https://pfspipe.ipmu.jp/jira This file is an XML representation of an issue en-us 8.3.4 803005 13-09-2019 https://pfspipe.ipmu.jp/jira/browse/PIPE2D-33 DRP 2-D Pipeline <p>Currently FiberTraceSet is a std::vector&lt;PTR&lt;FiberTrace&gt;&gt;. For a SpectrumSet I originally had the same concept, but I found that the only way to get rid of the compiler warning "Returning address of a stack object" (something like that) was to go from std::vector&lt;PTR&lt;Spectrum&gt;&gt; to std::vector&lt;Spectrum&gt;. Actually I now have an idea what I might have done wrong (probably created a shared pointer without the "new" in the argument list - should check that) so I think for consistency and to reduce copying I should probably change that back to std::vector&lt;PTR&lt;Spectrum&gt;&gt; and try again...)</p> PIPE2D-33

Sort out Spectrum containers

Bug Major Done Done aritter aritter Fri, 1 Jul 2016 18:21:04 +0000 Tue, 11 Oct 2016 09:59:00 +0000 Mon, 10 Oct 2016 00:55:56 +0000 0 2 <p>Changed SpectrumSet._spectra to PTR(Vector&lt;PTR(Spectrum)&gt;). Tests pass, but I'm still getting a compiler warning:<br/> python/pfs/drp/stella/stellaLib_wrap.cc:4601:9: warning: destination for this 'memset' call is a pointer to dynamic class 'Spectrum&lt;float, unsigned short, float, float&gt;'; vtable po<br/> inter will be overwritten <span class="error">&#91;-Wdynamic-class-memaccess&#93;</span></p> <p>We have a similar warning in LSST (ticket DM-1401). I just gonna leave it like that for now...</p> <p>Pretty sure the plan was you'd chat to JimB about this, then set it back to "in review" if he doesn't have any good suggestions and agrees that this warning is likely harmless.</p> <p>Jim Bosh didn't really have anything to say about this. Here is the function that contains the error:</p> <div class="code panel" style="border-width: 1px;"><div class="codeHeader panelHeader" style="border-bottom-width: 1px;"><b>stellaLib_wrap.cc</b></div><div class="codeContent panelContent"> <pre class="code-java"> template &lt;<span class="code-keyword">class </span>Type&gt; struct traits_as&lt;Type, pointer_category&gt; { <span class="code-keyword">static</span> Type as(PyObject *obj, bool throw_error) { Type *v = 0; <span class="code-object">int</span> res = (obj ? traits_asptr&lt;Type&gt;::asptr(obj, &amp;v) : SWIG_ERROR); <span class="code-keyword">if</span> (SWIG_IsOK(res) &amp;&amp; v) { <span class="code-keyword">if</span> (SWIG_IsNewObj(res)) { Type r(*v); delete v; <span class="code-keyword">return</span> r; } <span class="code-keyword">else</span> { <span class="code-keyword">return</span> *v; } } <span class="code-keyword">else</span> { <span class="code-comment">// Uninitialized <span class="code-keyword">return</span> value, no Type() constructor required. </span> <span class="code-keyword">static</span> Type *v_def = (Type*) malloc(sizeof(Type)); <span class="code-keyword">if</span> (!PyErr_Occurred()) { SWIG_Error(SWIG_TypeError, swig::type_name&lt;Type&gt;()); } <span class="code-keyword">if</span> (throw_error) <span class="code-keyword">throw</span> std::invalid_argument(<span class="code-quote">"bad type"</span>); memset(v_def,0,sizeof(Type)); <span class="code-keyword">return</span> *v_def; } } }; </pre> </div></div> <p>Seems like the </p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java">memset</pre> </div></div> <p> command is only called if </p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java">v</pre> </div></div> <p> is uninitialized, so as long as our code is correct I believe this should not cause a problem.</p> <p>Merged into master</p> Development Rank 0|02qpuc:zw Reviewers swinbank Sprint 2014-14 2014-15 2014-16