--- layout: fc_discuss_archives title: Message 4 from Frama-C-discuss on April 2012 ---
Hi St?phane, On Mon, Apr 2, 2012 at 10:09 PM, DUPRAT Stephane <stephane.duprat at atos.net> wrote: > As for us, we are interested in this information in two ways: > ?1/ to help user in analysing causes of warning in going back in the call tree > I previously used logs like " [value] computing for function f1 <- f2 <- main." processed by a awk file that displays the last path preceding a warning. Useful, when we have a lot of warnings at a memcpy call, for example. > So, the callstack in that form will be very useful for us and certainly for others that are using Frama-C/Value. > ?2/ The second reason will be to process them in the IDE integration (FCDT) in order to display a marker at each stage of a warning in the callstack, and perhaps, by providing navigation features to inspect warnings in the full project. > > In both usage, I just need to have all path for all warnings. I don't know if I need to get the information, each time an assertion is analysed. I have added callstacks to all warnings that give rise to ACSL assertions, which should improve things significantly. The remaining ones will have to be converted one by one; this will happen gradually before the release of Oxygen, as we clean up some parts of this code. Note that option -val-print-callstacks may outlive its usefulness with the Fluorine release (and might be subsequently removed), as we plan to rewrite major parts of the emission of alarms. If this succeeds, the good approach will be to use a lightweigtht plugin that will listen to all emitted alarms, and decide what it wants to see printed. -- Boris