labath added a comment.

I am glad you're finding lldb-test useful. Please see my inline comment about 
quoting.



================
Comment at: lldb/lit/SymbolFile/DWARF/compilercontext.ll:1
+; This simulates the debug info for a Clang module.
+source_filename = "/t.c"
----------------
We generally put test data into an `Inputs` subfolder. Though in this case, it 
looks like you could actually inline the data into the test file below. (Feel 
free to add `.ll` into the list of test suffixes.)


================
Comment at: lldb/tools/lldb-test/lldb-test.cpp:145
+    "compiler-context",
+    cl::desc("Specify a compiler context as \"[(type,name),...]\"."),
+    cl::value_desc("context"), cl::sub(SymbolsSubcommand));
----------------
JDevlieghere wrote:
> aprantl wrote:
> > JDevlieghere wrote:
> > > > ...as a JSON array...`
> > > 
> > > This makes me wonder though, wouldn't it be better to pass a file here 
> > > and have it live in inputs? Do you expect it to be different across tests 
> > > or could we reuse the context across tests?
> > I don't really expect a DeclContexts to have more than 3 entries in the 
> > forseeable future, so that would arguable just make the test harder to read.
> But would those entries be the same across tests or not? If they're the same 
> I think it'd be better to reuse them, even if it's only a few entries.
I actually think it is nice to have the entries present in the test file 
directly. However, what I think will be an issue is that you need to use two 
layers of quotes to pass a json string over the command line. I expect this 
will cause problems on windows, because quoting works differently there.

What I would actually suggest is to ditch JSON, and use a custom format for 
passing this. Maybe something like `--compiler-context 
Module:CModule,Module:SubModule,...`. That would be both shorter and completely 
avoid quotes. I don't think is should make parsing this signtificantly more 
difficult than what you have now.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D66453/new/

https://reviews.llvm.org/D66453



_______________________________________________
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits

Reply via email to