hocker@enuxha.eas.asu.edu (Charles C. Hocker) (11/14/90)
Hello,
I am having touble trying to created an array of objects of
class pnt using TC++. My code is as follows:
class pnt {
public:
int X, Y, Z;
pnt (int x = 0, int y = 0, int z = 0) {X=x; Y=y; Z=z;}
};
class pnts {
public:
int Count;
pnt Pnts [MAX_PNTS]; // pnt from above
pnts (int count = 0) {Count = count;}
};
int main (void)
{
pnts Points;
// ...
}
I recieve an error to the effect "cannot find pnt::pnt()" when I compile my
program. I have tried to use several different methods for getting the
code segment to work and have found not that will work. The ARM gives a
reference on page 289, section 12.6.1, but I still recieve the error
about "pnt::pnt() not found" when I use:
pnt Pnts [MAX_PNTS] = {pnt ()}
Although the ARM talks of initializing the array of objects, it talks
about not being able to initialize an array of objects with constructors.
Sincerly
Charles C. Hocker
hocker@enuxha.eas.asu.edurmartin@clear.com (Bob Martin) (11/15/90)
In article <1756@enuxha.eas.asu.edu> hocker@enuxha.eas.asu.edu (Charles C. Hocker) writes: > >Hello, > > I am having touble trying to created an array of objects of >class pnt using TC++. My code is as follows: > >class pnt { > public: > int X, Y, Z; > pnt (int x = 0, int y = 0, int z = 0) {X=x; Y=y; Z=z;} >}; > >class pnts { > public: > int Count; > pnt Pnts [MAX_PNTS]; // pnt from above > pnts (int count = 0) {Count = count;} >}; > >int main (void) >{ > pnts Points; >// ... > >} > >I recieve an error to the effect "cannot find pnt::pnt()" when >I compile my program. I'll take a stab at answering this. My manual (which happens to be the version of the ARM published by SUN) says in 12.1 that: "A default constructor for a class X is a constructor of the form X::X(). A default constructor will not be generated for a class X if any constructor has been declared for class X. Now, since you declared pnt::pnt(x,y,z). The compiler did _NOT_ generate the default constructor pnt::pnt(). Yet it needs to use the pnt::pnt() constructor when constructing the array in the pnts class. Note that the array cannot be constructed with pnt(x,y,z) since the compiler has no way of knowing what values should be assigned to x, y or z. So the solution is for you to declare and implement a pnt::pnt() constructor. Perhaps it should set the coordinates to zero... -- +-Robert C. Martin-----+:RRR:::CCC:M:::::M:| Nobody is responsible for | | rmartin@clear.com |:R::R:C::::M:M:M:M:| my words but me. I want | | uunet!clrcom!rmartin |:RRR::C::::M::M::M:| all the credit, and all | +----------------------+:R::R::CCC:M:::::M:| the blame. So there. |
miron@fornax.UUCP (Miron Cuperman) (11/24/90)
rmartin@clear.com (Bob Martin) writes: >>class pnt {... >> pnt (int x = 0, int y = 0, int z = 0) {X=x; Y=y; Z=z;} >>}; >>I recieve an error to the effect "cannot find pnt::pnt()" when >>I compile my program. >So the solution is for you to declare and implement a pnt::pnt() >constructor. Perhaps it should set the coordinates to zero... No. pnt::pnt() is defined, since all the arguments to pnt(int, int, int) are optional. It may be that there is a bug in your compiler, or that it is not up to date. If not you define pnt::pnt(), then it would be ambigous which version of pnt::pnt() to call.
mlm@cs.brown.edu (Moises Lejter) (11/26/90)
About the two constructors:
pnt::pnt( int x = 0, int y = 0, int y = 0 ) {};
pnt::pnt() {};
Under AT&T C++ 2.0, these were in fact two different constructors, and
in order to initialize an array of pnt's it was necessary to include
the second one. Under AT&T C++ 2.1, that restriction was lifted, and
the first constructor will be used to initialize arrays of pnt's, if
the second constructor is not supplied. Under either compiler, I
believe that including both versions of the constructor and then
defining a variable
pnt P;
will result in a compile-time error on that line, as the compiler will
be unable to disambiguate that call.
Moises
--
-----------------------------------------------------------------------------
Internet/CSnet: mlm@cs.brown.edu BITNET: mlm@browncs.BITNET
UUCP: ...!uunet!brunix!mlm Phone: (401)863-7664
USmail: Moises Lejter, Box 1910 Brown University, Providence RI 02912dcurtis@crc.ac.uk (Dr. David Curtis) (11/28/90)
In article <MLM.90Nov25233119@hector.cs.brown.edu> mlm@cs.brown.edu writes: > >About the two constructors: > pnt::pnt( int x = 0, int y = 0, int y = 0 ) {}; > pnt::pnt() {}; >Under AT&T C++ 2.0, these were in fact two different constructors, and >in order to initialize an array of pnt's it was necessary to include >the second one. Under AT&T C++ 2.1, that restriction was lifted, and >the first constructor will be used to initialize arrays of pnt's, if >the second constructor is not supplied. Under either compiler, I >believe that including both versions of the constructor and then >defining a variable > pnt P; >will result in a compile-time error on that line, as the compiler will >be unable to disambiguate that call. > > Moises Easy to get round this, just take out the first default argument to the first constructor: pnt::pnt( int x, int y = 0, int z = 0 ) { pnt_x=x; pnt_y=y; pnt_z=z; } pnt::pnt() { pnt_x=0; pnt_y=0; pnt_z=0; } Dave Curtis Academic Department of Psychiatry, Janet: dc@UK.AC.UCL.SM.PSYCH Middlesex Hospital, Elsewhere: dc@PSYCH.SM.UCL.AC.UK Mortimer Street, London W1N 8AA. EARN/Bitnet: dc%PSYCH.SM.UCL@UKACRL Tel 081-380 9475 Fax 081-323 1459 Usenet: ...!mcsun!ukc!mrccrc!D.Curtis
jimad@microsoft.UUCP (Jim ADCOCK) (12/04/90)
In article <1990Nov15.050358.16043@clear.com> rmartin@clear.com (Bob Martin) writes: |I'll take a stab at answering this. My manual (which happens to be |the version of the ARM published by SUN) says in 12.1 that: | | "A default constructor for a class X is a constructor of the | form X::X(). A default constructor will not be generated for | a class X if any constructor has been declared for class X. ACK! *My* manual [which happens to *the* version of ARM, published by Addison-Wesley in 1990 ISBN 0-201-51459-1 in an ugly reddish-brown cover] says: "A default constructor for a class X is a constructor of class X that can be called without an argument. [A default constructor is typically of the form X::X(), but a constructor that can be called with no arguments because it has default arguments, X::X(int=0), for example, is also a default constructor.]" My copy of the ANSI-C++ committee working paper X3J16/90-0091 dated Oct 23 1990 speaks similarly. Please do not quote an individual vendor's reference manuals as being "the" facts on C++. At best these manuals will tend to lag a year or so behind the standardizations effort. At worse, they may contain gratuitous incompatibilities and/or extensions. Please quote either directly from Ellis & Stroustrup's ARM [the reddish-brown book from Addison-Wesley ISBN 0-201-51459-1] or from the ANSI-C++ committee's working papers. By so quoting from the language standards, we put pressure on vendors to make their compilers conform to the standard, rather than allow arguments to develop as to who's compiler is "right." [There is no ARM except *the* ARM. Get it. :-]
steve@taumet.com (Stephen Clamage) (12/05/90)
jimad@microsoft.UUCP (Jim ADCOCK) writes: >Please quote either directly from Ellis & Stroustrup's ARM [the reddish-brown >book from Addison-Wesley ISBN 0-201-51459-1] or from the ANSI-C++ committee's >working papers. Please do not quote from the X3J16 C++ committee working papers, either. They carry no official weight, and are subject to change without notice. They reflect items under discussion by the committee, not a language standard. (A working paper might be just the opinions of a single person.) -- Steve Clamage, TauMetric Corp, steve@taumet.com
rmartin@clear.com (Bob Martin) (12/07/90)
In article <59511@microsoft.UUCP> jimad@microsoft.UUCP (Jim ADCOCK) writes: >In article <1990Nov15.050358.16043@clear.com> rmartin@clear.com (Bob Martin) writes: >|I'll take a stab at answering this. My manual (which happens to be >|the version of the ARM published by SUN) says in 12.1 that: >| >| "A default constructor for a class X is a constructor of the >| form X::X(). A default constructor will not be generated for >| a class X if any constructor has been declared for class X. > >ACK! *My* manual [...] *the* version of ARM says: > "A default constructor for a class X is a constructor of class > X that can be called without an argument. > >Please do not quote an individual vendor's reference manuals as being >"the" facts on C++. [There is no ARM except *the* ARM. Get it. :-] Sorry, point taken. *the* ARM says in 12.1: A default constructor is typically of the form X::X() [...] A default contructor will be generated for a class X only if no constructor has been declared for class X. -- +-Robert C. Martin-----+:RRR:::CCC:M:::::M:| Nobody is responsible for | | rmartin@clear.com |:R::R:C::::M:M:M:M:| my words but me. I want | | uunet!clrcom!rmartin |:RRR::C::::M::M::M:| all the credit, and all | +----------------------+:R::R::CCC:M:::::M:| the blame. So there. |
jimad@microsoft.UUCP (Jim ADCOCK) (12/13/90)
In article <536@taumet.com> steve@taumet.com (Stephen Clamage) writes: |jimad@microsoft.UUCP (Jim ADCOCK) writes: | |>Please quote either directly from Ellis & Stroustrup's ARM [the reddish-brown |>book from Addison-Wesley ISBN 0-201-51459-1] or from the ANSI-C++ committee's |>working papers. | |Please do not quote from the X3J16 C++ committee working papers, either. |They carry no official weight, and are subject to change without notice. |They reflect items under discussion by the committee, not a language |standard. (A working paper might be just the opinions of a single person.) Well, nothing has "official" weight right now. The ARM seems to be holding up pretty well, but for example, I'd hate for people to say that template and exceptions *aren't* in the language, because the ARM calls them experimental. I'd much rather people try to track the developments of the ANSI-C++ committee. For now, I think that either the ARM or the ANSI-C++ pronouncements [the working papers, for lack of anything better] are pretty darned good discriptions of what the language "is." And I'd hate to C++ to degenerate into multiple dialects, either because people feel there is no "official" description of the language, or because they're working from old references. The ANSI-C effort showed that nothing "is" until its over, but good efforts towards standardization were made before the standard became "official." I claim most of the benefits of standardization come from trying to reach a consensus, not from having made that consensus.