'a[n]': - 'a[i]' is equivalent to '*(a + i)' (and this is also equivalent to '*(i + a)' which in turn is equivalent to 'i[a]'; but this is a completely irrelevant sidetrack...): both yield a reference to an element for '0 <= i && i < n' and are invalid otherwise. - '&a[i]' is equivalent to '(a + i)' if 'i != n'. In case 'i == n', the expression '(a+i)' is valid, while '&a[i]' is not. If 'i < 0 || n <= i', both expressions are invalid while for '0 <= i && i < n' both expressions are valid. Often, the case 'a + n' is important: it is used to obtain a "past the end"-iterator for an array object. -- Phaidros eaSE - Easy Software Engineering: <http://www.phaidros.com/> [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Bjarne Stroustrup Guest Posted: Wed Dec 10, 2003 11:28 am    Post subject: Re: a[i] not equivalent to *(a+i)? [email]kanze (AT) gabi-soft (DOT) fr[/email] wrote in message news:<d6652001.0312080748.40b60eac (AT) posting (DOT) google.com>... Quote: "Michal Kowalski" <michal_kowalski (AT) hotmail (DOT) com> wrote in message news:<M2XAb.67706$zY1.30782 (AT) newssvr25 (DOT) news.prodigy.com>... I have a question regarding old good array and pointers. I remember reading a remark by Bjarne Stroustrup that a[i] is not equivalent to *(a+i) when it comes to reaching pass the end of an array. Say I have an array int a[100]. Then (a+100) is safe, but &a[100] is not. Master didn't explain what he meant by those two being not equivalent. I can kind of feel &a[100] is not quite right, but what does C++ standard say? Where does similarity end? Am I overlooking something obvious? Can someone comment on this issue please. I think that Stroustrup must have been anticipating a change which didn't take place. For built-in arrays, both in C++ and in C90, the expression is undefined behavior; in C99, there is a special rule which allows it ONLY as the operand of a unary & operator. For containers in the STL, it is also undefined behavior, and wht at least some implementations, it actually fails. I think this is a case where one must be very careful attributions. On my "issues" page http://www.research.att.com/~bs/3rd_issues.html I say: ______ pg 42 I use char vc2[200]; copy(&vc1[0],&vc1[200],&vc2[200]); The issue is whether taking the address of one-past-the-last element of an array is conforming C and C++. I could make the example clearly conforming by a simple rewrite: copy(vc1,vc1+200,vc2+200); However, I don't want to introduce addition to pointers at this point of the book. It is a surprise to most experienced C and C++ programmers that &vc2[200] isn't completely equivalent to vc2+200. In fact, it was a surprise to the C committee also and I expect it to be fixed in the upcoming revision of the standard. (resolved for C9x - bs 10/13/9. (this means that the example was ok in K&R C, in ARM C++, an error in C89 and ISO C++, and ok in C9x - a thorny issue). _____ I suspect that "a thorny issue" is a good characterization, and that anything less than a longish explanation will add to the confusion. - Bjarne Stroustrup; http://www.research.att.com/~bs [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top John Potter Guest Posted: Wed Dec 10, 2003 11:30 am    Post subject: Re: a[i] not equivalent to *(a+i)? On 9 Dec 2003 06:03:54 -0500, Maciej Sobczak <no.spam (AT) no (DOT) smap.com> wrote: Quote: Francis Glassborow wrote: In theory &a[100] is equivalent to &*(a+100) which has the singular pointer (a+100) Assuming that 24.1/5 applies to pointer to arrays: "Dereferenceable and past-the-end values are always nonsingular." It does not. A, IMO misguided DR, has changed that to allow singular past-the-end values. Quote: Which brings us to the question what does it really mean that the pointer (or iterator) is "singular". This word is used in the Standard but not explained beside few words stating that iterator with singular values "are not associated with any container". The original intent used uninitialized pointers as an example of singular. It was stated that no operation other than assignment is permitted. The DR now states that past-the-end iterators may be singular which means that they can not be read for any reason including comparison. All this because they wanted to allow a null pointer to be an off the end iterator for a singly linked list. Because singular was never well defined, the common use of null as a singular == unique value was used when it was intended to mean an unusable value like a singularity of a function. John [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Matthias Hofmann Guest Posted: Wed Dec 10, 2003 8:24 pm    Post subject: Re: a[i] not equivalent to *(a+i)? John Potter <jpotter (AT) falcon (DOT) lhup.edu> schrieb in im Newsbeitrag: [email]lapctvck2p3rmvtsq7arqcsndc7b29kt06 (AT) 4ax (DOT) com[/email]... Quote: On 9 Dec 2003 06:03:54 -0500, Maciej Sobczak wrote: Francis Glassborow wrote: In theory &a[100] is equivalent to &*(a+100) which has the singular pointer (a+100) Assuming that 24.1/5 applies to pointer to arrays: "Dereferenceable and past-the-end values are always nonsingular." It does not. A, IMO misguided DR, has changed that to allow singular past-the-end values. Which brings us to the question what does it really mean that the pointer (or iterator) is "singular". This word is used in the Standard but not explained beside few words stating that iterator with singular values "are not associated with any container". The original intent used uninitialized pointers as an example of singular. It was stated that no operation other than assignment is permitted. The DR now states that past-the-end iterators may be singular which means that they can not be read for any reason including comparison. All this because they wanted to allow a null pointer to be an off the end iterator for a singly linked list. Because singular was never well defined, the common use of null as a singular == unique value was used when it was intended to mean an unusable value like a singularity of a function. So when is a past-the-end value singular and when is it not? What about the ubiquitous for ( it = list.begin(); it != list.end(); ++it ); If no comparison with past-the-end values are allowed, is this illegal code in future versions of the standard? Matthias [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top kanze@gabi-soft.fr Guest Posted: Wed Dec 10, 2003 8:32 pm    Post subject: Re: a[i] not equivalent to *(a+i)? [email]bs (AT) research (DOT) att.com[/email] (Bjarne Stroustrup) wrote in message news:<188b3370.0312091746.3e22a207 (AT) posting (DOT) google.com>... Quote: kanze (AT) gabi-soft (DOT) fr wrote in message news:<d6652001.0312080748.40b60eac (AT) posting (DOT) google.com>... "Michal Kowalski" <michal_kowalski (AT) hotmail (DOT) com> wrote in message news:<M2XAb.67706$zY1.30782 (AT) newssvr25 (DOT) news.prodigy.com>... I have a question regarding old good array and pointers. I remember reading a remark by Bjarne Stroustrup that a[i] is not equivalent to *(a+i) when it comes to reaching pass the end of an array. Say I have an array int a[100]. Then (a+100) is safe, but &a[100] is not. Master didn't explain what he meant by those two being not equivalent. I can kind of feel &a[100] is not quite right, but what does C++ standard say? Where does similarity end? Am I overlooking something obvious? Can someone comment on this issue please. I think that Stroustrup must have been anticipating a change which didn't take place. For built-in arrays, both in C++ and in C90, the expression is undefined behavior; in C99, there is a special rule which allows it ONLY as the operand of a unary & operator. For containers in the STL, it is also undefined behavior, and wht at least some implementations, it actually fails. I think this is a case where one must be very careful attributions. Or simply more careful reading of the orginal posting. I just missed a not. With regards to the technical issues: Quote: I suspect that "a thorny issue" is a good characterization, and that anything less than a longish explanation will add to the confusion. I'd say that this more or less sums it up. The simple statement is that &a[nbElem] is undefined behavior in C++. Regardless of whether a is a built in array or an STL container. Anything beyond that enters into the realm of implementation techniques and C compatibility, with probably a good deal of personal opinions regarding style thrown in as well. -- James Kanze GABI Software mailto:kanze (AT) gabi-soft (DOT) fr Conseils en informatique orientée objet/ http://www.gabi-soft.fr Beratung in objektorientierter Datenverarbeitung 11 rue de Rambouillet, 78460 Chevreuse, France, +33 (0)1 30 23 45 16 [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top John Potter Guest Posted: Wed Dec 10, 2003 8:33 pm    Post subject: Re: a[i] not equivalent to *(a+i)? On 10 Dec 2003 06:28:42 -0500, [email]bs (AT) research (DOT) att.com[/email] (Bjarne Stroustrup) wrote: Quote: I think this is a case where one must be very careful attributions. On my "issues" page http://www.research.att.com/~bs/3rd_issues.html I say: ______ pg 42 I use char vc2[200]; copy(&vc1[0],&vc1[200],&vc2[200]); Issues-errata: To make the issue clear, pg 42 has the correct &vc2[0] and does not copy off the end of vc2. The issue with &vc1[200] will be clearer without the added error. John [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Old Wolf Guest Posted: Thu Dec 11, 2003 11:42 am    Post subject: Re: a[i] not equivalent to *(a+i)? Quote: The expression '&a[100]' yields undefined behavior, because the sub-expression 'a[100]' would access a non-existing element. To answer the question from your subject for built-in arrays 'a[n]': - 'a[i]' is equivalent to '*(a + i)' (and this is also equivalent to '*(i + a)' which in turn is equivalent to 'i[a]'; but this is a completely irrelevant sidetrack...): both yield a reference to an element for '0 <= i && i < n' and are invalid otherwise. - '&a[i]' is equivalent to '(a + i)' if 'i != n'. In case 'i == n', the expression '(a+i)' is valid, while '&a[i]' is not. If 'i < 0 || n <= i', both expressions are invalid while for '0 <= i && i < n' both expressions are valid. Often, the case 'a + n' is important: it is used to obtain a "past the end"-iterator for an array object. Does any of this apply to: extern char a[100]; extern char a[]; /* does that make &a[0] invalid? */ extern char *a; or even #include int put101(char *a) { char *p = a + 101; /* char *p = &a[101]; -- better? worse? identical? */ while (*a && a < p) putchar(*a++); putchar('n'); } int main() { char a[100] = "Hello, world!"; put101(a); } People (in C anyway) add and subtract pointers all the time. It would horrify me to learn that this program is undefined behaviour. Until now I have used the &a[i] syntax instead of (a+i) in situations like this (I find it more readable). Is this a bad habit? [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Display posts from previous: All Posts1 Day7 Days2 Weeks1 Month3 Months6 Months1 Year Oldest FirstNewest First         C++Talk.NET Forum Index -> C++ Language (Moderated) All times are GMT Goto page 1, 2, 3, 4, 5  Next Page 1 of 5   Jump to: Select a forum C/C++----------------C LanguageC Language (Moderated)C++ Language (Moderated)C++ language (comp.lang.c++)C++ language, library and standards Non-English C/C++ Newsgroups----------------C++ (German)C++ (French)  You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum Powered by phpBB © 2001, 2006 phpBB Group SEO toolkit © 2004-2006 webmedic.