C++Talk.NET C++ language newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  synthesized operators in C++ - a good thing? Goto page 1, 2  Next          C++Talk.NET Forum Index -> C++ Language (Moderated) View previous topic :: View next topic   Author Message Gene Bushuyev Guest Posted: Thu Nov 03, 2005 1:04 pm    Post subject: synthesized operators in C++ - a good thing? MS is adding sinthesized operators to Visual C++ (CLI?). Below is the exceprt from the language specification I stumbled upon: -------------------- The compound assignment operators (+=, -=, *=, /=, %=, >>=, <<=, ^=, &=, and Quote: =) are synthesized from other operators. For the expression x @= y (where @ denotes one of the operators listed above): If lookup for operator@= succeeds, the rules specified so far are applied. Otherwise, the expression x @= y is rewritten as x = x @ y, and the transformed expression is interpreted with the rules specified so far. Identify when synthesis would and would not occur. [[#56]] If no overload for operator@= applies after overload resolution or synthesis, the program is ill-formed. Synthesis shall not occur for operators defined inside native classes. --------------------- I hope that the last line means that normal C++ classes will not be affected by this mis-feature. I think it rarely (if ever) makes sense to define @= operators in terms of @ operators. The opposite can be the case. If MS adds those rules into their C++ compiler, there will be a need for another rule in the coding standards: beware operator synthesis. Am I overreacting? Can it be a good feature after all? -- Gene Bushuyev ~ Cadence Design Systems [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Valentin.Samko Guest Posted: Thu Nov 03, 2005 7:14 pm    Post subject: Re: synthesized operators in C++ - a good thing? Quote: I hope that the last line means that normal C++ classes will not be affected by this mis-feature. I think it rarely (if ever) makes sense to define @= operators in terms of @ operators. This does not make much sense to me either. The opposite would probably make sense (defining @ in terms of @+) but I still find this confusing. Also, this feature will probably lead to more problems when developers overuse operators. Quote: The opposite can be the case. If MS adds those rules into their C++ compiler, there will be a need for another rule in the coding standards: beware operator synthesis. Yes, as they may lead to performance problems when someone forgets to implement the correct operator (or even worse, if the correct operator is implemented, but is not found when the name lookup is performed). I would rather prefer to get a compile time error in such a case. Think of MyStringClass '+=' operator implemented in terms of the '+' operator. Quote: Am I overreacting? Can it be a good feature after all? Well, may be they may be usefull for "immutable" types, where you never modify your object (i.e. you throw away the old value set the new one). -- Valentin Samko - http://www.valentinsamko.com [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Gabriel Dos Reis Guest Posted: Thu Nov 03, 2005 7:34 pm    Post subject: Re: synthesized operators in C++ - a good thing? "Gene Bushuyev" <blackhole (AT) spamguard (DOT) com> writes: [...] Quote: I hope that the last line means that normal C++ classes will not be affected by this mis-feature. I think it rarely (if ever) makes sense to define @= operators in terms of @ operators. There are few cases like adding or multiplying polynomials (or like) where it makes sense. But as a whole, I believe it is a misfeature if done mechanically with no user control/input. -- Gabriel Dos Reis [email]gdr (AT) integrable-solutions (DOT) net[/email] [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Mirek Fidler Guest Posted: Thu Nov 03, 2005 7:35 pm    Post subject: Re: synthesized operators in C++ - a good thing? Gene Bushuyev wrote: Quote: MS is adding sinthesized operators to Visual C++ (CLI?). Below is the exceprt from the language specification I stumbled upon: -------------------- The compound assignment operators (+=, -=, *=, /=, %=, >>=, <<=, ^=, &=, and |=) are synthesized from other operators. For the expression x @= y (where @ denotes one of the operators listed above): If lookup for operator@= succeeds, the rules specified so far are applied. Otherwise, the expression x @= y is rewritten as x = x @ y, and the transformed expression is interpreted with the rules specified so far. Identify when synthesis would and would not occur. [[#56]] If no overload for operator@= applies after overload resolution or synthesis, the program is ill-formed. Synthesis shall not occur for operators defined inside native classes. --------------------- I hope that the last line means that normal C++ classes will not be affected by this mis-feature. I think it rarely (if ever) makes sense to define @= operators in terms of @ operators. The opposite can be the case. If MS adds those rules into their C++ compiler, there will be a need for another rule in the coding standards: beware operator synthesis. Am I overreacting? Can it be a good feature after all? No. It is horrible. The reason not to buy the next MS compiler. Mirek [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Tony Delroy Guest Posted: Thu Nov 03, 2005 7:36 pm    Post subject: Re: synthesized operators in C++ - a good thing? I think there's some sense behind this, but share your concerns, and think any facility should be explicitly controlled. Plenty of concise but readable notations could be envisaged, such as: auto X::operator+=, -=, *=, /=, %=; Tony [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top kanze Guest Posted: Fri Nov 04, 2005 11:11 am    Post subject: Re: synthesized operators in C++ - a good thing? Gene Bushuyev wrote: Quote: MS is adding sinthesized operators to Visual C++ (CLI?). Below is the exceprt from the language specification I stumbled upon: -------------------- The compound assignment operators (+=, -=, *=, /=, %=, >>=, =, ^=, &=, and |=) are synthesized from other operators. For the expression x @= y (where @ denotes one of the operators listed above): If lookup for operator@= succeeds, the rules specified so far are applied. Otherwise, the expression x @= y is rewritten as x = x @ y, and the transformed expression is interpreted with the rules specified so far. Identify when synthesis would and would not occur. [[#56]] If no overload for operator@= applies after overload resolution or synthesis, the program is ill-formed. Synthesis shall not occur for operators defined inside native classes. --------------------- I hope that the last line means that normal C++ classes will not be affected by this mis-feature. I think it rarely (if ever) makes sense to define @= operators in terms of @ operators. The opposite can be the case. If MS adds those rules into their C++ compiler, there will be a need for another rule in the coding standards: beware operator synthesis. Am I overreacting? Can it be a good feature after all? I don't see any real reason to call it a feature -- I don't think I've ever had a case where @= was derived from @, so its usefulness is extremely limited. On the other hand, I don't see any reason to be worried about it: IMHO, any code which defines and operator @ and does not define @= is already broken, at least at the design level. And if I understand correctly, if there is a @=, the compiler uses it, and doesn't synthesize. So the total effect of the "feature" is 0 (except maybe that it took developer time which could have been spent on something userful) -- it has absolutly no effect in reasonably designed and written code. -- James Kanze GABI Software Conseils en informatique orientée objet/ Beratung in objektorientierter Datenverarbeitung 9 place Sémard, 78210 St.-Cyr-l'École, France, +33 (0)1 30 23 00 34 [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Daniel Krügler Guest Posted: Sat Nov 05, 2005 11:38 am    Post subject: Re: synthesized operators in C++ - a good thing? kanze wrote: Quote: I don't see any real reason to call it a feature -- I don't think I've ever had a case where @= was derived from @, so its usefulness is extremely limited. The only real-world example for defining @= in terms of @ I'm aware of is operator overloading of enumeration types. In this case there is no alternative (I think) to first define @ using a cast to a reasonable integral type (e.g. the unsigned form of the promotion type) and after that defining @= in terms of the first. Returning to the OP's question: I can't see convincing reasons for the decision of this kind of autogenerated operators. Explicit compiler advice to do so as in Glassborow's proposal, e.g. http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2004/n1702.pdf http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2004/n1717.pdf makes much more sense. Greetings from Bremen, Daniel [ 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: Sat Nov 05, 2005 2:29 pm    Post subject: Re: synthesized operators in C++ - a good thing? In article <Rudaf.7607$BZ5.41 (AT) newssvr13 (DOT) news.prodigy.com>, Gene Bushuyev <blackhole (AT) spamguard (DOT) com> writes Quote: I hope that the last line means that normal C++ classes will not be affected by this mis-feature. I think it rarely (if ever) makes sense to define @= operators in terms of @ operators. The opposite can be the case. If MS adds those rules into their C++ compiler, there will be a need for another rule in the coding standards: beware operator synthesis. Am I overreacting? Can it be a good feature after all? It takes some companies a long time to learn the lesson that generated code causes problems (if we do not want it we have to find a way to turn it off) As most experienced C++ programmers know, the @= operators are actually more primitive than the corresponding @ ones. It is easy to automate generation of @ from @= and so, in general, I would define the @= and delegate to that to implement @. It seems to me that MS is often motivated by making things easier for poor quality programmers that make up a proportion (probably substantial) of its customer base. I find it difficult to see how a well designed user type could benefit from this miss-feature. However let me turn my attention to possible problems. Consider: mytype & operator+(int, mytype const &); How do they propose to synthesize an operator += from that? Does the return type have significance? The order of the parameters? Or what? The trouble is that once a compiler provides such a 'feature' it will be almost impossible to remove it without howls of agony from some customers. -- Francis Glassborow ACCU Author of 'You Can Do It!' see http://www.spellen.org/youcandoit For project ideas and contributions: http://www.spellen.org/ youcandoit/projects [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top kwikius Guest Posted: Sat Nov 05, 2005 6:37 pm    Post subject: Re: synthesized operators in C++ - a good thing? kanze wrote: Quote: and operator @ and does not define @= is already broken, at least at the design level. FWIW In a physical-quantities library, where a and b are dimensioned-quantities, a * b is a valid operation , but a *= b violates dimensional analysis rules. see http://www.servocomm.freeserve.co.uk/Cpp/physical_quantity/index.html regards Andy Little [ 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: Sat Nov 05, 2005 6:38 pm    Post subject: Re: synthesized operators in C++ - a good thing? In article <436b5029$0$303$4d4ebb8e (AT) businessnews (DOT) de.uu.net>, Daniel Krügler <dsp (AT) bdal (DOT) de> writes Quote: The only real-world example for defining @= in terms of @ I'm aware of is operator overloading of enumeration types. What is wrong with: enum type { /* enumerator list */ }; type& operator +=(type & t, int i){ return t = type (t + i); } Note that the + is just a builtin operator+(int, int). Yes you could define operator+(type const &, int i) first if you want but there is no need to. -- Francis Glassborow ACCU Author of 'You Can Do It!' see http://www.spellen.org/youcandoit For project ideas and contributions: http://www.spellen.org/youcandoit/projects [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Mirek Fidler Guest Posted: Sun Nov 06, 2005 12:11 pm    Post subject: Re: synthesized operators in C++ - a good thing? Quote: any reason to be worried about it: IMHO, any code which defines and operator @ and does not define @= is already broken, at least at the design level. Like streams and operator<< ? :) Mirek [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top jimmaureenrogers@worldnet Guest Posted: Mon Nov 07, 2005 9:34 am    Post subject: Re: synthesized operators in C++ - a good thing? Francis Glassborow wrote: Quote: In article <436b5029$0$303$4d4ebb8e (AT) businessnews (DOT) de.uu.net>, Daniel Krügler <dsp (AT) bdal (DOT) de> writes The only real-world example for defining @= in terms of @ I'm aware of is operator overloading of enumeration types. What is wrong with: enum type { /* enumerator list */ }; type& operator +=(type & t, int i){ return t = type (t + i); } Note that the + is just a builtin operator+(int, int). One problem I see with your example is that it does not account for enum overflow. Any concept of enum arithmetic should be modular. Jim Rogers [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top James Kanze Guest Posted: Mon Nov 07, 2005 10:45 am    Post subject: Re: synthesized operators in C++ - a good thing? Mirek Fidler wrote: Quote: any reason to be worried about it: IMHO, any code which defines and operator @ and does not define @= is already broken, at least at the design level. Like streams and operator<< ? Touché. Ostream::<< has some pretty special semantics, which means that even for a user defined class, I will define it. And no <<&, obviously. On the other hand, this is a very potent argument against the Microsoft extension. I certainly don't what a mess of <<= operators for ostream appearing out of nowhere. -- James Kanze mailto: [email]james.kanze (AT) free (DOT) fr[/email] Conseils en informatique orientée objet/ Beratung in objektorientierter Datenverarbeitung 9 pl. Pierre Sémard, 78210 St.-Cyr-l'École, France +33 (0)1 30 23 00 34 [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Daniel Krügler Guest Posted: Mon Nov 07, 2005 10:51 am    Post subject: Re: synthesized operators in C++ - a good thing? Francis Glassborow wrote: Quote: In article <436b5029$0$303$4d4ebb8e (AT) businessnews (DOT) de.uu.net>, Daniel Krügler <dsp (AT) bdal (DOT) de> writes The only real-world example for defining @= in terms of @ I'm aware of is operator overloading of enumeration types. What is wrong with: enum type { /* enumerator list */ }; type& operator +=(type & t, int i){ return t = type (t + i); } Note that the + is just a builtin operator+(int, int). Yes you could define operator+(type const &, int i) first if you want but there is no need to. While this is an alternative way to realize @= of an enumeration type, its essentially **does** implement @= in terms of an @ operation, although it is that of an integral type - I can't count that as a "pure" @= implementation ;-) By the way: In your example I would have expected something like type& operator +=(type & t, type i) { return t = type (t + int(i)); } but I assume you wanted just to demonstrate the basic idea? Greetings from Bremen, Daniel [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top kanze Guest Posted: Mon Nov 07, 2005 10:54 am    Post subject: Re: synthesized operators in C++ - a good thing? Francis Glassborow wrote: Quote: In article <436b5029$0$303$4d4ebb8e (AT) businessnews (DOT) de.uu.net>, Daniel Krügler <dsp (AT) bdal (DOT) de> writes The only real-world example for defining @= in terms of @ I'm aware of is operator overloading of enumeration types. What is wrong with: Nothing in itself but... Quote: enum type { /* enumerator list */ }; type& operator +=(type & t, int i){ return t = type (t + i); } Note that the + is just a builtin operator+(int, int). Yes you could define operator+(type const &, int i) first if you want but there is no need to. All of the coding guidelines I have seen insist that if @= is present, so is @ (and vice versa -- curiously, none of the coding guidelines I've seen made an exception for the << and >> of iostream, although it never occured to anyone to apply the guideline to them). Code which defines +=, but not +, on a type, is generally considered broken, or at least poorly defined. Given that, Daniel is right in that for enum types, it makes more sense to define the | and & operators first, and |= and &= in terms of them. -- James Kanze GABI Software Conseils en informatique orientée objet/ Beratung in objektorientierter Datenverarbeitung 9 place Sémard, 78210 St.-Cyr-l'École, France, +33 (0)1 30 23 00 34 [ 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  Next Page 1 of 2   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.