shejialuo <shejialuo@xxxxxxxxx> writes: > As discussed in [1], we need to use mmap mechanism to open large > "packed_refs" file to save the memory usage. This patch mainly does the > following things: > > 1: Fix an issue that we would report an error when the "packed-refs" > file is empty, which does not align with the runtime behavior. > 2-4: Extract some logic from the existing code and then use these > created helper functions to let fsck code to use mmap necessarily > > [1] https://lore.kernel.org/git/20250503133158.GA4450@xxxxxxxxxxxxxxxxxxxxxxx > > Really thank Peff and Patrick to suggest me to do above change. > ... > Change in v5: > > 1. Improve the commit message in the first patch to be more clear: > 1. Talk about the current behavior, what error we would report if > "packed-refs" is empty. > 2. To align with the runtime behavior, we should skip checking the > content of "packed-refs". > 3. Why do we need to report to the user when the "packed-refs" is > empty > 2. Fix grammar issue in the last patch. The thread has gone quiet on this topic. Is everybody happy with this version? Thanks.
