| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11
Comments:
<0> i just want a template cl***<T> that will insert some values into a hash_map<T,T>
<1> There's no T2 declared in line 24. Where do you expect it to come from?
<2> oh, i'm guessing it's all one file
<1> The error message looks pretty obvious to me.
<3> actually, I guess it's a one line file, but it demonstrates the problem
<2> why can't you just use a hash_map<T,T> ?
<0> _m_ sorry... in the prevous code i used T.. that not the error.
<2> it likely has some function to insert a value
<0> with hash_map<T,T> the error continues...
<1> the compiler doesn't see any hash function for std::string.
<2> i don't see the error message
<0> ../MySite.cpp:28: instantiated from `void C<T>::insert(T, T) [with T = std::string]'
<0> ../MySite.cpp:38: instantiated from here
<2> forget about C
<2> what's the problem with using a hash_map ?
<3> they're not standard.....yet
<1> And they're not portable, either.
<0> _m_ maybe thats the problem.. with int it compiles..
<1> s/maybe//
<1> add this to your code when using gcc: namespace __gnu_cxx { template<> struct hash<std::string> { size_t operator()(std::string const &s) const { return hash<char const *>()(s.c_str()); } }; }
<1> You should make sure to #include <string> when you use std::string, btw.
<0> _m_ many thanks!!! i will do that!
<0> ehhehe no errors... you are the man!
<3> _m_ you got a minute to test something?
<1> sure
<3> #include "boost/filesystem/path.hpp"
<3> put that in a c++ source file and try to compile with all the "conformance" options turned on
<3> I'm getting 5 warnings
<3> "level 1 warnings.... the ones that you can turn off only if you turn them all off
<1> I don't get any with g++ -W -Wall -Werror -ansi -pedantic
<3> c:\boost\include\boost-1_34\boost\filesystem\path.hpp(281) : warning C4180: qualifier applied to function type has no meaning; ignored
<3> ok, thanks
<3> it might be a compiler bug <sigh>
<1> Even no warning with -Weffc++
<3> well it's odd because line 281 is essentially the same as line 280 which does NOT get the warning
<3> it's bitching about some const on some of the methods
<1> hmm ... are talking about CVS head?
<1> I tested against 1.33.1
<3> yup
<3> it's after 1.33.1
<3> I even know what revision starts causing the "problem"
<3> revision 1.16
<3> date: 2005/12/16 16:40:34; author: beman_dawes; state: Exp; lines: +1278 -93
<3> merge from i18n branch - at last!
<1> +1278 ... holy ****
<3> yeah
<3> I noticed that
<3> I hope I don't need that many if/when I do some i18n
<3> I'm trying to figure out why I didn't notice it before yesterday
<1> no warnings for the new code, either.
<3> ok, must be vc8
<3> const string_type & string() const { return m_path; }
<3> const string_type file_string() const;
<3> const string_type directory_string() const { return file_string(); }
<3> there are lines 280, 281, 282
<3> gets warnings on 81,82
<1> The warnings are valid. The const on the return type is odd.
<1> I think these two consts make some sense, in this special case, though. The make the interface of file_string() and directory_string() consistent with string().
<1> lovely
<1> I own the channel now.
<0> heeheh
Return to
#c++
or
Go to some related
logs:
#AllNiteCafe
#linux
#AllNiteCafe
Shitzo dogs
#php
activewidget tabs size
#skype
#c++
#linux
sun v40z iommu
|
|