www.digitalmars.com         C & C++   DMDScript  

c++ - Using directive still does not work?

reply Wu Yongwei <adah netstd.com> writes:
I see that with stlport now I can use "using std::cout;", "using std::endl;",
and such like.  But "using namespace std;" still fails.  I vaguely remember it
has been a problem for quite some time.  Anyone knows when this problem could be
solved?

Best regards,

Wu Yongwei
Feb 28 2003
next sibling parent reply "Nic Tiger" <nictiger progtech.ru> writes:
I had this problem too, but it went away when I downloaded and used the
latest compiler beta.
Also I used the latest stlport.

Nic Tiger.

"Wu Yongwei" <adah netstd.com> сообщил/сообщила в новостях следующее:
news:b3n7ck$2f45$1 digitaldaemon.com...

 I see that with stlport now I can use "using std::cout;", "using


std::endl;",

 and such like.  But "using namespace std;" still fails.  I vaguely


remember it

 has been a problem for quite some time.  Anyone knows when this problem


could be

 solved?

 Best regards,

 Wu Yongwei
Feb 28 2003
parent reply Hans Wennborg <hans hanshq.net> writes:
It would be great if this problem could be solved without having to use 
the stlport, since it's only awailable for 32-bit compiling.


Hans

Nic Tiger wrote:

 I had this problem too, but it went away when I downloaded and used the
 latest compiler beta.
 Also I used the latest stlport.
 
 Nic Tiger.
 
 "Wu Yongwei" <adah netstd.com> сообщил/сообщила в новостях следующее:
 news:b3n7ck$2f45$1 digitaldaemon.com...
 

I see that with stlport now I can use "using std::cout;", "using


 
 std::endl;",
 

and such like.  But "using namespace std;" still fails.  I vaguely


 
 remember it
 

has been a problem for quite some time.  Anyone knows when this problem


 
 could be
 

solved?

Best regards,

Wu Yongwei
Mar 01 2003
parent Wu Yongwei <adah netstd.com> writes:
STLport could be used with other memory models.  Try specifying
"/D_STLP_NO_OWN_IOSTREAMS" on the command line.

Wu Yongwei

In article <b3qoop$26db$1 digitaldaemon.com>, Hans Wennborg says...

It would be great if this problem could be solved without having to use 
the stlport, since it's only awailable for 32-bit compiling.


Hans
Mar 02 2003
prev sibling parent reply "Jim Jennings" <jwjenn mindspring.com> writes:
"Wu Yongwei" <adah netstd.com> wrote in message
news:b3n7ck$2f45$1 digitaldaemon.com...

 I see that with stlport now I can use "using std::cout;", "using std::endl;",
 and such like.  But "using namespace std;" still fails.  I vaguely remember it
 has been a problem for quite some time.  Anyone knows when this problem could
be
 solved?

 Best regards,

 Wu Yongwei


    Are you sure about that? Check it out. I have version 8.32.17n, which I
downloaded on Feb. 20, 2003. I believe it is the latest
version.
    "using namespace std;" and "using std::cout" work, but "using std::endl;"
does not. I.e., you can use it, but you will get a
compile error: "Error: template endl<> is not instantiated", unless you also
say something like:

    cout << "Hello World" << std::endl;

but then it would be unnecessary to say "using std::endl;"  To save a lot of
frustration use: '\n' unless you have a special reason
for using "endl".

    Scott Meyers wrote an article for C++ Report, December, 1995, "The little
endl that couldn't" where he advised against the
gratuitous use of endl. You can find the article at:
    http://www.aristeia.com/publications_frames.html

Jim Jennings
Mar 03 2003
parent reply Richard Grant <fractal clark.net> writes:
In article <b3vrm9$1n7b$1 digitaldaemon.com>, Jim Jennings says...



    Scott Meyers wrote an article for C++ Report, December, 1995, "The little
endl that couldn't" where he advised against the
gratuitous use of endl. You can find the article at:
    http://www.aristeia.com/publications_frames.html


Not that I'm defending Scott, but he concludes that there is no real basis to
generally avoid using std::endl with the standard out streams
(w)cout/(w)cerr/(w)clog.

The problem with std:endl and "using std::endl" is a bug that is under review,
and while '\n' is a reasonable fix, it does not help in the instance where the
stream needs to be flushed. Any templatized use of the applicator/manipulator
pattern combined with the using declaration will fail. This includes std::endl
and std::flush.

In the rare instances that using std::endl appear, I tend to do:

#if defined (BROKEN_USING_DECLARATION)

#else
using std::endl;
#endif

Richard



 Mar 03 2003



  parent reply "Jim Jennings" <jwjenn mindspring.com>  writes:


"Richard Grant" <fractal clark.net> wrote in message
news:b3vuso$1pf6$1 digitaldaemon.com...

 In article <b3vrm9$1n7b$1 digitaldaemon.com>, Jim Jennings says...



    Scott Meyers wrote an article for C++ Report, December, 1995, "The little
endl that couldn't" where he advised against the
gratuitous use of endl. You can find the article at:
    http://www.aristeia.com/publications_frames.html


 Not that I'm defending Scott, but he concludes that there is no real basis to
 generally avoid using std::endl with the standard out streams
 (w)cout/(w)cerr/(w)clog.


He said that there is no reason to use it either -- since the standard calls
for unbuffered cout, except that it is easier to type
than '\n'.  He is against endl because "it is one of my pet peeves". But as you
say there is no reason to avoid it anymore.
Providing compilers comply with standard. That's why he took it out of 
"Effective C++".


 The problem with std:endl and "using std::endl" is a bug that is under review,
 and while '\n' is a reasonable fix, it does not help in the instance where the
 stream needs to be flushed. Any templatized use of the applicator/manipulator
 pattern combined with the using declaration will fail. This includes std::endl
 and std::flush.

 In the rare instances that using std::endl appear, I tend to do:

 #if defined (BROKEN_USING_DECLARATION)

 #else
 using std::endl;
 #endif


Preprocessors, and compiler flags are not my strong point. How do I (or can I)
enter #define BROKEN_USING_DECLARATION from the
command line? I have your "#if defined ...."  in a test source file, and have
tried the flag:
       -D #define BROKEN_USING_DECLARATION
with and without the parentheses when compiling. I get no complaint, but it
doesn't take, either. I guess my question is: how do I
turn the define on for dmc, and off for another compiler? Edit the header file
every time? I am using dmc almost exclusively now,
but just in case. And where can one fine some examples of how the compiler
flags work? I cannot figure them out.
Jim J



 Mar 03 2003



 next sibling parent reply "Rajiv Bhagwat" <dataflow vsnl.com>  writes:


The way to specify command line defines is:
-DBROKEN_USING_DECLARATION
and if you need to specify a value, it is:
-DBROKEN_USING_DECLARATION=1

This should be there in the documentation, somewhere.
- Rajiv

"Jim Jennings" <jwjenn mindspring.com> wrote in message
news:b415lj$2is1$1 digitaldaemon.com...

 "Richard Grant" <fractal clark.net> wrote in message


news:b3vuso$1pf6$1 digitaldaemon.com...

 In article <b3vrm9$1n7b$1 digitaldaemon.com>, Jim Jennings says...



    Scott Meyers wrote an article for C++ Report, December, 1995, "The






little endl that couldn't" where he advised against the

gratuitous use of endl. You can find the article at:
    http://www.aristeia.com/publications_frames.html


 Not that I'm defending Scott, but he concludes that there is no real




basis to

 generally avoid using std::endl with the standard out streams
 (w)cout/(w)cerr/(w)clog.


 He said that there is no reason to use it either -- since the standard


calls for unbuffered cout, except that it is easier to type

 than '\n'.  He is against endl because "it is one of my pet peeves". But


as you say there is no reason to avoid it anymore.

 Providing compilers comply with standard. That's why he took it out of


"Effective C++".

 The problem with std:endl and "using std::endl" is a bug that is under




review,

 and while '\n' is a reasonable fix, it does not help in the instance




where the

 stream needs to be flushed. Any templatized use of the




applicator/manipulator

 pattern combined with the using declaration will fail. This includes




std::endl

 and std::flush.

 In the rare instances that using std::endl appear, I tend to do:

 #if defined (BROKEN_USING_DECLARATION)

 #else
 using std::endl;
 #endif


 Preprocessors, and compiler flags are not my strong point. How do I (or


can I) enter #define BROKEN_USING_DECLARATION from the

 command line? I have your "#if defined ...."  in a test source file, and


have tried the flag:

        -D #define BROKEN_USING_DECLARATION
 with and without the parentheses when compiling. I get no complaint, but


it doesn't take, either. I guess my question is: how do I

 turn the define on for dmc, and off for another compiler? Edit the header


file every time? I am using dmc almost exclusively now,

 but just in case. And where can one fine some examples of how the compiler


flags work? I cannot figure them out.

 Jim J





 Mar 04 2003



  parent reply "Jim Jennings" <jwjenn mindspring.com>  writes:


"Rajiv Bhagwat" <dataflow vsnl.com> wrote in message
news:b41qf3$2u16$1 digitaldaemon.com...

 The way to specify command line defines is:
 -DBROKEN_USING_DECLARATION
 and if you need to specify a value, it is:
 -DBROKEN_USING_DECLARATION=1

 This should be there in the documentation, somewhere.
 - Rajiv


Rajiv,
OK, it works. I would swear that I tried that, and it didn't work. I must be
spooked.
Yes, it is in the documentation. But usually I have only a vague idea what the
exact form should be.
I.e.,   -Dmacro[=text]  what does that mean? Should I enter:
        -Dmacro[=BROKEN_USING_DECLARATION] ?  I know that is wrong.
       -D[BROKEN_USING_DECLARATION] ?  well, I know what [ . . . ] means.
       -D=BROKEN_USING_DECLARATION ?
       -D=#define BROKEN_USING_DECLARATION ?
        -DBROKEN_USING_DECLARATION =true (or on, yes, uh-huh!)
        -DDominus vobiscum.
And I can think of several more permutations and combinations (but never the
right one)  Is the first commandment
of Unix documentation "Thou shalt never give an example!"? I hate to keep
asking these neophyte questions.
Jim J



 Mar 04 2003



 next sibling parent  "Walter" <walter digitalmars.com>  writes:


"Jim Jennings" <jwjenn mindspring.com> wrote in message
news:b42jq2$ann$1 digitaldaemon.com...

 "Rajiv Bhagwat" <dataflow vsnl.com> wrote in message


news:b41qf3$2u16$1 digitaldaemon.com...

 The way to specify command line defines is:
 -DBROKEN_USING_DECLARATION
 and if you need to specify a value, it is:
 -DBROKEN_USING_DECLARATION=1

 This should be there in the documentation, somewhere.
 - Rajiv


 Rajiv,
 OK, it works. I would swear that I tried that, and it didn't work. I must


be spooked.

 Yes, it is in the documentation. But usually I have only a vague idea what


the exact form should be.

 I.e.,   -Dmacro[=text]  what does that mean? Should I enter:
         -Dmacro[=BROKEN_USING_DECLARATION] ?  I know that is wrong.
        -D[BROKEN_USING_DECLARATION] ?  well, I know what [ . . . ] means.
        -D=BROKEN_USING_DECLARATION ?
        -D=#define BROKEN_USING_DECLARATION ?
         -DBROKEN_USING_DECLARATION =true (or on, yes, uh-huh!)
         -DDominus vobiscum.


-D means:
    #define DEBUG 1
-Dmacro means:
    #define macro 1
-Dmacro=text means:
    #define macro text



 Mar 04 2003



prev sibling  parent  "Jim Jennings" <jwjenn mindspring.com>  writes:


Pardon the rant. I learned something today. Thank you all very much.
JIm

"Jim Jennings" <jwjenn mindspring.com> wrote in message
news:b42jq2$ann$1 digitaldaemon.com...

 "Rajiv Bhagwat" <dataflow vsnl.com> wrote in message
news:b41qf3$2u16$1 digitaldaemon.com...

 The way to specify command line defines is:
 -DBROKEN_USING_DECLARATION
 and if you need to specify a value, it is:
 -DBROKEN_USING_DECLARATION=1

 This should be there in the documentation, somewhere.
 - Rajiv


 Rajiv,
 OK, it works. I would swear that I tried that, and it didn't work. I must be
spooked.
 Yes, it is in the documentation. But usually I have only a vague idea what the
exact form should be.
 I.e.,   -Dmacro[=text]  what does that mean? Should I enter:
         -Dmacro[=BROKEN_USING_DECLARATION] ?  I know that is wrong.
        -D[BROKEN_USING_DECLARATION] ?  well, I know what [ . . . ] means.
        -D=BROKEN_USING_DECLARATION ?
        -D=#define BROKEN_USING_DECLARATION ?
         -DBROKEN_USING_DECLARATION =true (or on, yes, uh-huh!)
         -DDominus vobiscum.
 And I can think of several more permutations and combinations (but never the
right one)  Is the first commandment
 of Unix documentation "Thou shalt never give an example!"? I hate to keep
asking these neophyte questions.
 Jim J





 Mar 04 2003



prev sibling  parent reply Richard Grant <fractal clark.net>  writes:


In article <b415lj$2is1$1 digitaldaemon.com>, Jim Jennings says...


 I) enter #define BROKEN_USING_DECLARATION from the


#include <iostream>

using std::cout;
using std::endl;

int main() {
cout << endl;
}

Works fine with latest beta..

Richard



 Mar 10 2003



  parent reply "Jim Jennings" <jwjenn mindspring.com>  writes:


Richard,
Just now saw your post.
I downloaded STLport-4.5-0119. I do not see a makefile for Digital Mars,
however. I have read the Install instructions quickly, and it seems
complicated to me. Can you give me some instructions? Shall I query
Christof? Or just wait until a set of DM instructions are published? I can
survive the "endl" bug for awhile longer.
Jim J.

"Richard Grant" <fractal clark.net> wrote in message
news:b4iprg$bg7$1 digitaldaemon.com...

 In article <b415lj$2is1$1 digitaldaemon.com>, Jim Jennings says...


 I) enter #define BROKEN_USING_DECLARATION from the


 #include <iostream>

 using std::cout;
 using std::endl;

 int main() {
 cout << endl;
 }

 Works fine with latest beta..

 Richard





 Mar 12 2003



  parent reply Richard Grant <fractal clark.net>  writes:


In article <b4oog5$2p11$1 digitaldaemon.com>, Jim Jennings says...

Richard,
Just now saw your post.
I downloaded STLport-4.5-0119. I do not see a makefile for Digital Mars,


uh.. sorry. I mean the latest DM beta.

Richard



 Mar 13 2003



  parent  "Jim Jennings" <jwjenn mindspring.com>  writes:


Richard,
     OK, I got the beta, and using std::endl; works. I had downloaded the
zip some time ago, but did not know what to do with it. I did not realize
that scppn.exe was the compiler. After your message, the dawn broke.
     I'm way down on the learning curve, but I really like the compiler. It
is very fast, with smaller .exe files.
     Thanks very much for the tip.
Jim

"Richard Grant" <fractal clark.net> wrote in message
news:b4pf8q$7if$1 digitaldaemon.com...

 In article <b4oog5$2p11$1 digitaldaemon.com>, Jim Jennings says...

Richard,
Just now saw your post.
I downloaded STLport-4.5-0119. I do not see a makefile for Digital Mars,


 uh.. sorry. I mean the latest DM beta.

 Richard





 Mar 13 2003