C++Talk.NET C++ language newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  Portable types Goto page 1, 2  Next          C++Talk.NET Forum Index -> C++ Language (Moderated) View previous topic :: View next topic   Author Message Murali Guest Posted: Sun Oct 02, 2005 1:30 am    Post subject: Portable types Hi, Given that in C++ the sizes of fundamental types are implementation defined (subject to some restrictions), programmers who want their code to be portable normally define their own typedefs to obtain types of required size (for eg., Int8, UInt8, Int16 etc). My question is, why can't this mechanism be made a part of some standard header so that a common vocabulary can be used. For eg., we can have predefined types int8_t, uint8_t, int16_t, int32_t etc. Is this already available in some form? Was this considered during the standardization process? Thanks. Regards, Murali [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top albrecht.fritzsche Guest Posted: Sun Oct 02, 2005 4:12 pm    Post subject: Re: Portable types Murali wrote: Quote: Hi, Given that in C++ the sizes of fundamental types are implementation defined (subject to some restrictions), programmers who want their code to be portable normally define their own typedefs to obtain types of required size (for eg., Int8, UInt8, Int16 etc). Here the problem already starts - what the language guarantees is that sizeof(char) == 1 and 1 <= sizeof(the other types) but to the best of my knowledge that's it (apart from size_t being sufficiently large). So, in your naming, eg Int8, what does the 8 stands for? I guess it should mean the number of bits. But this is already contrary to sizeof(), which returns you the number of bytes since there is the possibility that on some machines 1 byte = 7 bits or some- thing completely different (even though I've never met such a machine). Quote: My question is, why can't this mechanism be made a part of some standard header so that a common vocabulary can be used. For eg., we can have predefined types int8_t, uint8_t, int16_t, int32_t etc. I guess that several compiler builders already offer you such types, so, if you know the range of your targetting OSs you might use those. But, in the end, using your own typedefs is the safest way - then targetting to a new OS can start as easily as adding the correct typedefs. Ali [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Stanimir Kabaivanov Guest Posted: Sun Oct 02, 2005 4:17 pm    Post subject: Re: Portable types Hi Murali, Quote: Given that in C++ the sizes of fundamental types are implementation defined (subject to some restrictions), programmers who want their code to be portable normally define their own typedefs to obtain types of required size (for eg., Int8, UInt8, Int16 etc). My question is, why can't this mechanism be made a part of some standard header so that a common vocabulary can be used. For eg., we can have predefined types int8_t, uint8_t, int16_t, int32_t etc. Is this already available in some form? Was this considered during the standardization process? Well, according to the C99 standard, all vendors that claim to have C99 compliant tools (compilers) should provide stdint.h header file, that defines exactly the types you mention. You can check your compiler standard headers and figure out if its there. On my system it looks like: /* 7.18.1.1 Exact-width integer types */ typedef signed char int8_t; typedef unsigned char uint8_t; typedef short int16_t; typedef unsigned short uint16_t; typedef int int32_t; typedef unsigned uint32_t; typedef long long int64_t; typedef unsigned long long uint64_t; 7.18.1.1 (http://dev.unicals.com/papers/c99-draft.html#7.18.1.1) is what you should be looking for to read the exact text. Besides there's a good article by Michael Barr that explains reasoning behind this decision (and drawbacks too ): http://www.embedded.com/shared/printableArticle.jhtml?articleID=17300092 HTH, Stanimir Kabaivanov [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Thomas Maeder Guest Posted: Sun Oct 02, 2005 4:18 pm    Post subject: Re: Portable types "Murali" <murdesi (AT) yahoo (DOT) com> writes: Quote: Given that in C++ the sizes of fundamental types are implementation defined (subject to some restrictions), programmers who want their code to be portable normally define their own typedefs to obtain types of required size (for eg., Int8, UInt8, Int16 etc). My question is, why can't this mechanism be made a part of some standard header so that a common vocabulary can be used. For eg., we can have predefined types int8_t, uint8_t, int16_t, int32_t etc. Is this already available in some form? Was this considered during the standardization process? Not exactly Standard, but very portable: http://www.boost.org/libs/integer/cstdint.htm [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Jeff Schwab Guest Posted: Sun Oct 02, 2005 4:18 pm    Post subject: Re: Portable types Murali wrote: Quote: Hi, Given that in C++ the sizes of fundamental types are implementation defined (subject to some restrictions), programmers who want their code to be portable normally define their own typedefs to obtain types of required size (for eg., Int8, UInt8, Int16 etc). My question is, why can't this mechanism be made a part of some standard header so that a common vocabulary can be used. For eg., we can have predefined types int8_t, uint8_t, int16_t, int32_t etc. Is this already available in some form? Was this considered during the standardization process? It's standard in C, and available in C++. http://boost.org/libs/integer/cstdint.htm [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Carl Barron Guest Posted: Sun Oct 02, 2005 4:19 pm    Post subject: Re: Portable types Murali <murdesi (AT) yahoo (DOT) com> wrote: Quote: Hi, Given that in C++ the sizes of fundamental types are implementation defined (subject to some restrictions), programmers who want their code to be portable normally define their own typedefs to obtain types of required size (for eg., Int8, UInt8, Int16 etc). My question is, why can't this mechanism be made a part of some standard header so that a common vocabulary can be used. For eg., we can have predefined types int8_t, uint8_t, int16_t, int32_t etc. Is this already available in some form? Was this considered during the standardization process? Thanks. c99 provides stdint.h. and there is motion if not already approved for a cstdint in c++0x. see section 7.18 of the 1999 C standard for a definition of the exact contents of stdint.h. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Ivan Vecerina Guest Posted: Sun Oct 02, 2005 4:20 pm    Post subject: Re: Portable types "Murali" <murdesi (AT) yahoo (DOT) com> wrote Quote: Hi, Given that in C++ the sizes of fundamental types are implementation defined (subject to some restrictions), programmers who want their code to be portable normally define their own typedefs to obtain types of required size (for eg., Int8, UInt8, Int16 etc). My question is, why can't this mechanism be made a part of some standard header so that a common vocabulary can be used. For eg., we can have predefined types int8_t, uint8_t, int16_t, int32_t etc. Is this already available in some form? Was this considered during the standardization process? Look for <stdint.h> in the C99 standard. [draft available here: http://www.vmunix.com/~gabor/c/draft.html] I haven't kept track of it, but this is likely to be incorporated in some form into the next C++ standard. Cheers, Ivan -- http://ivan.vecerina.com/contact/?subject=NG_POST <- email contact form Brainbench MVP for C++ <> http://www.brainbench.com [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Greg Herlihy Guest Posted: Sun Oct 02, 2005 4:21 pm    Post subject: Re: Portable types Murali wrote: Quote: Hi, Given that in C++ the sizes of fundamental types are implementation defined (subject to some restrictions), programmers who want their code to be portable normally define their own typedefs to obtain types of required size (for eg., Int8, UInt8, Int16 etc). My question is, why can't this mechanism be made a part of some standard header so that a common vocabulary can be used. For eg., we can have predefined types int8_t, uint8_t, int16_t, int32_t etc. Is this already available in some form? Was this considered during the standardization process? Thanks. Regards, Murali You should be able to find the "exact width" integer types in stdint.h. The stdint.h header is part of the C99 standard which many C++ compilers support. I would hazard a guess that stdint.h will become part of the C++ standard at some point; in the meantime I can see no reason not to include it. Greg [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Jack Klein Guest Posted: Sun Oct 02, 2005 4:21 pm    Post subject: Re: Portable types On 1 Oct 2005 21:30:30 -0400, "Murali" <murdesi (AT) yahoo (DOT) com> wrote in comp.lang.c++.moderated: Quote: Hi, Given that in C++ the sizes of fundamental types are implementation defined (subject to some restrictions), programmers who want their code to be portable normally define their own typedefs to obtain types of required size (for eg., Int8, UInt8, Int16 etc). My question is, why can't this mechanism be made a part of some standard header so that a common vocabulary can be used. For eg., we can have predefined types int8_t, uint8_t, int16_t, int32_t etc. Is this already available in some form? Was this considered during the standardization process? Thanks. Regards, Murali Sounds like you want the standard C header <stdint.h>, added to the C standard in 1999, to be added to the C++ standard. And it almost certainly will be, as <cstdint>, some day. I would imagine that almost all C++ compiler ship with a <stdint.h> right now, in fact. With the possible exception of the integer types with at least 64 bits, you can make a workable <stdint.h> header for an conforming C89 or C++ compiler yourself. But note that the exact width integer types may not be present. They are required if and only if an implementation has integer data types with exactly that width, no padding bits, and using a 2's complement representation for negative values in the signed versions. This is actually a very real issue, even if most desktop programmers don't realize it. DSPs (Digital Signal Processors) are becoming more widely used all the time, and the newer ones have C and C++ compilers. I am working right now a TI2812 DSP, which comes with a freestanding C++ compiler, although I am programming it in C. You can't have a <stdint.h> header that defines int8_t and uint8_t because no such type exists on this architecture. It accesses memory in 16 bit chunks, and the char, short, and int types are all 16 bits. Likewise there are DSPs where all the integer types are 32 bits. For maximum portability, you can use the int_least8_t or int_fast8_t, but each of these will may have more than 8 bits on some implementations. -- Jack Klein Home: http://JK-Technology.Com FAQs for comp.lang.c http://www.eskimo.com/~scs/C-faq/top.html comp.lang.c++ http://www.parashift.com/c++-faq-lite/ alt.comp.lang.learn.c-c++ http://www.contrib.andrew.cmu.edu/~ajo/docs/FAQ-acllc.html [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Jack Klein Guest Posted: Sun Oct 02, 2005 8:23 pm    Post subject: Re: Portable types On 2 Oct 2005 12:12:31 -0400, "albrecht.fritzsche" <albrecht.fritzsche (AT) arcor (DOT) de> wrote in comp.lang.c++.moderated: Quote: Murali wrote: Hi, Given that in C++ the sizes of fundamental types are implementation defined (subject to some restrictions), programmers who want their code to be portable normally define their own typedefs to obtain types of required size (for eg., Int8, UInt8, Int16 etc). Here the problem already starts - what the language guarantees is that sizeof(char) == 1 and 1 <= sizeof(the other types) but to the best of my knowledge that's it (apart from size_t being sufficiently large). So, in your naming, eg Int8, what does the 8 stands for? I guess it should mean the number of bits. But this is already contrary to sizeof(), which returns you the number of bytes since there is the possibility that on some machines 1 byte = 7 bits or some- thing completely different (even though I've never met such a machine). [snip] This last bit is incorrect. C++ and C require that a byte, that is a char, contain at least 8 bits. It can contain more than 8, but never less. There are both C and C++ compilers for platforms, mostly DSPs (Digital Signal Processors) where a char contains 16 or 32 bits. -- Jack Klein Home: http://JK-Technology.Com FAQs for comp.lang.c http://www.eskimo.com/~scs/C-faq/top.html comp.lang.c++ http://www.parashift.com/c++-faq-lite/ alt.comp.lang.learn.c-c++ http://www.contrib.andrew.cmu.edu/~ajo/docs/FAQ-acllc.html [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Jack Klein Guest Posted: Sun Oct 02, 2005 8:25 pm    Post subject: Re: Portable types On 2 Oct 2005 12:17:51 -0400, Stanimir Kabaivanov <kabaivanov (AT) bulgaria (DOT) com> wrote in comp.lang.c++.moderated: Quote: Hi Murali, Given that in C++ the sizes of fundamental types are implementation defined (subject to some restrictions), programmers who want their code to be portable normally define their own typedefs to obtain types of required size (for eg., Int8, UInt8, Int16 etc). My question is, why can't this mechanism be made a part of some standard header so that a common vocabulary can be used. For eg., we can have predefined types int8_t, uint8_t, int16_t, int32_t etc. Is this already available in some form? Was this considered during the standardization process? Well, according to the C99 standard, all vendors that claim to have C99 compliant tools (compilers) should provide stdint.h header file, that defines exactly the types you mention. All C99 conforming compilers must indeed provide the stdint.h header. Many C and C++ compilers that do not claim C99 conformance provide it as an extension. But the header is not required to provide the types the OP asked about. Quote: You can check your compiler standard headers and figure out if its there. On my system it looks like: /* 7.18.1.1 Exact-width integer types */ typedef signed char int8_t; typedef unsigned char uint8_t; typedef short int16_t; typedef unsigned short uint16_t; typedef int int32_t; typedef unsigned uint32_t; typedef long long int64_t; typedef unsigned long long uint64_t; Paragraph 3 of "7.18.1.1 Exact-width integer types" states specifically: "These types are optional. However, if an implementation provides integer types with widths of 8, 16, 32, or 64 bits, it shall define the corresponding typedef names." There are platforms in existence today, with C and C++ compilers, that do not provide all of these types, because the processor hardware does not support them. This is most common with DSPs. All the other types provided in <stdint.h> are required, but the exact-width types are not, if they don't exist on the target processor. -- Jack Klein Home: http://JK-Technology.Com FAQs for comp.lang.c http://www.eskimo.com/~scs/C-faq/top.html comp.lang.c++ http://www.parashift.com/c++-faq-lite/ alt.comp.lang.learn.c-c++ http://www.contrib.andrew.cmu.edu/~ajo/docs/FAQ-acllc.html [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Krakers Guest Posted: Mon Oct 03, 2005 1:10 pm    Post subject: Re: Portable types "Murali" <murdesi (AT) yahoo (DOT) com> wrote... Quote: Given that in C++ the sizes of fundamental types are implementation defined (subject to some restrictions), programmers who want their code to be portable normally define their own typedefs to obtain types of required size (for eg., Int8, UInt8, Int16 etc). Exact sizes of integer types are neither required nor sufficient to write portable code. There are two fundamental issues with the size of integer variables and portability: 1. Avoiding overflow. That is a matter of minimal (and not exact) sizes and C/C++ standard guarantees those: char is at least 8, short 16 and long is at least 32 bits. Int may be used instead of short, as it will not be smaller. 2. Writing to a file. Here, a popular, but not portable way is to take an address of a variable and write sizeof variable bytes to a file. A portable way is to design the file to use either big-endian or little-endian convention and write (or read) a fixed number of bytes, combining them into a number. Exact sizes are not required in this approach - again the variable should not be smaller, but may be larger than the number of bytes it needs to hold. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top albrecht.fritzsche Guest Posted: Mon Oct 03, 2005 1:14 pm    Post subject: Re: Portable types Jack Klein wrote: Quote: On 2 Oct 2005 12:12:31 -0400, "albrecht.fritzsche" [email]albrecht.fritzsche (AT) arcor (DOT) de[/email]> wrote in comp.lang.c++.moderated: since there is the possibility that on some machines 1 byte = 7 bits or some- thing completely different (even though I've never met such a machine). This last bit is incorrect. C++ and C require that a byte, that is a char, contain at least 8 bits. Yes, I stand corrected - "A byte is at least large enough to contain any member of the basic execution character set and is composed of a contigous sequence of bits, the number of which is implementation- defined." (1.7.1) That's exactly what makes reading the Standard a (black) science in its own - why can't there be a note like: there are X members of that very character set and hence a byte requires at least Y bits. Cheers, Ali [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Carl Barron Guest Posted: Mon Oct 03, 2005 1:17 pm    Post subject: Re: Portable types Jack Klein <jackklein (AT) spamcop (DOT) net> wrote: Quote: Paragraph 3 of "7.18.1.1 Exact-width integer types" states specifically: "These types are optional. However, if an implementation provides integer types with widths of 8, 16, 32, or 64 bits, it shall define the corresponding typedef names." There are platforms in existence today, with C and C++ compilers, that do not provide all of these types, because the processor hardware does not support them. This is most common with DSPs. All the other types provided in <stdint.h> are required, but the exact-width types are not, if they don't exist on the target processor. At least if you use have stdint.h and use it, then any non-existant tyoe will be a compiler error and not some wild runtime bug, that only surfaces occasionally depending on the data. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Niklas Matthies Guest Posted: Tue Oct 04, 2005 1:43 pm    Post subject: Re: Portable types On 2005-10-03 13:14, albrecht.fritzsche wrote: Quote: Jack Klein wrote: On 2 Oct 2005 12:12:31 -0400, "albrecht.fritzsche" wrote: : since there is the possibility that on some machines 1 byte = 7 bits or some- thing completely different (even though I've never met such a machine). This last bit is incorrect. C++ and C require that a byte, that is a char, contain at least 8 bits. Yes, I stand corrected - "A byte is at least large enough to contain any member of the basic execution character set and is composed of a contigous sequence of bits, the number of which is implementation- defined." (1.7.1) That's exactly what makes reading the Standard a (black) science in its own - why can't there be a note like: there are X members of that very character set and hence a byte requires at least Y bits. Because that's not the reason. The basic execution character set has only 95 required characters (or 96 if you count the null character), which would fit into 7 bits. Instead it follows from the minimum value of CHAR_BIT. -- Niklas Matthies [ 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.