On 09/04/2007 09:47 AM, Andy Ross wrote:
> .... the patch ....
I assume we are still talking about:
http://sourceforge.net/tracker/index.php?func=detail&aid=1584727&group_id=382&atid=100382
> .... strchr()
> returns a pointer into the *same* memory it got. The constness needs
> to be synchronized between the pointers, which is outside the
> capabilities of the C language.
It is not outside the capabilities of the c++ language ... which is
the language of the patch. The capability is demonstrated in the
example below.
The whole line of argument is pointless anyway, because the constness
is a property of the pointer, not a property of the memory.
> So programmers have to choose between
> a slightly "unsafe" function that drops const and one that requires a
> cast to use with a non-const string.
No, in general they don't need to choose, as explained below.
================================
Returning to the higher-level discussion, it is not necessary to do a
strdup in this situation, as the following constructive suggestion
illustrates.
using namespace std;
#include <iostream>
const char* strchr(const char* foo){
cout << "const" << endl;
return foo+3;
}
char* strchr(char* foo){
cout << "not const, implemented in terms of: " ;
const char* bar(foo);
return foo + (strchr(bar) - foo);
}
int main(){
char aa[100];
const char* xx(aa);
const char* yy(aa);
yy = strchr(xx); // const
yy = strchr(aa); // non const
}
The output of this program is:
const
not const, implemented in terms of: const
Notes:
1) The trick of foo+(bar-foo) is a standard way of transferring information
from a const pointer (bar) to a non-const pointer (foo).
2) I chose to embody this trick in a wrapper function, partly to make
it easier to insert the explanatory messages. It would be well within
the reasonable envelope to put the foo+(bar-foo) trick inline, with no
wrapper.
==================================================================================
> Oh, good grief:
>
> $ echo 'void foo(const char* p){*(char*)p=0;}' > tmp.c
> $ gcc -S -c -o - tmp.c
Good grief yourself. Do you really think I don't know that ...
especially given that I explicitly mentioned it a few lines farther
down?
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Flightgear-devel mailing list
Flightgear-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/flightgear-devel
