determining which files have the most Git commits
Use git rev-list --objects --all | awk '$2' | sort -k2 | uniq -cf1 | sort -rn | head
.
output:
1058 fffcba193374a85fd6a3490f800c6901218a950b src
715 ffffe0f08798e95b66cc4ad4ff22cf10734d045e src/lib
450 ffcfe596031a5985664e35937fff4ac9ff38dcca src/zfs-fuse
367 ffc5d5340f95360fc9f7b739c5593dd3f92fced0 src/lib/libzpool
202 ff92db000792044d45eec21c57a3cd21618631e7 src/lib/libsolkerncompat
183 ff1a44edae3fd121ddd86864b589e5ab2f9ff99b src/lib/libzfscommon
178 fec6b3a789e578983c2242b3aa5adf217cb8b887 src/lib/libzfs
168 ffeefc9e81222d7c471bdb0911d8b98f23cff050 src/cmd
167 fbd60bd3430765863648c52db7ceb3ffa15d5e50 src/lib/libzfscommon/include
155 ff225f6b41f9557d683079c5f9276f497bcb06bd src/lib/libzfscommon/include/sys
explanation
git rev-list --objects --all
gives all objects from all revisions in all branches.awk '$2'
ignores any results without a path.sort -k2
sorts them by path.uniq -cf1
makes them unique (ignoring the blob hash), prefixing lines with duplication count.sort -rn
sorts descending on duplication count.head
shows topmost lines.
Alternatively, to see only file blobs, use:
git rev-list --objects --all | awk '$2' | sort -k2 | uniq -cf1 | sort -rn |
while read frequency sample path
do
[ "blob" == "$(git cat-file -t $sample)" ] && echo -e "$frequency\t$path";
done
output:
135 src/zfs-fuse/zfs_operations.c
84 src/zfs-fuse/zfs_ioctl.c
79 src/zfs-fuse/zfs_vnops.c
73 src/lib/libzfs/libzfs_dataset.c
67 src/lib/libzpool/spa.c
66 src/zfs-fuse/zfs_vfsops.c
62 src/cmd/zdb/zdb.c
62 CHANGES
60 src/cmd/ztest/ztest.c
60 src/lib/libzpool/arc.c
Alternatively, to see only specifc range of revisions, use:
git rev-list --after=2011-01-01 --until='two weeks ago' \
tag1...remote/hotfix ^master
This will use only revisions in the specified date range, that are in the symmetric set difference for tag1
and remote/hotfix
and are not in master.
licensing
Some rights reserved: CC BY-SA 3.0. Includes significant content from an answer on Stack Overflow by sehe with changes made, including converting the original HTML to Markdown.