fgets() vs std::getline() performance
in [C++]
|
From: Earl Purple on 28 Sep 2006 08:58 kanze wrote: > > There is no way to access the string's already-allocated buffer > from outside the string, so this would require getline to be a > friend of std::string. I rather doubt that it is in most > implementations. (A quick grep on the preprocessor output of a > program which included <string> showed no friend in the g++ > implementation.) It wouldn't need to write to it directly, it would call one of the string methods insert() or append() that takes a const char * and a length - you'd want to pass the length so string won't need to look for the null character. > > Besides that, I would assume that both methods would write to > > a buffer before writing directly to the string / char-array. > > I presume you are talking about the buffering done in filebuf. For ostream. And for FILE * it also uses a buffer. > > I would hope that it reads ahead to find the terminating '\n' > > character before checking the allocation length of the > > std::string. > > I'm not sure how you propose to implement this. getline doesn't > have access to the internals of filebuf, any more than it does > to the internals of std::string. And the buffer in filebuf has > a fixed length anyway, so there's no guarantee that you'd find > the '\n' in it. No and I don't know the methods of streambuf etc but I do know char_traits<char> has a find method which returns a null pointer if the character is not found. If there is no way for ostream to do this there is something lacking, if not in the standard then in this particular implementation (as the implementer of fstream can surely use any specific additions underneath that they have added to its streambuf). -- [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ]
From: kanze on 29 Sep 2006 08:42 Earl Purple wrote: > kanze wrote: > > There is no way to access the string's already-allocated > > buffer from outside the string, so this would require > > getline to be a friend of std::string. I rather doubt that > > it is in most implementations. (A quick grep on the > > preprocessor output of a program which included <string> > > showed no friend in the g++ implementation.) > It wouldn't need to write to it directly, it would call one of > the string methods insert() or append() that takes a const > char * and a length - you'd want to pass the length so string > won't need to look for the null character. Yes. I think I was thinking more in terms of stringbuf. Of course, accessing the already allocated buffer only works if there is one; the user must have called string::reserve before. > > > Besides that, I would assume that both methods would write > > > to a buffer before writing directly to the string / > > > char-array. > > I presume you are talking about the buffering done in filebuf. > For ostream. And for FILE * it also uses a buffer. > > > I would hope that it reads ahead to find the terminating '\n' > > > character before checking the allocation length of the > > > std::string. > > I'm not sure how you propose to implement this. getline > > doesn't have access to the internals of filebuf, any more > > than it does to the internals of std::string. And the > > buffer in filebuf has a fixed length anyway, so there's no > > guarantee that you'd find the '\n' in it. > No and I don't know the methods of streambuf etc but I do know > char_traits<char> has a find method which returns a null > pointer if the character is not found. If there is no way for > ostream to do this there is something lacking, if not in the > standard then in this particular implementation (as the > implementer of fstream can surely use any specific additions > underneath that they have added to its streambuf). There's no way for an external class to access the streambuf buffer. A derived class can do so, but all of the functions which do so directly are protected, so the istream and getline can't. -- James Kanze GABI Software Conseils en informatique orient?e objet/ Beratung in objektorientierter Datenverarbeitung 9 place S?mard, 78210 St.-Cyr-l'?cole, 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! ]
From: Earl Purple on 29 Sep 2006 11:45 kanze wrote: > There's no way for an external class to access the streambuf > buffer. A derived class can do so, but all of the functions > which do so directly are protected, so the istream and getline > can't. It is true that there is no standard "find" method in streambuf, but as it is the same entity that implements streambuf as the one who implements getline, i.e. the writer of this STL, they are free to put in hidden extras for their own use. Thus they can add in some kind of find method and allow getline to access it so that it doesn't have to read from the streambuf one character at a time until the terminator is found. Instead it could call the internal find method and then do one big copy, assuming that to be more efficient. -- [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ]
From: Alex Vinokur on 23 Oct 2006 01:09 "crhras" <crhras(a)sbcglobal.net> wrote in message news:TkoOg.1329$e66.993(a)newssvr13.news.prodigy.com... > > Wow ! I just used two different file IO methods and the performance > difference was huge. Is there something that I am doing wrong? or is > fgets() just that much faster than getline()? > > Here's the code I used : > > // SLOOOOOOW > // ----------------- > std::string line; > std::ifstream in(filename.c_str()); > > while (std::getline(in, line,'\n')) > { > } > > // FAAAAAASSSSSTTTTT > // --------------------------- > FILE * fp; > fp = fopen(filename.c_str(), "r"); > > while (fgets(line, 512, fp) != NULL) > { > } There are a lot of methods of copying files. Various test cases can be seen at: http://groups.google.com/group/sources/browse_frm/thread/64e45316dde35021 Results of comparative performance measurement can be seen here: http://groups.google.com/group/log-files/browse_frm/thread/508ee2a154042a0c -- Alex Vinokur email: alex DOT vinokur AT gmail DOT com http://mathforum.org/library/view/10978.html http://sourceforge.net/users/alexvn [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ]
First
|
Prev
|
Pages: 1 2 3 4 5 6 Prev: Sanity check: public/private Next: Base Class undefined but header right above it! |