C++Talk.NET C++ language newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  I long for an int Goto page 1, 2  Next          C++Talk.NET Forum Index -> C++ Language (Moderated) View previous topic :: View next topic   Author Message Guest Posted: Thu Jun 15, 2006 3:26 am    Post subject: I long for an int Here's my dilemma: class X { unsigned long file_size; // files can be big unsigned int record_size; // record's not that big unsigned int number_of_records; // not that many records .. }; Some code: number_of_records = files_size / record_size; This line gives me a warning on my 64-bit compiler - "possible loss of digits through conversion of long type into int type". Being somewhat anal about warnings, should I: 1. forget about it? There are more important things in life to worry about. 2. find some #pragma that will turn off the warning? 3. try to cast away the warning? 4. change the type of number_or_records to unsigned long? I grew up in the old days when memory wasn't cheap. And what if I had a bunch of unsigned int record_numbers that I had to keep track of? Thanks for your input. Joe [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Zara Guest Posted: Thu Jun 15, 2006 7:59 pm    Post subject: Re: I long for an int On 14 Jun 2006 18:26:17 -0400, joe.foxhound (AT) gmail (DOT) com wrote: Quote: Here's my dilemma: class X { unsigned long file_size; // files can be big unsigned int record_size; // record's not that big unsigned int number_of_records; // not that many records .. }; Some code: number_of_records = files_size / record_size; This line gives me a warning on my 64-bit compiler - "possible loss of digits through conversion of long type into int type". Being somewhat anal about warnings, should I: 1. forget about it? There are more important things in life to worry about. 2. find some #pragma that will turn off the warning? 3. try to cast away the warning? 4. change the type of number_or_records to unsigned long? I grew up in the old days when memory wasn't cheap. And what if I had a bunch of unsigned int record_numbers that I had to keep track of? You may cast away the warn, but test for the reason of the warning: #include <cassert> <...> unsigned long calculated_number_of_records = files_size / record_size; number_of_records = static_cast<unsigned int>( calculated_number_of_records ); assert(calculated_number_of_records == calculated_number_of_records); Regards, Zara [ 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: Thu Jun 15, 2006 8:01 pm    Post subject: Re: I long for an int joe.foxhound (AT) gmail (DOT) com wrote: Quote: Here's my dilemma: class X { unsigned long file_size; // files can be big unsigned int record_size; // record's not that big unsigned int number_of_records; // not that many records .. }; Some code: number_of_records = files_size / record_size; This line gives me a warning on my 64-bit compiler - "possible loss of digits through conversion of long type into int type". Being somewhat anal about warnings, should I: 1. forget about it? There are more important things in life to worry about. 2. find some #pragma that will turn off the warning? 3. try to cast away the warning? 4. change the type of number_or_records to unsigned long? I grew up in the old days when memory wasn't cheap. And what if I had a bunch of unsigned int record_numbers that I had to keep track of? Verify, then cast. Something like: unsigned long tmp = file_size / record_size ; if ( tmp > std::numeric_limits< unsigned int >::max() ) { // handle error... } else { number_of_records = static_cast< unsigned int >( tmp ) ; } Without the verification, I can't imagine the code ever passing code review, and an explicit conversion makes it a lot clearer as to what is going on. If you still get a warning, just ignore it. -- 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 ThosRTanner Guest Posted: Thu Jun 15, 2006 8:12 pm    Post subject: Re: I long for an int joe.foxhound (AT) gmail (DOT) com wrote: Quote: Here's my dilemma: class X { unsigned long file_size; // files can be big unsigned int record_size; // record's not that big unsigned int number_of_records; // not that many records .. }; Some code: number_of_records = files_size / record_size; This line gives me a warning on my 64-bit compiler - "possible loss of digits through conversion of long type into int type". Being somewhat anal about warnings, should I: 1. forget about it? There are more important things in life to worry about. Well, probably there are more important things, but: Consider your file might be 256Gb long, and contain 32 byte records. Then your number of records clearly doesn't fit into an unsigned int. So, no, don't forget about it. It's a potential issue. Quote: 2. find some #pragma that will turn off the warning? No, never do that. Quote: 3. try to cast away the warning? Well, that's a way of switching off warnings, certainly. But perhaps you should test that the number of records will fit into an unsigned int, die horribly if not, and then cast it. Quote: 4. change the type of number_or_records to unsigned long? I grew up in the old days when memory wasn't cheap. That's an option Quote: And what if I had a bunch of unsigned int record_numbers that I had to keep track of? Change them to unsigned long. Really you should have a type for your record numbers, so you can easily identify them. Possibly even a Boost::strong_typedef. Me, I'm of the school of thought that 'int fred' is 'magic typing' and about as unmaintanable as 'magic numbers' [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Jeffrey Schwab Guest Posted: Thu Jun 15, 2006 8:13 pm    Post subject: Re: I long for an int joe.foxhound (AT) gmail (DOT) com wrote: Quote: Here's my dilemma: class X { unsigned long file_size; // files can be big unsigned int record_size; // record's not that big unsigned int number_of_records; // not that many records .. }; Some code: number_of_records = files_size / record_size; This line gives me a warning on my 64-bit compiler - "possible loss of digits through conversion of long type into int type". Being somewhat anal about warnings, should I: 1. forget about it? There are more important things in life to worry about. 2. find some #pragma that will turn off the warning? 3. try to cast away the warning? 4. change the type of number_or_records to unsigned long? I grew up in the old days when memory wasn't cheap. And what if I had a bunch of unsigned int record_numbers that I had to keep track of? Are you sure the warning is not relevant? If your file_size requires sixty-four bits, and your record_size is "not that big," then couldn't file_size divided by record_size result in something requiring more than thirty-two bits? [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Rob Guest Posted: Thu Jun 15, 2006 8:13 pm    Post subject: Re: I long for an int joe.foxhound (AT) gmail (DOT) com wrote: Quote: Here's my dilemma: class X { unsigned long file_size; // files can be big unsigned int record_size; // record's not that big unsigned int number_of_records; // not that many records .. }; Some code: number_of_records = files_size / record_size; This line gives me a warning on my 64-bit compiler - "possible loss of digits through conversion of long type into int type". Being somewhat anal about warnings, should I: 1. forget about it? There are more important things in life to worry about. I don't recommend this option. Too many people make a habit of ignoring compiler warnings, and then get bitten by some behaviour that the compiler had warned them about. I therefore tend to advocating doing anything legitimate that will convince the compiler not to complain. And, by your analysis above, you know that the code is safe. So being nagged about it is counter-productive. Quote: 2. find some #pragma that will turn off the warning? I wouldn't do this either. The #pragma is compiler dependent (by definition) and you will have to do this with every compiler. It also has the problem that it will disable the warning in other code. And, unless you know that all other code that triggers the warning is actually safe, you are inviting a future problem. Quote: 3. try to cast away the warning? This is one option. Your problem is tailor made for a cast: the compiler is able to do the conversion (long to int) safely in this case. So use the cast to tell the compiler to shut up. More generally, the disadvantage of this is that it may be problematical when porting (eg if different compilers support longs and ints of different sizes from your current compiler). I would suggest that is unlikely in your particular case, as any value that can be stored in an (unsigned) int can be stored in an (unsigned) long. An advantage is that its effect is localised: you are not accidentally disabling the warning in any other code. And it is a technique that any programmer will understand. Quote: 4. change the type of number_or_records to unsigned long? I grew up in the old days when memory wasn't cheap. And what if I had a bunch of unsigned int record_numbers that I had to keep track of? While I don't believe in jumping through hoops to minimise memory usage (unless I have to), I don't believe in going out of my way to use more than I have to either. This is a valid option, apart from the problem of memory usage. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Milind Guest Posted: Fri Jun 16, 2006 1:29 am    Post subject: Re: I long for an int Hi Joe, What do you do if you have record_size = 1; // some really intresting record i donno how.. The point to be notated is that during this division that you are going to do, the quotient is going to be in the same range as that of the divident; which asks us to keep the quotient variable capable of holding that range. In which case option 4 would be your best bet. if you are really so sure that the number of records are going to be that large and that the record size is really not going to be that large, may be you should cast. But that would be a assumption that you would have to make based on the statistics that you have at your hand. But we have to remember that things might just not work right sometime down the line. For me options 1, 2, are a strict No. 3 is based on the data, but is very risky. 4. is the best option. Lemme know if you disagree. Regards, ~M joe.foxhound (AT) gmail (DOT) com wrote: Quote: Here's my dilemma: class X { unsigned long file_size; // files can be big unsigned int record_size; // record's not that big unsigned int number_of_records; // not that many records .. }; Some code: number_of_records = files_size / record_size; This line gives me a warning on my 64-bit compiler - "possible loss of digits through conversion of long type into int type". Being somewhat anal about warnings, should I: 1. forget about it? There are more important things in life to worry about. 2. find some #pragma that will turn off the warning? 3. try to cast away the warning? 4. change the type of number_or_records to unsigned long? I grew up in the old days when memory wasn't cheap. And what if I had a bunch of unsigned int record_numbers that I had to keep track of? Thanks for your input. Joe [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Guest Posted: Fri Jun 16, 2006 1:31 am    Post subject: Re: I long for an int joe.foxhound (AT) gmail (DOT) com wrote: Quote: 1. forget about it? There are more important things in life to worry about. 2. find some #pragma that will turn off the warning? 3. try to cast away the warning? 4. change the type of number_or_records to unsigned long? I grew up in the old days when memory wasn't cheap. And what if I had a bunch of unsigned int record_numbers that I had to keep track of? I prefer 4. If the system on which your program runs has different size between int and long, it is quite possible that int is too small for number_or_records. Ingoring this might result in disaster in the future. By the way, an extra four bytes is acceptable nowadays. On the other hand, if long and int has the same size, using long instead of int still has no problem. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Oleksandr Yefremov Guest Posted: Fri Jun 16, 2006 1:35 am    Post subject: Re: I long for an int I can recomend you two possibilities: first is to use long for both record_size and number_of_records and second if you are shure that you will fit in 32 bit with your variables: long long_number_of_records = files_size / record_size; if (long_number_of_records < MAX_INT) number_of_records = (int)long_number_of_records; else { // process special case (should never happens) throw unsupported_number_of_records(); } joe.foxhound (AT) gmail (DOT) com wrote: Quote: Here's my dilemma: class X { unsigned long file_size; // files can be big unsigned int record_size; // record's not that big unsigned int number_of_records; // not that many records .. }; Some code: number_of_records = files_size / record_size; This line gives me a warning on my 64-bit compiler - "possible loss of digits through conversion of long type into int type". Being somewhat anal about warnings, should I: 1. forget about it? There are more important things in life to worry about. 2. find some #pragma that will turn off the warning? 3. try to cast away the warning? 4. change the type of number_or_records to unsigned long? I grew up in the old days when memory wasn't cheap. And what if I had a bunch of unsigned int record_numbers that I had to keep track of? Thanks for your input. Joe [ 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: Fri Jun 16, 2006 1:37 am    Post subject: Re: I long for an int Victor Bazarov wrote: Quote: It all comes down to what your model (world) is like. From the C++ point of view, when you say 'int', you really mean "no more than 32767" because that's the lower limit. There is no such "lower limit" for int in C++. int may have the same size as char. -- 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 Robert Seacord Guest Posted: Fri Jun 16, 2006 2:19 pm    Post subject: Re: I long for an int Joe, I would make all of these variables size_t. It is possible that the sizeof(unsigned long) == unsigned(unsigned int) and both are too small to represent the size of a large file. If you really have to do the downcast I would add some to the effect that: unsigned long temp = file_size / record_size; if ( temp > UINT_MAX) { // throw exception } else { number_of_records = temp; } Also you have "file_size" in one location and "file_sizes" in the other. I'm assuming this was a typo. You may also want to check on how really big your file sizes can get. You may need to use a long long int type which will most likely be 64 bits in length. rCs {Please would posters learn how to respond in context rather than defaulting to top posting. -mod/fg} Quote: Here's my dilemma: class X { unsigned long file_size; // files can be big unsigned int record_size; // record's not that big unsigned int number_of_records; // not that many records .. }; Some code: number_of_records = files_size / record_size; This line gives me a warning on my 64-bit compiler - "possible loss of digits through conversion of long type into int type". Being somewhat anal about warnings, should I: 1. forget about it? There are more important things in life to worry about. 2. find some #pragma that will turn off the warning? 3. try to cast away the warning? 4. change the type of number_or_records to unsigned long? I grew up in the old days when memory wasn't cheap. And what if I had a bunch of unsigned int record_numbers that I had to keep track of? {quoted banner and signature removed mod/fg} [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Alf P. Steinbach Guest Posted: Fri Jun 16, 2006 2:20 pm    Post subject: Re: I long for an int * Valentin Samko: Quote: Victor Bazarov wrote: It all comes down to what your model (world) is like. From the C++ point of view, when you say 'int', you really mean "no more than 32767" because that's the lower limit. There is no such "lower limit" for int in C++. There is. See the FAQ item "What are the criteria for choosing between short / int / long data types?", currently at <url: http://www.parashift.com/c++-faq-lite/newbie.html#faq-29.5>. Not mentioned in that FAQ item: these lower limits stem from C++ being compatible with C, which imposes these limits (I'd rather the C++ standard explicitly stated these limits, but unfortunately it doesn't). Quote: int may have the same size as char. Yes, because there is no upper limit: a char may very well be 16 bits or more, depending on the system. -- A: Because it messes up the order in which people normally read text. Q: Why is it such a bad thing? A: Top-posting. Q: What is the most annoying thing on usenet and in e-mail? [ 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: Fri Jun 16, 2006 2:21 pm    Post subject: Re: I long for an int On 15 Jun 2006 16:37:54 -0400, Valentin Samko <c++.moderated (AT) digiways (DOT) com> wrote in comp.lang.c++.moderated: Quote: Victor Bazarov wrote: It all comes down to what your model (world) is like. From the C++ point of view, when you say 'int', you really mean "no more than 32767" because that's the lower limit. There is no such "lower limit" for int in C++. int may have the same size as char. There is of course a lower limit for the size of an int in C++, that is 1 byte. This is only possible on a platform where the value of the macro CHAR_BIT is greater than or equal to 16. There are such platforms, most of the recent ones are Digital Signal Processors (DSPs). But there is indeed a lower limit on the range of values that an int must be able to hold. These are available in <limits> and also as in C in <limits.h> or <climits>. These are the limits that Victor is referring to. He said nothing about size, just about the minimum positive value that a signed int must be able to contain. The value of INT_MAX must be greater than or equal to 32767. The value of INT_MIN must be less than or equal to -32767. -- Jack Klein Home: http://JK-Technology.Com FAQs for comp.lang.c http://c-faq.com/ 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 Victor Bazarov Guest Posted: Fri Jun 16, 2006 2:21 pm    Post subject: Re: I long for an int Valentin Samko wrote: Quote: Victor Bazarov wrote: It all comes down to what your model (world) is like. From the C++ point of view, when you say 'int', you really mean "no more than 32767" because that's the lower limit. There is no such "lower limit" for int in C++. int may have the same size as char. I am not sure I understand to what in my statement you object. The requirement in the language comes from C Standard, which specifies that the range of a value of type 'int' is _at_least_ -32767..32767. What it has to do with 'char', I am not sure. Yes, 'int' and 'char' can be of the same size. That's absolutely acceptable, and only means that 'char' has a wider range than required by the C Standard, not that 'int' has a narrower range. V -- Please remove capital 'A's when replying by e-mail I do not respond to top-posted replies, please don't ask [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Rob Guest Posted: Fri Jun 16, 2006 2:22 pm    Post subject: Re: I long for an int Valentin Samko wrote: Quote: Victor Bazarov wrote: It all comes down to what your model (world) is like. From the C++ point of view, when you say 'int', you really mean "no more than 32767" because that's the lower limit. There is no such "lower limit" for int in C++. int may have the same size as char. Incorrect. The C++ standard includes references back to the C standard, which says (among other things) that; 1) INT_MAX is an implementation defined value that represents the maximum value that can be stored in an int. 2) The implementation defined value of INT_MAX shall be equal or greater than 32767 [ 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.