| Item | File | Description | Consensus | Who | Status | Notes |
|---|
| 1 | isolineparams.cpp | IsolineParams::validateIsoControls() accesses every 2D and 3D variable in the data set whether they are needed or not | y | Alan | Done | Variable bounds are obtained as needed. |
| 2 | isolineparams.cpp | replace new3DIsoControls[i]->setVarNum(i) with setVarName(). Don't use variable indices when their meaning can change from data set to data set, session to session. | y | Alan | Done | Alan did evaluate. |
| 3 | isolineparams.cpp | Does there need to be a distinction made between 2D and 3D IsoControls? A lot of redundant code could be collapsed if they were not treated separately. | y | Alan | Done | |
| 4 | isolineparams.cpp | Type 1 validation in IsolineParams::validateIsoControls() appears to be reconciling different variables by using their integer offset in the data set. Shouldn't we be using the variables names, not the offset? I.e. The variable "temp" probably has the same meaning in two different data sets, but may have different offsets. | ? | Alan | Done | Alan will evaluate. |
| 5 | isolinerenderer.cpp | Move responsibility for cache management inside of isolinerender.cpp. Clients of the renderer classes shouldn't have to know what changes necessitate cache invalidation. They shouldn't even have to know that a renderer uses a cache. Renderers that maintain a cache should retain parameter values for parameters that would force an update of the cache. | y | Alan | Done | |
| 6 | isolinerenderer.cpp | Renderer's should be subordinate to ControlExec. I.e. They shouldn't have access to it. | y | Alan | open | discuss ion context of control exec review |
| 7 | isolineeventrouter.cpp | Memory leaks: GLIsolineWindow, Histo, | y | Alan | Done | Qt is freeing the GLIsolineWindow memory, but not Histo. |
| 8 | isolineeventrouter.cpp | Can _dataMgr ever be NULL? Should the render event routers be enabled when data aren't loaded? | y | Alan | Done | Change tests for existence to asserts, or remove unnecessary tests. |
| 9 | isolineeventrouter.cpp | isolineeventrouter is really long and complex (~1200 lines), but the GUI it supports is fairly typical of all of the tabs. What can be done to simplify so that other render event routers aren't as long and complex? | y | Alan | Done | Reorganize the cpp file to place common elements together (e.g. by subtabs) and document them a such |
| 10 | isovalueeditor.cpp | Lots of memory leaks. | n | | | All the QObjects in the editor get deleted when their parent is deleted. Read the Qt documentation on QObjects. |
| 11 | isolinerender.cpp | Unnecessary ControlExec access. E.g. uses ControlExec to get DataMgr inside buildLineCache(), which is passed DataMgr as an argument. | y | | Done | |
| 12 | isolinerender.cpp | Make invalidateLineCache(), buildLineCache(), etc. private | y | | Done | |
| 13 | isolineeventrouter.cpp, isolinerenderer.cpp | Check for NULL before deleting pointers (e.g. in destructor) | y | | Done | |
| 14 | isolineeventrouter.cpp | Remove ForceRenderer requirement. Specialize QWidget::Event()?? | y | | Done | ForceRender has been removed. Now various Qt events will trigger rerender. However transfer function edits do not necessarily trigger a Qt event so these now cause a rerender upon completion. |
| 15 | glisolinewindow.cpp | Rendering code duplicates isolinerenderer. Have isolinerenderer perform rendering for tab and eliminate/reduce glisolinewindow.cpp | ? | | postpone | |
| 16 | isolineparams.cpp | Resource mgt: memory allocated in validateIsoControls is never freed | n | | | These are ParamsBase instances that get destroyed with the Root ParamNode. |