llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT--> @llvm/pr-subscribers-lldb Author: Med Ismail Bennani (medismailben) <details> <summary>Changes</summary> Starting `27f27d1`, we changed the image loading logic to conform to the various options (`-a|--load-all` & `-c|--crashed-only`) and set the `resolve` attribute of the matching images. Then, we would try to load all those images concurrently. However, the `symbolicator.Image.add_module` method didn't check the `resolve` attribute to skip unwanted images, causing all images to be loaded in the target everytime. This matches the `-a|--load-all` option behaviour but it can cause lldb to load thousands of images, which can take a very long time if the images are downloaded over the network. This patch fixes that issue by checking the `Image.resolve` attribute to conform to the user selection, and reinstates the expected default behaviour, by only loading the crashed thread images and skipping all the others. rdar://123694062 --- Full diff: https://github.com/llvm/llvm-project/pull/91109.diff 1 Files Affected: - (modified) lldb/examples/python/symbolication.py (+6) ``````````diff diff --git a/lldb/examples/python/symbolication.py b/lldb/examples/python/symbolication.py index f6dcc8b9a79437..f247ebf3da733a 100755 --- a/lldb/examples/python/symbolication.py +++ b/lldb/examples/python/symbolication.py @@ -399,6 +399,12 @@ def add_module(self, target, obj_dir=None): if not self.path and self.uuid == uuid.UUID(int=0): return "error: invalid image" + if not self.resolve: + # Since this method get called concurrently on every crashlog image, + # we should only add the images marked to be resolved and skip the + # others. + return None + if target: # Try and find using UUID only first so that paths need not match # up `````````` </details> https://github.com/llvm/llvm-project/pull/91109 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
