jimingham wrote: I don't understand what this address plus address offset breakpoint type is supposed to be?
We have "address plus offset" already - where the initial address is a Section Offset - that's what Address-es are. Addresses are very handy for address breakpoint setting because they allow you to track the container (the section) if it unloads and reloads either because the library containing the section was unloaded & reloaded or because you reran. Internally, lldb will try to turn a "load address" breakpoint into an Address, and it will only try to reload on re-run address breakpoints if they were resolvable to an Address. That's because you have to be careful about setting raw address breakpoints, if you put them in the wrong place they can cause subtle and hard to diagnose debugger-caused bugs. Why could you not use Address-es to set this two-form breakpoint? Given a module and an lldb::addr_t that is in the module, you can always resolve the addr_t to it's section+offset form, so you can always make an Address. What do you gain with your new breakpoint setting form? Maybe you should write the documentation for your proposed new API so we can see what it is supposed to be for? https://github.com/llvm/llvm-project/pull/148061 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
