From ce795d52adcf670656b661e0899a4b554c01b00a Mon Sep 17 00:00:00 2001
From: Romuald Conty <romuald@opus-codium.fr>
Date: Tue, 5 Oct 2021 21:45:24 +0200
Subject: [PATCH] GitService: Improve hostname extraction

This commit adds support for more repository notations and provides unit
tests.
---
 lib/modulesync/git_service.rb            | 33 ++++++++++++++++---
 lib/modulesync/source_code.rb            | 10 ++++--
 spec/unit/modulesync/git_service_spec.rb | 40 ++++++++++++++++++++++++
 3 files changed, 77 insertions(+), 6 deletions(-)

diff --git a/lib/modulesync/git_service.rb b/lib/modulesync/git_service.rb
index e653c7ff..cee3506a 100644
--- a/lib/modulesync/git_service.rb
+++ b/lib/modulesync/git_service.rb
@@ -89,12 +89,11 @@ def self.endpoint_for(sourcecode:, type:)
     end
 
     # This method attempts to guess the git service endpoint based on remote and type
-    # FIXME: It only supports "git@hostname:repo_path" scheme
     def self.guess_endpoint_from(remote:, type:)
-      pattern = /^git@(.*):(.*)(\.git)*$/
-      return nil unless remote.match?(pattern)
+      hostname = extract_hostname(remote)
+      return nil if hostname.nil?
 
-      endpoint = remote.sub(pattern, 'https://\1')
+      endpoint = "https://#{hostname}"
       endpoint += '/api/v4' if type == :gitlab
       endpoint
     end
@@ -121,5 +120,31 @@ def self.token_for(sourcecode:, type:)
 
       token
     end
+
+    # This method extracts hostname from URL like:
+    #
+    # - ssh://[user@]host.xz[:port]/path/to/repo.git/
+    # - git://host.xz[:port]/path/to/repo.git/
+    # - [user@]host.xz:path/to/repo.git/
+    # - http[s]://host.xz[:port]/path/to/repo.git/
+    # - ftp[s]://host.xz[:port]/path/to/repo.git/
+    #
+    # Returns nil if
+    # - /path/to/repo.git/
+    # - file:///path/to/repo.git/
+    # - any invalid URL
+    def self.extract_hostname(url)
+      #pattern = /^((?<user>.*)@)*(?<hostname>.*):(?<path>[a-zA-Z].*)*$/
+      return nil if url.start_with?('/') || url.start_with?('file://') # local path (e.g. file:///path/to/repo)
+
+      unless url.start_with?(/[a-z]+:\/\//) # SSH notation does not contain protocol (e.g. user@server:path/to/repo/)
+        pattern = /^(.*@)?(?<hostname>[\w|.]*):(.*)$/ # SSH path (e.g. user@server:repo)
+        return url.match(pattern)[:hostname] if url.match?(pattern)
+      end
+
+      return URI.parse(url).host
+    rescue URI::InvalidURIError => e
+      return nil
+    end
   end
 end
diff --git a/lib/modulesync/source_code.rb b/lib/modulesync/source_code.rb
index 10d72633..248bd420 100644
--- a/lib/modulesync/source_code.rb
+++ b/lib/modulesync/source_code.rb
@@ -48,9 +48,15 @@ def path(*parts)
       File.join(working_directory, *parts)
     end
 
+    def git_service
+      @git_service ||= GitService.instantiate(**git_service_configuration)
+    end
+
+    def git_service_configuration
+      @git_service_configuration ||= GitService.configuration_for(sourcecode: self)
+    end
+
     def open_pull_request
-      git_service_options = GitService.configuration_for(sourcecode: self)
-      git_service = GitService.instantiate(**git_service_options)
       git_service.open_pull_request(
         repo_path: repository_path,
         namespace: repository_namespace,
diff --git a/spec/unit/modulesync/git_service_spec.rb b/spec/unit/modulesync/git_service_spec.rb
index d4b115d8..9f372363 100644
--- a/spec/unit/modulesync/git_service_spec.rb
+++ b/spec/unit/modulesync/git_service_spec.rb
@@ -1,3 +1,4 @@
+require 'modulesync'
 require 'modulesync/git_service'
 
 describe ModuleSync::GitService do
@@ -164,4 +165,43 @@
       end
     end
   end
+
+  RSpec.shared_examples 'hostname_extractor' do |url, hostname|
+    context "with '#{url}' URL" do
+      subject { ModuleSync::GitService.extract_hostname(url) }
+      it "should extract '#{hostname}' as hostname" do
+        expect(subject).to eq(hostname)
+      end
+    end
+  end
+
+  context '#extract_hostname' do
+    [
+      %w[ssh://user@host.xz:4444/path/to/repo.git/ host.xz],
+      %w[ssh://user@host.xz:/path/to/repo.git/     host.xz],
+      %w[ssh://host.xz/path/to/repo.git/           host.xz],
+
+      %w[git://host.xz/path/to/repo.git/           host.xz],
+      %w[git://host.xz/path/to/repo/               host.xz],
+      %w[git://host.xz/path/to/repo                host.xz],
+
+      %w[user@host.xz:path/to/repo.git/            host.xz],
+      %w[user@host.xz:path/to/repo.git             host.xz],
+      %w[user@host.xz:path/to/repo                 host.xz],
+      %w[host.xz:path/to/repo.git/                 host.xz],
+
+      %w[https://host.xz:8443/path/to/repo.git/    host.xz],
+      %w[https://host.xz/path/to/repo.git/         host.xz],
+
+      %w[ftp://host.xz/path/to/repo/               host.xz],
+
+      ['/path/to/repo.git/',                       nil],
+
+      ['file:///path/to/repo.git/',                nil],
+
+      ['something-invalid',                        nil],
+    ].each do |url, hostname|
+      it_should_behave_like 'hostname_extractor', url, hostname
+    end
+  end
 end