Alexander_Droste marked an inline comment as done.
================
Comment at: tools/clang/lib/StaticAnalyzer/Core/MemRegion.cpp:653
@@ +652,3 @@
+ llvm::SmallString<2> intValAsString;
+ IndexInArray.toString(intValAsString);
+ std::string idx{intValAsString.begin(), intValAsString.end()};
----------------
xazax.hun wrote:
> What will be the content of IndexInArray in case the index is not a concrete
> int? Will the above code work at all in that case? In case the index is a
> variable do we want to include the name of the variable in the index?
Right, `getAs<..::ConcreteInt>()` might return a `nullptr`.
What do you think of including the number if it is a `nonloc::ConcreteInt`,
the name if it is a `nonloc::SymbolValKind`
(`getAs<nonloc::SymbolVal>().getSymbol()`)
and else keeping the brackets empty?
================
Comment at: tools/clang/lib/StaticAnalyzer/Core/MemRegion.cpp:656
@@ +655,3 @@
+
+ ArrayIndices = "[" + idx + "]" + ArrayIndices;
+
----------------
xazax.hun wrote:
> I do not really like this part. You could either use Twine to avoid some
> temporal values (and you could also eliminate the string copy this way) or
> you could construct the indices using the reverse of idx and than reverse the
> whole string after the loop.
So would be invoking `push_back` within and `reverse` after the loop sufficient?
http://reviews.llvm.org/D16044
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
