Boost C++ Libraries: Ticket #2991: Broadcast of bool leaks memory https://svn.boost.org/trac10/ticket/2991 <p> Good morning.<br /> Running valgrind on the attached test program, reveals that broadcasting a single boolean variable leaks memory. The new created data-type is not freed after its usage (and this causes the leak).<br /> Demonstration: if you try to re-build the attached test program with -DLEAK_WORKAROUND, and then run it under valgrind again, the leak is disappeared.<br /> <br /> May you fix this bug, please ?<br /> <br /> Best regards.<br /> <br /> Michele De Stefano </p> en-us Boost C++ Libraries /htdocs/site/boost.png https://svn.boost.org/trac10/ticket/2991 Trac 1.4.3 micdestefano@… Tue, 05 May 2009 08:12:58 GMT attachment set https://svn.boost.org/trac10/ticket/2991 https://svn.boost.org/trac10/ticket/2991 <ul> <li><strong>attachment</strong> → <span class="trac-field-new">bool_bcast.cpp</span> </li> </ul> <p> Test program that demonstrates the bug and its possible solution </p> Ticket Matthias Troyer Tue, 05 May 2009 16:36:17 GMT status, version, severity changed; resolution set https://svn.boost.org/trac10/ticket/2991#comment:1 https://svn.boost.org/trac10/ticket/2991#comment:1 <ul> <li><strong>status</strong> <span class="trac-field-old">new</span> → <span class="trac-field-new">closed</span> </li> <li><strong>version</strong> <span class="trac-field-old">Boost 1.38.0</span> → <span class="trac-field-new">Boost Development Trunk</span> </li> <li><strong>resolution</strong> → <span class="trac-field-new">duplicate</span> </li> <li><strong>severity</strong> <span class="trac-field-old">Problem</span> → <span class="trac-field-new">Cosmetic</span> </li> </ul> <p> This is a duplicate of <a class="closed ticket" href="https://svn.boost.org/trac10/ticket/2594" title="#2594: Bugs: Bug into all_reduce (closed: fixed)">#2594</a>. It is a cosmetic issue that comes up with valgrind but has no other impact. </p> Ticket anonymous Wed, 06 May 2009 05:59:42 GMT status changed; resolution deleted https://svn.boost.org/trac10/ticket/2991#comment:2 https://svn.boost.org/trac10/ticket/2991#comment:2 <ul> <li><strong>status</strong> <span class="trac-field-old">closed</span> → <span class="trac-field-new">reopened</span> </li> <li><strong>resolution</strong> <span class="trac-field-deleted">duplicate</span> </li> </ul> <p> Replying to <a class="ticket" href="https://svn.boost.org/trac10/ticket/2991#comment:1" title="Comment 1">troyer</a>: </p> <blockquote class="citation"> <p> This is a duplicate of <a class="closed ticket" href="https://svn.boost.org/trac10/ticket/2594" title="#2594: Bugs: Bug into all_reduce (closed: fixed)">#2594</a>. It is a cosmetic issue that comes up with valgrind but has no other impact.<br /> </p> </blockquote> <p> <br /> Excuse me, but ticket <a class="closed ticket" href="https://svn.boost.org/trac10/ticket/2594" title="#2594: Bugs: Bug into all_reduce (closed: fixed)">#2594</a> shows something totally different ... please, read carefully my comments for that ticket (the problem for that ticket is absolutely not cosmetic and this ticket is not a duplicate of that one ... for ticket <a class="closed ticket" href="https://svn.boost.org/trac10/ticket/2594" title="#2594: Bugs: Bug into all_reduce (closed: fixed)">#2594</a>, if you don't fix the bug the code does not get built at all, because you receive an error from the compiler).<br /> <br /> For this ticket, I repeat, if you free the new data type, no leak is generated (please, look at the example code I sent). I don't think this is cosmetic: who does free the new data type if you never call MPI_Type_free ? </p> Ticket anonymous Wed, 06 May 2009 06:57:57 GMT status changed; resolution set https://svn.boost.org/trac10/ticket/2991#comment:3 https://svn.boost.org/trac10/ticket/2991#comment:3 <ul> <li><strong>status</strong> <span class="trac-field-old">reopened</span> → <span class="trac-field-new">closed</span> </li> <li><strong>resolution</strong> → <span class="trac-field-new">duplicate</span> </li> </ul> <p> Sorry, I meant 2586. This is cosmetic since the memory would be released by a destructor call of a static member variable. This destructor would only called at the end of the program execution anyway, and the leak thus has no effect besides being detected by valgrind. Hence I call it cosmetic. </p> Ticket anonymous Wed, 06 May 2009 07:31:28 GMT <link>https://svn.boost.org/trac10/ticket/2991#comment:4 </link> <guid isPermaLink="false">https://svn.boost.org/trac10/ticket/2991#comment:4</guid> <description> <p> Replying to <a class="ticket" href="https://svn.boost.org/trac10/ticket/2991#comment:3" title="Comment 3">anonymous</a>: </p> <blockquote class="citation"> <p> Sorry, I meant 2586. This is cosmetic since the memory would be released by a destructor call of a static member variable. This destructor would only called at the end of the program execution anyway, and the leak thus has no effect besides being detected by valgrind. Hence I call it cosmetic. </p> </blockquote> <p> Ok. Clear.<br /> Thank you very much. </p> </description> <category>Ticket</category> </item> </channel> </rss>