Prev: C++0x variadic tuple data members
Next: Conversion to non-scalar type
From: itaj sherman on 3 Mar 2010 03:33
Is there a good reason the following code causes access violation in
MSVC6?
Using std::map::erase, like: aMap.erase( aMap.rbegin().base() );

#include <string>
#include <map>

typedef std::string cKey;
typedef std::string cElement;
typedef std::map< cKey, cElement > cMap;

int main(int argc, char* argv[])
{

cMap aMap;

aMap[ cKey("k1") ] = cElement("v1");

while( !aMap.empty() ) {
aMap.erase( aMap.rbegin().base() ); //this line causes access
violation in MSVC6
}

return 0;
}

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: Jared Wein on 3 Mar 2010 17:21
On Mar 3, 3:33 pm, itaj sherman <itajsher...(a)gmail.com> wrote:
> Is there a good reason the following code causes access violation in
> MSVC6?
> Using std::map::erase, like: aMap.erase( aMap.rbegin().base() );
>
> #include <string>
> #include <map>
>
> typedef std::string cKey;
> typedef std::string cElement;
> typedef std::map< cKey, cElement > cMap;
>
> int main(int argc, char* argv[])
> {
>
> cMap aMap;
>
> aMap[ cKey("k1") ] = cElement("v1");
>
> while( !aMap.empty() ) {
> aMap.erase( aMap.rbegin().base() ); //this line causes access
> violation in MSVC6
> }
>
> return 0;
>
> }
>
> --
> [ Seehttp://www.gotw.ca/resources/clcm.htmfor info about ]
> [ comp.lang.c++.moderated. First time posters: Do this! ]

Reproduced in VC10.

I believe this would happen because std::map::rbegin() will be
pointing to the position *after* the last element. If you switch the
code to use std::map::rend() then your problem will go away.

Another way to look at the problem is that if you try using
std::map::end() the same AV will happen. Swapping in std::map::begin()
will work though.

HTH
- jared


--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

From: itaj sherman on 4 Mar 2010 06:18
On Mar 4, 12:21 pm, Jared Wein <weinja...(a)gmail.com> wrote:
> On Mar 3, 3:33 pm, itaj sherman <itajsher...(a)gmail.com> wrote:
>
> > Is there a good reason the following code causes access violation in
> > MSVC6?
> > Using std::map::erase, like: aMap.erase( aMap.rbegin().base() );
>
> > #include <string>
> > #include <map>
>
> > typedef std::string cKey;
> > typedef std::string cElement;
> > typedef std::map< cKey, cElement > cMap;
>
> > int main(int argc, char* argv[])
> > {
>
> > cMap aMap;
>
> > aMap[ cKey("k1") ] = cElement("v1");
>
> > while( !aMap.empty() ) {
> > aMap.erase( aMap.rbegin().base() ); //this line causes access
> > violation in MSVC6
> > }
>
> > return 0;
>
> > }
>
> Reproduced in VC10.
>
> I believe this would happen because std::map::rbegin() will be
> pointing to the position *after* the last element. If you switch the
> code to use std::map::rend() then your problem will go away.
>
> Another way to look at the problem is that if you try using
> std::map::end() the same AV will happen. Swapping in std::map::begin()
> will work though.

{ edits: quoted block of blank lines, signature and two quoted banners removed.
please keep readers in mind when quoting. -mod }

Right, thanks, I should have done the following:

while( !aMap.empty() ) {
aMap.erase( (++aMap.rbegin()).base() );
}

I need to remove the last element in the map, not the first.
I had a few problems with VC6 lately, so I tend to blame it for
everything now.
Sadly management can't schedule upgrade now.

itaj


--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]

 | 
Pages: 1
Prev: C++0x variadic tuple data members
Next: Conversion to non-scalar type