labath wrote: > > * what about commands (I don't know if we have any) whose output consists > > of more than formatting a single value (which, I guess would be lost if the > > IDE decides to ignore the output). Are we fine with not supporting that? > > (we could say that if a command wants to print some extra output, it should > > use the error stream for that) > > In the longer term, we can also add an optional SBStructuredData for commands > that return structured data, like `image lists` or `image lookup` so the UI > can print these as tables or something else nice. So if we end up having > fancier commands that have both a ValueObject return plus extra annotation, > we could add an SBValue type to what an SBStructured data can hold, and have > it emit `before` and `after` entries for the extra text. > > I think that would be better than trying to dump some of the output to stderr.
That sort of makes sense, but it doesn't sound like you're planning to do that soon, so I'm asking about what should we do until (and if) that happens. The reason I suggested stderr is because that made sense for the commands I'm thinking of. For example, the expression command can produce a result and some warnings and fixit-applied notes. I don't know if these currently go to stderr, but I think it would be okay if they did... https://github.com/llvm/llvm-project/pull/127566 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
