Version: (using Devel) OS: Linux Installed from: Compiled sources The Valgrind doc might benefit from a section for people used to Purify. For example: Projects that use both Purify and Valgrind (say, one on Windows and one on Mac and Linux) might like to use a uniform set of names for memory access errors. It'd be nice if the language purify and valgrind used to describe their warnings were more uniform, and in particular, if valgrind could be told to tag each error using the same TLA as Purify. A few purify error types are UMR (Uninitialized memory read) UMW (Uninitialized memory write) MLK (Memory leak) FMR (Free memory read) FUM (Freeing unallocated memory) FNH (Freeing non-heap memory) NPW (Null pointer write) NPR (Null pointer read) This is petty because projects in this situation can trivially write a sed script to do this, and impossible because there isn't a one-to-one mapping between the errors found by the two products. But perhaps a best effort script could be given in the documentation.
Why pick on purify for this special treatment? What about other free and/or commercial products... Should we tag each message with codes to match each of them? That might quickly get somewhat out of hand...
I'd be happy with a section in the doc for each competing tool, welcoming its users to valgrind. Note that my suggested resolution was a sed script in the documentation, and we could provide one of those for each competing tool. I only mention purify because that's the competing tool we happen to use alongside valgrind. (We also use Coverity, but that's not quite a competitor.)
I think that this is now obsolete.