public inbox for gcc-prs@sourceware.org
help / color / mirror / Atom feed
* libstdc++/10794: std::map's erase method invalidates reverse_iterators
@ 2003-05-14 23:46 gandy
  0 siblings, 0 replies; 2+ messages in thread
From: gandy @ 2003-05-14 23:46 UTC (permalink / raw)
  To: gcc-gnats


>Number:         10794
>Category:       libstdc++
>Synopsis:       std::map's erase method invalidates reverse_iterators
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed May 14 23:46:01 UTC 2003
>Closed-Date:
>Last-Modified:
>Originator:     gandy@abacho.de
>Release:        libstdc++5-3.3
>Organization:
>Environment:
debian linux:
ii  libstdc++5                                     3.3-0pre9                                     
ii  libstdc++5-3.3-dev                             3.3-0pre9                                      
ii  libstdc++5-3.3-doc                             3.3-0pre9                                      
ii  libstdc++5-dev                                 3.2.3-2                                        
ii  g++                                            3.2.3-1                                        
ii  g++-3.2                                        3.2.3-2                                        
ii  g++-3.3                                        3.3-0pre9                                     
>Description:
Programm output (g++-3.2 and g++-3.3):
-90
270
-90
-90

Expected output:
-90
-90
-90
-90

Note that the reverse_iterator jt is NOT pointing to the element which is erased.

Citation from SGI STL Docu:
Erasing an element from a map also does not invalidate any iterators, except, of course, for iterators that actually point to the element that is being erased.

I would think that applies to reverse_iterators, too.
>How-To-Repeat:
#include <iostream>
#include <map>

using namespace std;

int main()
{
        map<double, double> byloc;
        map<double, double>::reverse_iterator jt;

        byloc.insert(pair<double, double>(90, 1));
        byloc.insert(pair<double, double>(-90, 1));
        byloc.insert(pair<double, double>(270, 1));
        byloc.insert(pair<double, double>(-270, 1));
        jt = byloc.rbegin();
        jt++;
        jt++;
        cout << jt->first << endl;
        byloc.erase(90);
        cout << jt->first << endl;

        byloc.clear();

        byloc.insert(pair<double, double>(90, 1));
        byloc.insert(pair<double, double>(-90, 1));
        byloc.insert(pair<double, double>(270, 1));
        jt = byloc.rbegin();
        jt++;
        jt++;
        cout << jt->first << endl;
        byloc.erase(90);
        cout << jt->first << endl;

        return 0;
}
>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted:


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: libstdc++/10794: std::map's erase method invalidates reverse_iterators
@ 2003-05-15 16:01 paolo
  0 siblings, 0 replies; 2+ messages in thread
From: paolo @ 2003-05-15 16:01 UTC (permalink / raw)
  To: gandy, gcc-bugs, gcc-prs, nobody

Synopsis: std::map's erase method invalidates reverse_iterators

State-Changed-From-To: open->closed
State-Changed-By: paolo
State-Changed-When: Thu May 15 16:01:23 2003
State-Changed-Why:
    Not a bug: see 24.4.1.1 (or Stroustrup, p. 557). In a
    nutshell, reverse_iterator is implemented in terms of a
    plain iterator (called 'current' in the standard) which
    points one position forward. Therefore, in your testcase, upon erase(90), jt 'current' iterator (which points to 90)
    is invalidated.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=10794


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2003-05-15 16:01 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-05-14 23:46 libstdc++/10794: std::map's erase method invalidates reverse_iterators gandy
2003-05-15 16:01 paolo

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).