git release notes #git #awk #release #notes

git_log_format.awk

#! /usr/bin/awk -f

BEGIN {
        FS="\][ ]*[-]*[ ]*"
}
/^\[.*\]/ {
        sub(/\[/, "", $1)
        if (NR == 1) {
                previous_commit_tag = $1
                commit_message = "\t- " $2
                print previous_commit_tag "\n" commit_message
                next
        }

        commit_message = "\t- " $2
        commit_tag = $1
         if(previous_commit_tag != commit_tag) {
                 print"\n" commit_tag "\n" commit_message
         } else {
                 print commit_message
         }
         previous_commit_tag = commit_tag
}
END {}

git_release_notes.sh

#!/bin/bash

set -e
PAST_TAG=$(git describe --tags --abbrev=0)
REF=${1:-$PAST_TAG}
git log $REF..HEAD --pretty=format:"%s" | ./format_git_log

Usage

Put both scripts in same location.

Get formatted output from latest git tag to HEAD:

./git_release_notes.sh

Get formatter output from a give commit to HEAD:

./git_release_notes.sh COMMIT_HASH

Output

Process git log where commits have been written in the format. of:

[JIRA_TICKET_ID] COMMIT_MSG or [JIRA_TICKET_ID] - COMMIT_MSG

If multiple lines have the same ticket id these will be summarized in a section like. this

JIRA_TICKET_ID
    - COMMIT_MSG_1
   - COMMIT_MSG_2

JIRA_TICKET_ID
    -  COMMIT_MSG_1
    - COMMIT_MSG_2

TODO, first parse a git commit. template.