Allow to configure the compress algorithm in mysqldump and pg_dump

CHANGELOG.md

@@ -10,6 +10,14 @@ This project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html
 ### Breaking changes
 
 
+## 12.6.0 2025-09-22
+
+### Compatible changes
+
+- `geordi dump`: Allow to forward the compression option to the underlying `dumple` command, e.g. `geordi dump --compress=zstd:3` (for PostgreSQL) or `geordi dump --compress` (for MySQL).
+- `dumple`: Allow to specify a compression algorithm for PostgreSQL, e.g. `dumple --compress=zstd:3`. The already supported compression for MySQL `dumple --compress` is kept untouched.
+
+
 ## 12.5.0 2025-09-09
 
 ### Compatible changes

Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    geordi (12.5.0)
+    geordi (12.6.0)
       highline
       thor (~> 1)
 

README.md

@@ -65,8 +65,8 @@ Example: `geordi chromedriver_update`
 This command will find and install the matching chromedriver for the currently
 installed Chrome.
 
-Setting `auto_update_chromedriver` to `true` in your global Geordi config file 
-(`~/.config/geordi/global.yml`), will automatically update chromedriver before 
+Setting `auto_update_chromedriver` to `true` in your global Geordi config file
+(`~/.config/geordi/global.yml`), will automatically update chromedriver before
 cucumber tests if a newer chromedriver version is available.
 
 
@@ -98,8 +98,8 @@ servers. When passed a number, directly connects to the selected server.
 IRB flags can be given as `irb_flags: '...'` in the global or local Geordi config file
 (`~/.config/geordi/global.yml` / `./.geordi.yml`). If you define irb_flags in both files, the local config file will be
 used. For IRB >=1.2 in combination with Ruby <3 geordi automatically sets the `--nomultiline` flag, to prevent slow
-pasting. You can override this behavior by setting `--multiline` in the global config file or by defining `irb_flags` 
-in the local config file. The latter will always turn off the automatic behavior, even if you don't set any values for 
+pasting. You can override this behavior by setting `--multiline` in the global config file or by defining `irb_flags`
+in the local config file. The latter will always turn off the automatic behavior, even if you don't set any values for
 the irb_flags key.
 
 **Options**
@@ -191,8 +191,8 @@ and offer to delete them. Excluded are databases that are whitelisted. This come
 in handy when you're keeping your currently active projects in the whitelist files
 and perform regular housekeeping with Geordi.
 
-Per default, Geordi will try to connect to the databases as a local user without 
-password authorization.  
+Per default, Geordi will try to connect to the databases as a local user without
+password authorization.
 
 Geordi will ask for confirmation before actually dropping databases and will
 offer to edit the whitelist instead.
@@ -234,6 +234,7 @@ not match, please issue separate commands for dumping (`dump -d`) and sourcing
 **Options**
 - `-l, --load=[DUMP_FILE]`: Load a dump
 - `-d, --database=NAME`: Target database, if there are multiple databases
+- `-c, --compress=[ALGORITHM]`: Compress the dump file (default for PSQL)
 
 
 ### `geordi help [COMMAND]`
@@ -341,7 +342,7 @@ Run all employed tests.
 When running `geordi tests` without any arguments, all unit tests, rspec specs
 and cucumber features will be run.
 
-When passing file paths or directories as arguments, Geordi will forward them to `rspec` and `cucumber`. 
+When passing file paths or directories as arguments, Geordi will forward them to `rspec` and `cucumber`.
 All rspec specs and cucumber features matching the given paths will be run.
 
 
@@ -389,7 +390,9 @@ Stores a timestamped database dump for the given Rails environment in `~/dumps`:
 
 **Options**
 - `-i`: Print disk usage of `~/dumps`
-- `--compress`: After dumping, run gzip to compress the dump in place
+- `--fail-gently`: On error, do not crash but print a warning and exit(0)
+- `--for-download`: Dump to `~/dumps/dump_for_download.dump`
+- `--compress`: Compress the dump (default for PostgreSQL) and optionally set the compression algorithm (only available for PostgreSQL)
 
 
 Contributing

exe/dumple

@@ -50,7 +50,7 @@ def cd_to_project_root(fail_gently)
   end
 end
 
-def dump_command(dump_file, config)
+def dump_command(dump_file, config, compress)
   host = config['host']
   port = config['port']
 
@@ -79,6 +79,7 @@ def dump_command(dump_file, config)
     command << " pg_dump #{config['database']}"
     command << " --clean"
     command << " --format=custom"
+    command << " --compress=#{compress}" if compress.is_a?(String)
     command << " --file=#{dump_file}"
     command << " --username=\"#{config['username']}\""
     command << " --host=#{host}" if host
@@ -124,7 +125,7 @@ def prepare_dump_path(config)
     run "chmod 700 #{DUMPS_DIR}"
   end
 
-  if ARGV.include? '--for_download'
+  if ARGV.include?('--for_download') || ARGV.include?('--for-download')
     "#{DUMPS_DIR}/dump_for_download.dump"
   else
     "#{DUMPS_DIR}/#{config['database']}_#{Time.now.strftime("%Y%m%d_%H%M%S")}.dump"
@@ -133,21 +134,32 @@ end
 
 begin
   fail_gently = ARGV.include?("--fail-gently")
-  compress = ARGV.include?("--compress")
-  environment, database = ARGV.reject { |arg| arg[0].chr == '-' }
+  compress = ARGV.find do |argument|
+    if argument == '--compress'
+      break true
+    elsif argument.start_with?('--compress=')
+      break argument.split('=').last
+    end
+  end
+  environment, database = ARGV.reject { |argument| argument.start_with?('-') }
 
   cd_to_project_root(fail_gently)
   config = find_database_config(DB_CONFIG_PATH, environment, database)
   dump_path = prepare_dump_path(config)
 
   # Dump!
   given_database = database ? %(#{database} ) : ""
-  command = dump_command(dump_path, config)
+  command = dump_command(dump_path, config, compress)
   puts "> Dumping #{given_database}database for \"#{environment}\" environment ..."
   run command or raise "x Creating the dump failed."
   run "chmod 600 #{dump_path}"
 
-  if compress
+  if config['adapter'] == 'mysql' && compress.is_a?(String)
+    puts "> Cannot compress a MySQL dump with #{compress}, falling back to gzip."
+  end
+
+  # For PostgreSQL, #dump_command will do the compression. MySQL needs manual compression.
+  if config['adapter'] == 'mysql' && compress
     puts "> Compressing the dump ..."
     # gzip compresses in place
     compress_success = run "gzip #{dump_path}"

features/dump.feature

@@ -17,9 +17,7 @@ Feature: The dump command
 
   Scenario: Creating a dump of a remote database
     Given a file named "Capfile" with "Capfile exists"
-    And a file named "config/deploy.rb" with:
-    """
-    """
+    And a file named "config/deploy.rb" with "deploy.rb exists"
     And a file named "config/deploy/staging.rb" with:
     """
     set :rails_env, 'staging'
@@ -32,7 +30,7 @@ Feature: The dump command
     When I run `geordi dump staging`
     Then the output should contain "# Dumping the database of staging"
       And the output should contain "> Connecting to www.example.com"
-      And the output should contain "Util.run! ssh, [email protected], -t, cd /var/www/example.com/current && bash --login -c 'dumple staging --for_download'"
+      And the output should contain "Util.run! ssh, [email protected], -t, cd /var/www/example.com/current && bash --login -c 'dumple staging --for-download'"
       And the output should contain "> Downloading remote dump to tmp/staging.dump"
       # Omitting the absolute path in this regex (.*)
       And the output should match:
@@ -62,7 +60,7 @@ Feature: The dump command
 
     When I run `geordi dump staging --load`
     Then the output should contain "# Dumping the database of staging"
-      And the output should contain "Util.run! ssh, [email protected], -t, cd /var/www/example.com/current && bash --login -c 'dumple staging --for_download'"
+      And the output should contain "Util.run! ssh, [email protected], -t, cd /var/www/example.com/current && bash --login -c 'dumple staging --for-download'"
       And the output should contain "> Dumped the staging database to tmp/staging.dump"
 
       # Loading the dump
@@ -85,7 +83,7 @@ Feature: The dump command
 
     When I run `geordi dump staging --database primary`
     Then the output should contain "# Dumping the database of staging (primary database)"
-      And the output should contain "Util.run! ssh, [email protected], -t, cd /var/www/example.com/current && bash --login -c 'dumple staging primary --for_download'"
+      And the output should contain "Util.run! ssh, [email protected], -t, cd /var/www/example.com/current && bash --login -c 'dumple staging primary --for-download'"
       And the output should contain "> Dumped the primary staging database to tmp/staging.dump"
 
 
@@ -110,7 +108,7 @@ Feature: The dump command
 
     When I run `geordi dump staging --database primary --load`
     Then the output should contain "# Dumping the database of staging (primary database)"
-      And the output should contain "Util.run! ssh, [email protected], -t, cd /var/www/example.com/current && bash --login -c 'dumple staging primary --for_download'"
+      And the output should contain "Util.run! ssh, [email protected], -t, cd /var/www/example.com/current && bash --login -c 'dumple staging primary --for-download'"
       And the output should contain "> Dumped the primary staging database to tmp/staging.dump"
 
       # Loading the dump
@@ -176,6 +174,7 @@ Feature: The dump command
         And the output should contain "Util.run! dropdb --if-exists test-other && createdb test-other && pg_restore --no-owner --no-acl --dbname=test-other tmp/production.dump"
       But the output should not contain "Sourcing dump into the test-primary db"
 
+
   Scenario: Sourcing a dump into one of multiple databases without specifying a db
     Given a file named "tmp/production.dump" with "some content"
       And a file named "config/database.yml" with:
@@ -193,3 +192,32 @@ Feature: The dump command
         And the output should contain "Source file: tmp/production.dump"
         And the output should contain "Util.run! dropdb --if-exists test-primary && createdb test-primary && pg_restore --no-owner --no-acl --dbname=test-primary tmp/production.dump"
       But the output should not contain "Sourcing dump into the test-other db"
+
+
+  Scenario: Enforcing compression
+    When I run `geordi dump --compress`
+    Then the output should contain "Util.run! dumple development --compress"
+      And the output should contain "Successfully dumped the development database"
+
+
+  Scenario: Setting a custom compression algorithm
+    When I run `geordi dump --compress=zstd:3`
+    Then the output should contain "Util.run! dumple development --compress=zstd:3"
+    And the output should contain "Successfully dumped the development database"
+
+
+  Scenario: Setting a custom compression algorithm for a remote target
+    Given a file named "Capfile" with "Capfile exists"
+    And a file named "config/deploy.rb" with "deploy.rb exists"
+    And a file named "config/deploy/staging.rb" with:
+    """
+    set :rails_env, 'staging'
+    set :deploy_to, '/var/www/example.com'
+    set :user, 'user'
+
+    server 'www.example.com'
+    """
+
+    When I run `geordi dump staging --compress=zstd:3`
+    Then the output should contain "Util.run! ssh, [email protected], -t, cd /var/www/example.com/current && bash --login -c 'dumple staging --compress=zstd:3 --for-download'"
+      And the output should contain "> Dumped the staging database to tmp/staging.dump"

features/dumple.feature

@@ -1,6 +1,6 @@
 Feature: Creating Rails database dumps with the "dumple" script
 
-  The --for_download option generates a deterministic file name.
+  The --for-download option generates a deterministic file name.
 
   Scenario: Execution outside of a Rails application
     When I run `dumple development`
@@ -20,7 +20,7 @@ Feature: Creating Rails database dumps with the "dumple" script
     Since dumple won't actually dump in tests, we prepare the dump file in advance.
     """
 
-    When I run `dumple development --for_download`
+    When I run `dumple development --for-download`
     Then the output should match /Dumping database for "development" environment/
       And the output should contain "> Dumped to /home/"
       And the output should contain "/geordi/tmp/aruba/dumps/dump_for_download.dump (0 KB)"
@@ -36,7 +36,7 @@ Feature: Creating Rails database dumps with the "dumple" script
     """
     And a mocked home directory
 
-    When I run `dumple development --for_download`
+    When I run `dumple development --for-download`
     Then the output should match %r<system mysqldump -u"user" -p"password"  -r /home/.*/aruba/dumps/dump_for_download.dump --single-transaction --quick -hlocalhost>
 
 
@@ -50,7 +50,7 @@ Feature: Creating Rails database dumps with the "dumple" script
     """
     And a mocked home directory
 
-    When I run `dumple development --for_download`
+    When I run `dumple development --for-download`
     Then the output should match %r<system PGPASSWORD="password" pg_dump  --clean --format=custom --file=/home/.*/aruba/dumps/dump_for_download.dump --username="user">
 
 
@@ -133,3 +133,60 @@ Feature: Creating Rails database dumps with the "dumple" script
 
     When I run `dumple development sub`
     Then the output should match /Could not select "sub" database in a single-db environment/
+
+
+  Scenario: Instructing MySQL to compress the dump
+    Given a file named "config/database.yml" with:
+    """
+    development:
+      adapter: mysql
+      database: geordi_development
+    """
+    And a mocked home directory
+    And a file named "dumps/dump_for_download.dump.gz" with:
+    """
+    Since dumple won't actually dump in tests, we prepare the dump file in advance.
+    """
+
+    When I run `dumple development --for-download --compress`
+    Then the output should match /Dumping database for "development" environment/
+      And the output should contain "> Compressing the dump ..."
+      And the output should contain "system gzip"
+      And the output should contain "> Dumped to /home/"
+      And the output should contain "/geordi/tmp/aruba/dumps/dump_for_download.dump.gz (0 KB)"
+
+
+  Scenario: Instructing MySQL to compress the dump with a custom compression algorithm shows a warning
+    Given a file named "config/database.yml" with:
+    """
+    development:
+      adapter: mysql
+      database: geordi_development
+    """
+    And a mocked home directory
+    And a file named "dumps/dump_for_download.dump.gz" with:
+    """
+    Since dumple won't actually dump in tests, we prepare the dump file in advance.
+    """
+
+    When I run `dumple development --for-download --compress=zstd:3`
+    Then the output should match /Dumping database for "development" environment/
+      And the output should contain "> Cannot compress a MySQL dump with zstd:3, falling back to gzip."
+      And the output should contain "> Compressing the dump ..."
+      And the output should contain "system gzip"
+      And the output should contain "> Dumped to /home/"
+      And the output should contain "/geordi/tmp/aruba/dumps/dump_for_download.dump.gz (0 KB)"
+
+
+  Scenario: Setting a custom compression algorithm for PostgreSQL
+    Given a file named "config/database.yml" with:
+    """
+    development:
+      adapter: postgres
+      username: user
+      password: password
+    """
+    And a mocked home directory
+
+    When I run `dumple development --for-download --compress=zstd:3`
+    Then the output should match %r<system PGPASSWORD="password" pg_dump  --clean --format=custom --compress=zstd:3 --file=/home/.*/aruba/dumps/dump_for_download.dump --username="user">

lib/geordi/commands/dump.rb

@@ -26,11 +26,11 @@
 
 option :load, aliases: '-l', type: :string, desc: 'Load a dump', banner: '[DUMP_FILE]'
 option :database, aliases: '-d', type: :string, desc: 'Target database, if there are multiple databases', banner: 'NAME'
+option :compress, aliases: '-c', type: :string, desc: 'Compress the dump file (default for PSQL)', banner: '[ALGORITHM]'
 
 def dump(target = nil, *_args)
   require 'geordi/dump_loader'
   require 'geordi/remote'
-  database = options[:database] ? "#{options[:database]} " : ''
 
   if target.nil? # Local …
     if options.load # … dump loading
@@ -46,14 +46,17 @@ def dump(target = nil, *_args)
 
     else # … dump creation
       Interaction.announce 'Dumping the development database'
-      Util.run!("dumple development #{database}")
+      Util.run!(Util.dumple_command('development', options))
+
+      database = "#{options[:database]} " if options[:database]
       Interaction.success "Successfully dumped the #{database}development database."
     end
 
   else # Remote dumping …
-    database_label = options[:database] ? " (#{database}database)" : ""
+    database_label = target.dup
+    database_label << " (#{options[:database]} database)" if options[:database]
 
-    Interaction.announce "Dumping the database of #{target}#{database_label}"
+    Interaction.announce "Dumping the database of #{database_label}"
     dump_path = Geordi::Remote.new(target).dump(options)
 
     if options.load # … and dump loading
@@ -65,7 +68,7 @@ def dump(target = nil, *_args)
       Util.run! "rm #{dump_path}"
       Interaction.note "Dump file removed"
 
-      Interaction.success "Your #{loader.config['database']} database has now the data of #{target}#{database_label}."
+      Interaction.success "Your #{loader.config['database']} database has now the data of #{database_label}."
     end
   end
 

lib/geordi/remote.rb

@@ -33,11 +33,9 @@ def select_server
     end
 
     def dump(options = {})
-      database = options[:database] ? " #{options[:database]}" : ''
       # Generate dump on the server
-      shell options.merge({
-        remote_command: "dumple #{@config.env}#{database} --for_download",
-      })
+      dumple = Util.dumple_command(@config.env, options.merge(for_download: true))
+      shell(options.merge(remote_command: dumple))
 
       destination_directory = File.join(@config.root, 'tmp')
       FileUtils.mkdir_p destination_directory
@@ -48,6 +46,7 @@ def dump(options = {})
       server = @config.primary_server
       Util.run!("scp -C #{@config.user(server)}@#{server}:#{REMOTE_DUMP_PATH} #{destination_path}")
 
+      database = " #{options[:database]}" if options[:database]
       Interaction.success "Dumped the#{database} #{@stage} database to #{relative_destination}."
 
       destination_path