munola 0.5.2 → 0.6.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 634f7369a35be4b04097f3539bdf6c41b0419836242450170c07d0937de85c76
-  data.tar.gz: 916ec77f29fefc76fabb0229c10078677ba01407d02fa02e71dd6973933961a2
+  metadata.gz: 035c3780ee982f7ae5ec0dc3613ae0d09d4e7ae6fc1a6c3bbe7d446da45548b5
+  data.tar.gz: 7f17a853e60b770fe08a90f9f055f1891f1c3f7d7a59c188ba44e8e0744a3d10
 SHA512:
-  metadata.gz: 044d6b03723fe94d321d1024927e9691c82cdf5a763aecd9a64709b6c4e6ffab4cb92b1555a15e812e7647b496cd1852f7d5dfb7cb3e470a789a64b0851be7f5
-  data.tar.gz: 2dc4d4e98eb5d60d948913b0dd3c69bb93077f372e76b4eddc11e2d5f96091cf7e32c7c9ded1759c1e5b96f6d5783d11a87459bea58f95cf43d874b866b2c56d
+  metadata.gz: a50a5ba6d22316da864d6a1c15b5eee82a681e9e87d7de00590a211453ec7fa512cd6fae628ee94e800977f46a18e26201ffc5f4f98a8f7a537586e6fb7f96e7
+  data.tar.gz: 0c07b025c1f109e040f3af93de83be72e84b34de591956d95a1b91a25bd1daf7cbdaae9dbbc8fb49fc777d956dec5c0e64b010311e447dbafcf0cd20b8af325a

data/CHANGELOG.md

@@ -1,3 +1,38 @@
+## munola 0.6.1 (2026-08-24)
+
+munola drives channel release `0.4.6.1`, cut from the promoted current at
+upstream enola 0.4.6. The prefix walk fix is in the binary this gem fetches,
+so the expiry the 0.6.0 note named is met and Rubydex is on by default again.
+The channel still carries the exclusion filter, which upstream merged an hour
+after 0.4.6 was published, so an excluded document is skipped before its
+definitions are read: 440 seconds to 193 on a Rails monolith, facts identical.
+
+The 0.6.0 note said munola waits on Prism alone. On the Rails install path it
+did not. `munola init` reuses enola-rb's installer, which writes
+`mcp-arch.yaml` first, and munola's own writer yields to a file that already
+exists, so the config every Rails project got was the enola gem's, with both
+providers on, against a channel binary built before the fix. Two things follow.
+The config and the binary now agree, which is what closes it. And munola's
+writer has a test that reads munola's writer, rather than one that reads the
+file enola-rb wrote and reports on munola.
+
+## enola 0.5.3, enola-rb 0.5.3 and munola 0.6.0 (2026-08-24)
+
+Both providers are on again. enola-labs shipped the prefix walk fix in 0.4.6,
+which is the release the enola and enola-rb gems now drive, so the config they
+write carries Prism and Rubydex as it did before the hang was found.
+
+munola waits. Its binary is a fork release cut before that fix, so it stays on
+Prism alone until the channel is re-cut, and its version file says which
+upstream it is built on rather than implying the newest.
+
+munola-rb is gone, and was never published. The railtie and generator move
+into munola, which is why munola is 0.6.0: one gem for one channel, since the
+split that makes sense for enola and enola-rb, letting a Rubyist take the
+binary wrapper without the Rails layer, does not describe anyone's use of
+munola. Requiring munola still loads no Rails; the generator registers when
+the application has already loaded it.
+
 ## enola 0.5.2, enola-rb 0.5.2, munola 0.5.2 and munola-rb 0.5.2 (2026-08-23)
 
 Rubydex is off in the config these gems write, and named in a comment beside

data/README.md

@@ -4,10 +4,11 @@ Four pure-Ruby gems, one repository, for running
 [enola](https://github.com/enola-labs/enola), the architecture-graph tool,
 from Ruby. None of them carries a binary or compiles anything.
 
-Two channels, the same layering on both: `enola` wraps the released upstream
-binary and `enola-rb` adds the Rails layer on it; `munola` wraps the munola
-binary and `munola-rb` adds the Rails layer on that. The Rails code is written
-once, in `enola-rb`, and the munola gems add only what is munola's own.
+Two channels. `enola` wraps the released upstream binary and `enola-rb` adds
+the Rails layer on it, split so a Rubyist can take the wrapper without Rails.
+`munola` is one gem for its own channel, wrapper and Rails layer together,
+because nobody installs it to avoid Rails. The Rails code is written once, in
+`enola-rb`, and munola adds only what is munola's own.
 
 ## `enola`
 
@@ -109,14 +110,9 @@ publishes; `MUNOLA_BINARY=/path/to/enola` names one to drive instead, and
 every command says which answered. `munola` depends on `enola` at the same
 minor and on `enola-guides` 0.3.1 or later; requiring it loads no Rails.
 
-## `munola-rb`
+## munola under Rails
 
-The Rails layer over `munola`, what `enola-rb` is to `enola`. It depends on
-both at the same minor and adds one generator.
-
-```ruby
-gem "munola-rb"
-```
+munola carries its own generator, and the rake tasks are enola-rb's:
 
 ```sh
 bin/rails generate munola:install --tenant-column company_id
@@ -124,7 +120,8 @@ bin/rake enola:snapshot
 bin/rake enola:check
 ```
 
-The generator runs the same install `munola init` does. The rake tasks are
-enola-rb's and need no munola copy: `munola` installs its own resolver when it
-loads, so `enola:snapshot` and `enola:check` drive the munola binary in an app
-that has this gem, and the upstream binary in one that does not.
+The generator runs the same install `munola init` does. `munola` installs its
+own resolver when it loads, so `enola:snapshot` and `enola:check` drive the
+munola binary in an app that has this gem, and the upstream binary in one that
+does not. Requiring munola loads no Rails; the generator registers when the
+application has already loaded it.

data/lib/generators/munola/install/install_generator.rb

@@ -0,0 +1,19 @@
+# frozen_string_literal: true
+
+require "rails/generators"
+require "munola"
+
+module Munola
+  module Generators
+    class InstallGenerator < Rails::Generators::Base
+      desc "Writes enola/constraints/ and the munola recipe catalogue, binding what the tree shows a need for"
+      class_option :tenant_column, type: :string, default: nil, desc: "The column most tables carry; confirmed against db/schema.rb"
+
+      def install
+        Munola::Installer.new(destination_root, tenant_column: options[:tenant_column]).install.lines.each { |line| say line }
+      rescue Enola::Error => e
+        raise Thor::Error, "munola: #{e.message}"
+      end
+    end
+  end
+end

data/lib/munola/providers_config.rb

@@ -27,10 +27,7 @@ module Munola
       else
         lines << "  # prism: the enola gem on this machine carries no provider script; add one here to run it"
       end
-      lines += ["  # rubydex is off until the channel release carries the walk fix:",
-        "  # a reference spanning lines could be its own predecessor, so a tree with",
-        "  # vendored gems could snapshot without ever finishing. Uncomment to run it.",
-        "  # - name: rubydex", "  #   expected_version: \"#{RUBYDEX_VERSION}\""]
+      lines += ["  - name: rubydex", "    expected_version: \"#{RUBYDEX_VERSION}\""]
       "#{lines.join("\n")}\n"
     end
 

data/lib/munola/railtie.rb

@@ -0,0 +1,10 @@
+# frozen_string_literal: true
+
+module Munola
+  # The enola:* tasks come from the enola-rb railtie and drive whatever channel
+  # is installed; requiring munola installs its resolver, so under this gem
+  # they run the munola binary. Only the generator is munola's own.
+  class Railtie < Rails::Railtie
+    generators { require_relative "../generators/munola/install/install_generator" }
+  end
+end

data/lib/munola/version.rb

@@ -3,7 +3,7 @@
 module Munola
   # UPSTREAM_VERSION names the enola release this channel is built on,
   # CHANNEL_VERSION the fork release the gem fetches.
-  UPSTREAM_VERSION = "0.4.4"
-  CHANNEL_VERSION = "0.4.4.2"
-  VERSION = "0.5.2"
+  UPSTREAM_VERSION = "0.4.6"
+  CHANNEL_VERSION = "0.4.6.1"
+  VERSION = "0.6.1"
 end

data/lib/munola.rb

@@ -8,6 +8,10 @@ require_relative "munola/detector"
 require_relative "munola/providers_config"
 require_relative "munola/installer"
 require_relative "munola/cli"
+# munola is a wrapper first. enola-rb requires rails because it is the Rails
+# layer; munola only registers its generator when the application has already
+# loaded Rails, so driving the binary alone costs nothing.
+require_relative "munola/railtie" if defined?(Rails::Railtie)
 
 Enola.channel = Munola::CHANNEL
 Enola.resolver_factory = -> { Munola::Resolver.new(channel: Munola::CHANNEL) }

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: munola
 version: !ruby/object:Gem::Version
-  version: 0.5.2
+  version: 0.6.1
 platform: ruby
 authors:
 - Muhamed Isabegovic
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-08-23 00:00:00.000000000 Z
+date: 2026-08-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: enola
@@ -16,28 +16,28 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.5.0
+        version: 0.5.3
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.5.0
+        version: 0.5.3
 - !ruby/object:Gem::Dependency
   name: enola-rb
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.5.0
+        version: 0.5.3
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 0.5.0
+        version: 0.5.3
 - !ruby/object:Gem::Dependency
   name: enola-guides
   requirement: !ruby/object:Gem::Requirement
@@ -56,8 +56,8 @@ description: |
   The binary this gem fetches is not an enola-labs release. It is a build
   cut from https://github.com/misabegovic/enola, a fork of enola, whose
   release notes name what differs from upstream release by release; this
-  gem drives channel release 0.4.4.2, built on enola
-  0.4.4. Everything else is the enola gem's wrapper,
+  gem drives channel release 0.4.6.1, built on enola
+  0.4.6. Everything else is the enola gem's wrapper,
   unchanged: fetched on first use, verified against the checksums the
   release publishes, cached per user, every command and exit code
   forwarded. Use the enola gem to run stock upstream instead.
@@ -66,7 +66,7 @@ description: |
   enola-guides gem carries (Ember, data ownership, API boundaries,
   background work, a tenant foreign key) into the project, binds the
   recipes its tree justifies, and turns both Ruby providers on by default.
-  The Rails generator lives in munola-rb. Offered upstream where it fits;
+  Under Rails it adds `rails generate munola:install`; the enola:* rake tasks come from enola-rb and drive munola's binary. Offered upstream where it fits;
   no binary here, nothing compiled.
 email:
 - m.isabegovic@hotmail.com
@@ -79,6 +79,7 @@ files:
 - LICENSE
 - README.md
 - exe/munola
+- lib/generators/munola/install/install_generator.rb
 - lib/munola.rb
 - lib/munola/catalogue.rb
 - lib/munola/channel.rb
@@ -86,6 +87,7 @@ files:
 - lib/munola/detector.rb
 - lib/munola/installer.rb
 - lib/munola/providers_config.rb
+- lib/munola/railtie.rb
 - lib/munola/version.rb
 homepage: https://github.com/misabegovic/enola-rb
 licenses: