zhyty wrote: > I wonder if instead of a boolean we should pass down a function that > schedules the module loading as suggested in #10746. > > That gives more flexibility in that we could immediately schedule the preload > to happen in parallel (as done in the other PR), or for your use case we > could accumulate the preload callbacks into a vector and then schedule all > the preload jobs to run in parallel once all the modules have been loaded. > > Any thoughts @DmT021?
I just finished writing a small blurb explaining why I went with this approach, so maybe that helps justify this approach over that one? > That gives more flexibility in that we could immediately schedule the preload > to happen in parallel (as done in the other PR) Immediately scheduling the preload to happen in parallel shouldn't be thread-safe due to the ABBA deadlock. I don't think that flexibility is good, then, if it leads callers into this deadlock. https://github.com/llvm/llvm-project/pull/166480 _______________________________________________ lldb-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
