From 963dfbb0538a21f0e834dac3395a2c89fd246a89 Mon Sep 17 00:00:00 2001
From: Patryk Obara <dreamer.tan@gmail.com>
Date: Sun, 15 Sep 2019 00:12:44 +0200
Subject: [PATCH] Add initial version of import script

The script creates new Git repository with complete SVN history of
DOSBox project, using correct authorship information and importing tags
as real Git tags.
---
 scripts/import-from-svn/.gitignore         |   1 +
 scripts/import-from-svn/copied-from.xslt   |  11 ++
 scripts/import-from-svn/import-from-svn.sh | 122 +++++++++++++++++++++
 scripts/import-from-svn/svn-dosbox-authors |   9 ++
 4 files changed, 143 insertions(+)
 create mode 100644 scripts/import-from-svn/.gitignore
 create mode 100644 scripts/import-from-svn/copied-from.xslt
 create mode 100755 scripts/import-from-svn/import-from-svn.sh
 create mode 100644 scripts/import-from-svn/svn-dosbox-authors

diff --git a/scripts/import-from-svn/.gitignore b/scripts/import-from-svn/.gitignore
new file mode 100644
index 00000000..9d9c5001
--- /dev/null
+++ b/scripts/import-from-svn/.gitignore
@@ -0,0 +1 @@
+dosbox-git-svn-*
diff --git a/scripts/import-from-svn/copied-from.xslt b/scripts/import-from-svn/copied-from.xslt
new file mode 100644
index 00000000..407727ab
--- /dev/null
+++ b/scripts/import-from-svn/copied-from.xslt
@@ -0,0 +1,11 @@
+<?xml version="1.0" encoding="utf-8"?>
+<!--
+This file transforms svn log xml into path@rev representing copied-from information.
+You can use it with xsltproc(1): $ svn log -xml -v -l 1 | copied-from.xslt -
+-->
+<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+<xsl:output method="text"/>
+<xsl:template match="/">
+<xsl:value-of select="log/logentry/paths/path/@copyfrom-path"/>@<xsl:value-of select="log/logentry/paths/path/@copyfrom-rev"/>
+</xsl:template>
+</xsl:stylesheet>
diff --git a/scripts/import-from-svn/import-from-svn.sh b/scripts/import-from-svn/import-from-svn.sh
new file mode 100755
index 00000000..0e1448a9
--- /dev/null
+++ b/scripts/import-from-svn/import-from-svn.sh
@@ -0,0 +1,122 @@
+#!/bin/bash
+
+readonly svn_url=https://svn.code.sf.net/p/dosbox/code-0/dosbox
+
+echo_err () {
+	echo "$@" 1>&2
+}
+
+# SVN repository revision. By comparing repo revision we can detect if there
+# was any activity on any svn path.
+#
+svn_get_repo_revision () {
+	svn info "$svn_url" | grep "Revision:" | cut -d' ' -f2
+}
+
+# Example usage:
+#
+# svn_get_copied_from branches/0_74_3
+#
+svn_get_copied_from () {
+	local -r repo_path=$1
+	svn log -v --stop-on-copy --xml "$svn_url/$repo_path" \
+		| xsltproc copied-from.xslt -
+}
+
+# Full import takes ~40 minutes
+#
+git_svn_clone_dosbox () {
+	local -r repo_name=$1
+
+	git svn init \
+		--stdlayout \
+		"$svn_url" \
+		"$repo_name"
+
+	local -r authors_file=$PWD/svn-dosbox-authors
+	git -C "$repo_name" svn fetch \
+		--authors-file="$authors_file"
+}
+
+# Remove UUID of SVN server to avoid issues in case SourceForge will change
+# it's infrastructure.
+#
+git_rewrite_import_links () {
+	local -r repo_name=$1
+	git -C "$repo_name" filter-branch \
+		--msg-filter 'sed "s|git-svn-id: \([^ ]*\) .*|Imported-from: \1|"' \
+		-- --all
+}
+
+list_svn_branch_paths () {
+	echo trunk
+	for branch in $(svn ls "$svn_url/branches") ; do
+		echo "${branch///}"
+	done
+}
+
+list_svn_tag_paths () {
+	for branch in $(svn ls "$svn_url/tags") ; do
+		echo "${branch///}"
+	done
+}
+
+# Go through SVN "branches", that were not removed from SVN yet and create
+# Git branches out of them  using name prefix "svn/" in local repository.
+#
+name_active_branches () {
+	local -r repo_name=$1
+	for path in $(list_svn_branch_paths) ; do
+		git -C "$repo_name" branch "svn/$path" "origin/$path"
+	done
+}
+
+find_tagged_git_commit () {
+	git -C "$1" log \
+		--grep="$svn_copied_path" \
+		--branches=svn/* \
+		--format=%H
+}
+
+# Import SVN "tags" as proper Git tags
+#
+# Hopefully upstream never committed to tags…
+#
+import_svn_tagpaths_as_git_tags () {
+	local -r repo_name=$1
+	local svn_copied_path
+	local svn_tagged_commit
+	for path in $(list_svn_tag_paths) ; do
+		svn_copied_path=$(svn_get_copied_from "tags/$path")
+		svn_tagged_commit=$(find_tagged_git_commit "$repo_name")
+		git_tag_name="svn/$path"
+		echo "$path: $svn_copied_path -> $svn_tagged_commit -> $git_tag_name"
+		git -C "$repo_name" tag \
+			"$git_tag_name" \
+			"$svn_tagged_commit"
+	done
+}
+
+# Remove any unnecessary refs left behind in imported repo
+#
+cleanup () {
+	git -C "$1" checkout svn/trunk
+	git -C "$1" branch -D master
+}
+
+# main
+#
+main () {
+	readonly repo=dosbox-git-svn-$(svn_get_repo_revision)
+	if [ -e "$repo" ] ; then
+		echo_err "Repository '$repo' exists already."
+		exit 1
+	fi
+	git_svn_clone_dosbox "$repo"
+	git_rewrite_import_links "$repo"
+	name_active_branches "$repo"
+	import_svn_tagpaths_as_git_tags "$repo"
+	cleanup "$repo"
+}
+
+main "$@"
diff --git a/scripts/import-from-svn/svn-dosbox-authors b/scripts/import-from-svn/svn-dosbox-authors
new file mode 100644
index 00000000..a7cb6486
--- /dev/null
+++ b/scripts/import-from-svn/svn-dosbox-authors
@@ -0,0 +1,9 @@
+qbix79 = Peter Veenstra <qbix79@users.sourceforge.net>
+harekiet = Sjoerd van der Berg <harekiet@users.sourceforge.net>
+finsterr = Ulf Wohlers <finsterr@users.sourceforge.net>
+canadacow = Dean Beeler <canadacow@users.sourceforge.net>
+yot = Felix Jakschitsch <yot@users.sourceforge.net>
+c2woody = Sebastian Strohhäcker <c2woody@users.sourceforge.net>
+h-a-l-9000 = Ralf Grillenberger <h-a-l-9000@users.sourceforge.net>
+ripsaw8080 = ripsaw8080 <ripsaw8080@users.sourceforge.net>
+uid83799 = uid83799 <uid83799@users.sourceforge.net>