Use gitnamespaces for efficient storage

Instead of using one Git repository per package, use a single large
object storage for space efficiency. The refs of the individual package
bases are divided using gitnamespaces(7) which allows for exposing each
namespace as an independent repository easily. Also, git-serve is
modified to create a branch for each package, allowing to browse the
large repository with cgit.

Helped-by: Florian Pritz <bluewind@xinu.at>
Helped-by: Johannes Löthberg <johannes@kyriasis.com>
Signed-off-by: Lukas Fleischer <lfleischer@archlinux.org>

4 files changed, 39 insertions, 115 deletions

diff --git a/scripts/git-integration/gen-templates.py b/scripts/git-integration/gen-templates.py
deleted file mode 100755
index e451b53e..00000000
--- a/scripts/git-integration/gen-templates.py
+++ /dev/null
@@ -1,29 +0,0 @@
-#!/usr/bin/python3
-
-import configparser
-import os
-import shutil
-import sys
-
-config = configparser.RawConfigParser()
-config.read(os.path.dirname(os.path.realpath(__file__)) + "/../../conf/config")
-
-template_path = config.get('serve', 'template-path')
-git_update_hook = config.get('serve', 'git-update-hook')
-
-def die(msg):
-    sys.stderr.write("%s\n" % (msg))
-    exit(1)
-
-if os.path.exists(template_path):
-    shutil.rmtree(template_path)
-
-os.mkdir(template_path)
-os.chdir(template_path)
-os.mkdir("branches")
-os.mkdir("hooks")
-os.mkdir("info")
-os.symlink(git_update_hook, template_path + 'hooks/update')
-
-with open("description", 'w') as f:
-    f.write("Unnamed repository; push to update the description.\n")
diff --git a/scripts/git-integration/git-serve.py b/scripts/git-integration/git-serve.py
index de9cb59b..81222e8a 100755
--- a/scripts/git-integration/git-serve.py
+++ b/scripts/git-integration/git-serve.py
@@ -17,28 +17,23 @@ aur_db_user = config.get('database', 'user')
 aur_db_pass = config.get('database', 'password')
 aur_db_socket = config.get('database', 'socket')
 
-repo_base_path = config.get('serve', 'repo-base')
+repo_path = config.get('serve', 'repo-path')
 repo_regex = config.get('serve', 'repo-regex')
 git_shell_cmd = config.get('serve', 'git-shell-cmd')
 ssh_cmdline = config.get('serve', 'ssh-cmdline')
 template_path = config.get('serve', 'template-path')
 
-def repo_path_validate(path):
-    if not path.startswith(repo_base_path):
-        return False
-    if path.endswith('.git'):
-        repo = path[len(repo_base_path):-4]
-    elif path.endswith('.git/'):
-        repo = path[len(repo_base_path):-5]
-    else:
-        return False
-    return re.match(repo_regex, repo)
-
-def repo_path_get_pkgbase(path):
-    pkgbase = path.rstrip('/').rpartition('/')[2]
-    if pkgbase.endswith('.git'):
-        pkgbase = pkgbase[:-4]
-    return pkgbase
+def pkgbase_exists(pkgbase):
+    db = mysql.connector.connect(host=aur_db_host, user=aur_db_user,
+                                 passwd=aur_db_pass, db=aur_db_name,
+                                 unix_socket=aur_db_socket)
+    cur = db.cursor()
+
+    cur.execute("SELECT COUNT(*) FROM PackageBases WHERE Name = %s ",
+                [pkgbase])
+
+    db.close()
+    return (cur.fetchone()[0] > 0)
 
 def list_repos(user):
     db = mysql.connector.connect(host=aur_db_host, user=aur_db_user,
@@ -57,19 +52,17 @@ def list_repos(user):
         print((' ' if row[1] else '*') + row[0])
     db.close()
 
-def setup_repo(repo, user):
-    if not re.match(repo_regex, repo):
-        die('%s: invalid repository name: %s' % (action, repo))
+def setup_repo(pkgbase, user):
+    if not re.match(repo_regex, pkgbase):
+        die('%s: invalid repository name: %s' % (action, pkgbase))
+    if pkgbase_exists(pkgbase):
+        die('%s: package base already exists: %s' % (action, pkgbase))
 
     db = mysql.connector.connect(host=aur_db_host, user=aur_db_user,
                                  passwd=aur_db_pass, db=aur_db_name,
                                  unix_socket=aur_db_socket)
     cur = db.cursor()
 
-    cur.execute("SELECT COUNT(*) FROM PackageBases WHERE Name = %s ", [repo])
-    if cur.fetchone()[0] > 0:
-        die('%s: package base already exists: %s' % (action, repo))
-
     cur.execute("SELECT ID FROM Users WHERE Username = %s ", [user])
     userid = cur.fetchone()[0]
     if userid == 0:
@@ -77,7 +70,7 @@ def setup_repo(repo, user):
 
     cur.execute("INSERT INTO PackageBases (Name, SubmittedTS, ModifiedTS, " +
                 "SubmitterUID, MaintainerUID) VALUES (%s, UNIX_TIMESTAMP(), " +
-                "UNIX_TIMESTAMP(), %s, %s)", [repo, userid, userid])
+                "UNIX_TIMESTAMP(), %s, %s)", [pkgbase, userid, userid])
     pkgbase_id = cur.lastrowid
 
     cur.execute("INSERT INTO CommentNotify (PackageBaseID, UserID) " +
@@ -86,8 +79,11 @@ def setup_repo(repo, user):
     db.commit()
     db.close()
 
-    repo_path = repo_base_path + '/' + repo + '.git/'
-    pygit2.init_repository(repo_path, True, 48, template_path=template_path)
+    repo = pygit2.Repository(repo_path)
+    repo.create_reference('refs/heads/' + pkgbase,
+                          'refs/namespaces/' + pkgbase + '/refs/heads/master')
+    repo.create_reference('refs/namespaces/' + pkgbase + '/HEAD',
+                          'refs/namespaces/' + pkgbase + '/refs/heads/master')
 
 def check_permissions(pkgbase, user):
     db = mysql.connector.connect(host=aur_db_host, user=aur_db_user,
@@ -125,19 +121,25 @@ action = cmdargv[0]
 if action == 'git-upload-pack' or action == 'git-receive-pack':
     if len(cmdargv) < 2:
         die_with_help("%s: missing path" % (action))
-    path = repo_base_path.rstrip('/') + cmdargv[1]
-    if not repo_path_validate(path):
+
+    path = cmdargv[1].rstrip('/')
+    if not path.startswith('/') or not path.endswith('.git'):
         die('%s: invalid path: %s' % (action, path))
-    pkgbase = repo_path_get_pkgbase(path)
-    if not os.path.exists(path):
+    pkgbase = path[1:-4]
+    if not re.match(repo_regex, pkgbase):
+        die('%s: invalid repository name: %s' % (action, repo))
+
+    if not pkgbase_exists(pkgbase):
         setup_repo(pkgbase, user)
+
     if action == 'git-receive-pack':
         if not check_permissions(pkgbase, user):
             die('%s: permission denied: %s' % (action, user))
+
     os.environ["AUR_USER"] = user
-    os.environ["AUR_GIT_DIR"] = path
     os.environ["AUR_PKGBASE"] = pkgbase
-    cmd = action + " '" + path + "'"
+    os.environ["GIT_NAMESPACE"] = pkgbase
+    cmd = action + " '" + repo_path + "'"
     os.execl(git_shell_cmd, git_shell_cmd, '-c', cmd)
 elif action == 'list-repos':
     if len(cmdargv) > 1:
diff --git a/scripts/git-integration/git-update.py b/scripts/git-integration/git-update.py
index 161d42fe..2c159124 100755
--- a/scripts/git-integration/git-update.py
+++ b/scripts/git-integration/git-update.py
@@ -19,6 +19,8 @@ aur_db_user = config.get('database', 'user')
 aur_db_pass = config.get('database', 'password')
 aur_db_socket = config.get('database', 'socket')
 
+repo_path = config.get('serve', 'repo-path')
+
 def extract_arch_fields(pkginfo, field):
     values = []
 
@@ -166,12 +168,11 @@ sha1_new = sys.argv[3]
 
 user = os.environ.get("AUR_USER")
 pkgbase = os.environ.get("AUR_PKGBASE")
-git_dir = os.environ.get("AUR_GIT_DIR")
 
 if refname != "refs/heads/master":
     die("pushing to a branch other than master is restricted")
 
-repo = pygit2.Repository(git_dir)
+repo = pygit2.Repository(repo_path)
 walker = repo.walk(sha1_new, pygit2.GIT_SORT_TOPOLOGICAL)
 if sha1_old != "0000000000000000000000000000000000000000":
     walker.hide(sha1_old)
@@ -245,6 +246,6 @@ db.close()
 
 pkglist = list(srcinfo.GetPackageNames())
 if len(pkglist) > 0:
-    with open(git_dir + '/description', 'w') as f:
+    with open(repo_path + '/description', 'w') as f:
         pkginfo = srcinfo.GetMergedPackage(pkglist[0])
         f.write(pkginfo['pkgdesc'])
diff --git a/scripts/git-integration/init-repos.py b/scripts/git-integration/init-repos.py
deleted file mode 100755
index 5c4fcfe2..00000000
--- a/scripts/git-integration/init-repos.py
+++ /dev/null
@@ -1,50 +0,0 @@
-#!/usr/bin/python3
-
-import configparser
-import mysql.connector
-import os
-import pygit2
-import re
-import shlex
-import sys
-
-config = configparser.RawConfigParser()
-config.read(os.path.dirname(os.path.realpath(__file__)) + "/../../conf/config")
-
-aur_db_host = config.get('database', 'host')
-aur_db_name = config.get('database', 'name')
-aur_db_user = config.get('database', 'user')
-aur_db_pass = config.get('database', 'password')
-aur_db_socket = config.get('database', 'socket')
-
-repo_base_path = config.get('serve', 'repo-base')
-repo_regex = config.get('serve', 'repo-regex')
-template_path = config.get('serve', 'template-path')
-
-def die(msg):
-    sys.stderr.write("%s\n" % (msg))
-    exit(1)
-
-db = mysql.connector.connect(host=aur_db_host, user=aur_db_user,
-                             passwd=aur_db_pass, db=aur_db_name,
-                             unix_socket=aur_db_socket)
-cur = db.cursor()
-
-cur.execute("SELECT Name FROM PackageBases")
-repos = [row[0] for row in cur]
-db.close()
-
-for repo in repos:
-    if not re.match(repo_regex, repo):
-        die('invalid repository name: %s' % (repo))
-
-i = 1
-n = len(repos)
-
-for repo in repos:
-    print("[%s/%d] %s" % (str(i).rjust(len(str(n))), n, repo))
-
-    repo_path = repo_base_path + '/' + repo + '.git/'
-    pygit2.init_repository(repo_path, True, 48, template_path=template_path)
-
-    i += 1