From 48ba1f4ced543147287bb9043dd10b8230ab742f Mon Sep 17 00:00:00 2001
From: Alexander Kuzmenkov <akuzm@yandex-team.ru>
Date: Tue, 14 Apr 2020 14:28:27 +0300
Subject: [PATCH] simple changelog script

---
 utils/simple-backport/changelog.sh        | 26 +++++++++++++++++++----
 utils/simple-backport/format-changelog.py |  7 +++++-
 2 files changed, 28 insertions(+), 5 deletions(-)

diff --git a/utils/simple-backport/changelog.sh b/utils/simple-backport/changelog.sh
index b2f95f22533..d5a4e68900a 100755
--- a/utils/simple-backport/changelog.sh
+++ b/utils/simple-backport/changelog.sh
@@ -3,8 +3,9 @@ set -e
 
 from="$1"
 to="$2"
+log_command=(git log "$from..$to" --first-parent)
 
-git log "$from..$to" --first-parent > "changelog-log.txt"
+"${log_command[@]}" > "changelog-log.txt"
 
 # NOTE keep in sync with ./backport.sh.
 # Search for PR numbers in commit messages. First variant is normal merge, and second
@@ -14,11 +15,18 @@ find_prs=(sed -n "s/^.*Merge pull request #\([[:digit:]]\+\).*$/\1/p;
                   s/^.*back[- ]*port[ed of]*#\([[:digit:]]\+\).*$/\1/Ip;
                   s/^.*cherry[- ]*pick[ed of]*#\([[:digit:]]\+\).*$/\1/Ip")
 
-"${find_prs[@]}" "changelog-log.txt" | sort -rn > "changelog-prs.txt"
-
+"${find_prs[@]}" "changelog-log.txt" | sort -rn | uniq > "changelog-prs.txt"
 
 echo "$(wc -l < "changelog-prs.txt") PRs added between $from and $to."
 
+if "${log_command[@]}" --oneline --grep "Merge branch '" | grep ''
+then
+    # DO NOT ADD automated handling of diamond merges to this script.
+    # It is an unsustainable way to work with git, and it MUST be visible.
+    echo Warning: suspected diamond merges above.
+    echo Some commits will be missed, review these manually.
+fi
+
 function github_download()
 {
     local url=${1}
@@ -39,6 +47,7 @@ function github_download()
     fi
 }
 
+rm changelog-prs-filtered.txt &> /dev/null ||:
 for pr in $(cat "changelog-prs.txt")
 do
     # Download PR info from github.
@@ -51,6 +60,13 @@ do
         continue
     fi
 
+    # Filter out PRs by bots.
+    user_login=$(jq -r .user.login "$file")
+    if echo "$user_login" | grep "\[bot\]$" > /dev/null
+    then
+        continue
+    fi
+
     # Download author info from github.
     user_id=$(jq -r .user.id "$file")
     user_file="user$user_id.json"
@@ -61,9 +77,11 @@ do
         >&2 echo "Got wrong data for user #$user_id (please check and remove '$user_file')."
         continue
     fi
+
+    echo "$pr" >> changelog-prs-filtered.txt
 done
 
 echo "### ClickHouse release $to FIXME as compared to $from
 " > changelog.md
-./format-changelog.py changelog-prs.txt >> changelog.md
+./format-changelog.py changelog-prs-filtered.txt >> changelog.md
 cat changelog.md
diff --git a/utils/simple-backport/format-changelog.py b/utils/simple-backport/format-changelog.py
index 8f6966eba02..356ed48b6fd 100755
--- a/utils/simple-backport/format-changelog.py
+++ b/utils/simple-backport/format-changelog.py
@@ -30,6 +30,11 @@ def parse_one_pull_request(item):
                 i += 1
                 if i >= len(lines):
                     break
+                # Can have one empty line between header and the category itself. Filter it out.
+                if not lines[i]:
+                    i += 1
+                    if i >= len(lines):
+                        break
                 category = re.sub(r'^[-*\s]*', '', lines[i])
                 i += 1
             elif re.match(r'(?i)^\**\s*(Short description|Change\s*log entry)', lines[i]):
@@ -91,7 +96,7 @@ def print_category(category):
         user_name = user["name"] if user["name"] else user["login"]
 
         # Substitute issue links
-        pr["entry"] = re.sub(r'#([0-9]{4,})', r'[#\1](https://github.com/ClickHouse/ClickHouse/issues/\1)', pr["entry"])
+        pr["entry"] = re.sub(r'([^[])#([0-9]{4,})', r'\1[#\2](https://github.com/ClickHouse/ClickHouse/issues/\2)', pr["entry"])
 
         print(f'* {pr["entry"]} [#{pr["number"]}]({pr["html_url"]}) ([{user_name}]({user["html_url"]})).')