Updating the .mailmap file

Updating the .mailmap file

[Stefan Beller]

Hello,

I was just browsing the shortlog and it looked to 
me as if there are many double entries, i.e. the same person
being there with multiple email addresses.

So I wrote a script, which finds double names and adds them 
to the .mailmap file. However manual checking, whether these 
are really the same person is required.

---
#!/bin/bash

git shortlog -sne > shortlog.txt

# find double names in the shortlog
cat shortlog.txt |awk '{ NF--; $1=""; print }' |sort |uniq -d | sed 's/^ *//g' > names.txt

while read line
do
    echo "process name $line"
    # get the current mailmap file in a state where each person is listed
    # with their first line:
    rm mailmap_tmp*
    cat .mailmap |grep -v "^[# ]"|grep -v "^$" | sort > mailmap_tmp1
    while read <&9 line2
    do
        nameonly="$(echo "$line2" |awk '{NF--; print}')"
        grep -m1 "$nameonly" mailmap_tmp1 >> mailmap_tmp2
    done 9< "mailmap_tmp1"
    cat mailmap_tmp2 |uniq > mailmap_tmp3

    # find all occurences of one name
    cat shortlog.txt |grep "$line" | awk '{ $1=""; print }' | sed 's/^ *//g' > currentname.txt

    # get that guy alphabetically sorted into the mailmap file:
    echo $line > name
    cat mailmap_tmp3 name |sort > mailmap_tmp4

    # find the place where we need to insert this
    insertbeforeline=$(cat mailmap_tmp4 |grep -A1 "$line" |tail -n1)
    if [ "$insertbeforeline" != "" ] ; then
        # insert all in the middle of .mailmap:
        while read <&9 line2; do
            echo "adding line $line2 to mailmap before $insertbeforeline"
            sed "/$insertbeforeline/i $line2" .mailmap > mailmap_tmp5
            mv mailmap_tmp5 .mailmap
        done 9< "currentname.txt"
    else
        # must appended to end
        cat currentname.txt >> .mailmap
    fi

    git commit -a -m "Updating .mailmap file: Adding ${line}" > /dev/null
    echo "---"
done < "names.txt"

rm mailmap_tmp*
rm shortlog.txt
rm names.txt

Re: Updating the .mailmap file

[Joe Perches]

On Sat, 2012-12-01 at 19:50 +0100, Stefan Beller wrote:
> Hello,
> 
> I was just browsing the shortlog and it looked to 
> me as if there are many double entries, i.e. the same person
> being there with multiple email addresses.
> 
> So I wrote a script, which finds double names and adds them 
> to the .mailmap file. However manual checking, whether these 
> are really the same person is required.

You should get together with Philipp Riemer.
http://lkml.org/lkml/2012/5/3/98

I think it's a problem not really worth solving.