We are running into the case where we setup a default clang::ASTContext for every shared library. We could argue that as soon as you create the clang::ClangAST that you must then search for a "char" type by name and see if it is unsigned and if so, modify the clang::ASTContext. But we currently try to parse things only when needed in LLDB so I don't want to have to pull in all of the DWARF accelerator tables and load a type just to find out if chars should be signed. Why? Because on MacOSX we use DWARF in .o files and one shared library might have 1000 .o files. Now we try to make the clang::ASTContext for this shared library and we must pull in all 1000 .o files just to check if any have a unsigned char.
So I would vote to continue with the current functionality and just call "char" that are actually unsigned "unsigned char" and fix the test that have problems with that. http://reviews.llvm.org/D8636 EMAIL PREFERENCES http://reviews.llvm.org/settings/panel/emailpreferences/ _______________________________________________ lldb-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/lldb-commits
