#!/bin/bash

set -u

base=$1
next=$2
rc=$3

# to stabilise the sorts
export LANG=C

sd=../stats

[ -d $sd ] || mkdir $sd || {
	echo "No stats dir" 1>&2
	exit 1
}

git rev-list --no-merges $base..$next | sort >$sd/$next.sha1s
git rev-list --no-merges $base..$rc | sort >$sd/$rc.sha1s

# filter out the common commits
comm -12 $sd/$next.sha1s $sd/$rc.sha1s >$sd/common_sha1
comm -23 $sd/$next.sha1s $sd/$rc.sha1s >$sd/$next.sha1s.1
comm -13 $sd/$next.sha1s $sd/$rc.sha1s >$sd/$rc.sha1s.1

# now find commits with the same patch-id
git diff-tree -p --stdin <$sd/$next.sha1s.1 |
	git patch-id |
	sort >$sd/$next.patch_ids
git diff-tree -p --stdin <$sd/$rc.sha1s.1 |
	git patch-id |
	sort >$sd/$rc.patch_ids
join -j 1 $sd/$next.patch_ids $sd/$rc.patch_ids >$sd/common_id
join -v 1 -j 1 $sd/$next.patch_ids $sd/$rc.patch_ids |
	sed 's/^.* //' >$sd/$next.sha1s.2
join -v 2 -j 1 $sd/$next.patch_ids $sd/$rc.patch_ids |
	sed 's/^.* //' >$sd/$rc.sha1s.2

# now find commits with the same subject line
git log --stdin --no-walk --format='%s|%H' <$sd/$next.sha1s.2 |
	sed 's/|\(.*\)|/_\1|/g' |
	sort -f -k 1,1 -t '|' >$sd/$next.subj
git log --stdin --no-walk --format='%s|%H' <$sd/$rc.sha1s.2 |
	sed 's/|\(.*\)|/_\1|/g' |
	sort -f -k 1,1 -t '|' >$sd/$rc.subj
join -i -j 1 -t '|' $sd/$next.subj $sd/$rc.subj >$sd/common_subj
join -v 1 -i -j 1 -t '|' $sd/$next.subj $sd/$rc.subj |
	sed 's/^.*|//' >$sd/$next.sha1s.3
join -v 2 -i -j 1 -t '|' $sd/$next.subj $sd/$rc.subj |
	sed 's/^.*|//' >$sd/$rc.sha1s.3

nrc=$(wc -l <$sd/$rc.sha1s)
nnext=$(wc -l <$sd/$next.sha1s)
ncommsha1=$(wc -l <$sd/common_sha1)
ncommid=$(wc -l <$sd/common_id)
ncommsubj=$(wc -l <$sd/common_subj)
ncomm=$((ncommsha1+ncommid+ncommsubj))
pcomm=$((ncomm*100/nrc))

cat <<EOF
Hi all,

As usual, the executive friendly graph is at
http://neuling.org/linux-next-size.html :-)

(No merge commits counted, $next was the first linux-next after
the merge window opened.)

EOF
printf "%-50s %5d\n" "Commits in $rc (relative to $base):" $nrc
printf "%-50s %5d\n" "Commits in $next:" $nnext
printf "%-50s %5d\n" "Commits with the same SHA1:" $ncommsha1
printf "%-50s %5d (1)\n" "Commits with the same patch_id:" $ncommid
printf "%-50s %5d (1)\n" "Commits with the same subject line:" $ncommsubj

cat <<EOF

(1) not counting those in the lines above.

EOF

printf "%-50s %5d %2d%%\n" "So commits in -rc1 that were in $next:" $ncomm $pcomm

cat <<EOF

Some breakdown of the list of extra commits (relative to $next)
in -rc1:

Top ten first word of commit summary:

$(git log --no-walk --stdin --pretty="format:%s" <$sd/$rc.sha1s.3 |
	sed -e 's/^\[//;s/[]:( /,;].*$//' |
	tr 'A-Z' 'a-z' | sort | uniq -c | sort -nr | head -n 10)

Top ten authors:

$(git log --no-walk --stdin --pretty="format:%aE" <$sd/$rc.sha1s.3 |
	tr 'A-Z' 'a-z' | sort | uniq -c | sort -nr | head -n 10)

Top ten commiters:

$(git log --no-walk --stdin --pretty="format:%cE" <$sd/$rc.sha1s.3 |
	tr 'A-Z' 'a-z' | sort | uniq -c | sort -nr | head -n 10)

There are also $(wc -l <$sd/$next.sha1s.3) commits in $next that didn't make it into
$rc.

Top ten first word of commit summary:

$(git log --no-walk --stdin --pretty="format:%s" <$sd/$next.sha1s.3 |
	sed -e 's/^\[//;s/[]:( /,;].*$//' |
	tr 'A-Z' 'a-z' | sort | uniq -c | sort -nr | head -n 10)

Top ten authors:

$(git log --no-walk --stdin --pretty="format:%aE" <$sd/$next.sha1s.3 |
	tr 'A-Z' 'a-z' | sort | uniq -c | sort -nr | head -n 10)

Some of Andrew's patches are fixes for other patches in his tree (and
have been merged into those).

Top ten commiters:

$(git log --no-walk --stdin --pretty="format:%cE" <$sd/$next.sha1s.3 |
	tr 'A-Z' 'a-z' | sort | uniq -c | sort -nr | head -n 10)

Those commits by me are from Andrew's mmotm tree.
EOF