[cfe-dev] Clang Static Analyzer Question about Plists

[Artem Dergachev via cfe-dev]

Yeah, plist format is slightly different. Instead of grey notes that 
describe control flow as text, it describes control flow as "arrows" 
from one point in the source code to another. Eg., instead of writing 
"Taking true branch..." it'd list two (file, line, column) triples: one 
at the end of the branch condition, the other at the beginning of the 
true branch, and you can imagine this as an arrow from the first source 
location to the second source location. You can see how it's supposed to 
look on the top screenshot at http://clang-analyzer.llvm.org/

Hmm, there doesn't seem to be a way to override this decision right now. 
You can try tweaking PlistDiagnostics::getGenerationScheme(), i.e. 
change Extensive to Minimal, but i can't promise it won't crash when it 
tries to generate anything.



On 1/16/19 12:45 PM, Trevor Huang via cfe-dev wrote:
> Hello all,
>
> I am Trevor Huang, a student at University of Pennsylvania and one of 
> my classmates, Gianluca Gross, recently asked some questions about 
> Clang Static Analyzer for our senior design project. He forwarded us 
> some responses from Alexey Sidorin and Gabor Horvath, so thank you so 
> much for the assistance!
>
> One recommendation given to help us in our project is to use the plist 
> output from scan-build instead of the html output for our parsing. I 
> was looking through the contents of the plist and found that 
> statements such "Taking true branch," and other statements in the bug 
> trace that are in gray (instead of yellow) in the html report does not 
> appear to be present in the plist format. I was wondering where those 
> statements come from and if there is an easy way for me to include 
> them if I only read the plist output.
>
> Any and all help would be greatly appreciated. Thank you again!
>
> Sincerely,
> Trevor
>
> _______________________________________________
> cfe-dev mailing list
> cfe-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20190118/22bd31cf/attachment.html>