Welcome to phenaproxima who will be giving this module some much-needed love over the next couple of months! :)

Here's our rough battleplan/timeline, to provide transparency on what's being worked on when. In general, we're prioritizing the stuff that was already in the project and simply needs to be updated first, and putting off APIs that are less baked (entity/field, theme) until later on. Each of these bullets will eventually be updated with links to issues for each given API.

We plan to work in 2 week "sprints" and each sprint having a particular couple of goals. It remains to be seen if these goals are realistic or not. :) So please note that this schedule is very subject to change, and we'll be updating it every so often as reality hits. :) We'll post updates periodically, so feel free to subscribe.

If you can provide help on any of these API conversions (or others), and/or you have a good understanding of PHP_CodeSniffer, please make yourselves known! The more people involved, the more awesome and helpful this project will be. :D

Sprint Goals
Sprint 1: August 11 - August 22 (v0.1.0)
Sprint 2: August 25 - September 12
Sprint 3: September 15 - September 22
Sprint 4: September 22 - October 3
(DrupalCon Amsterdam)
- Automated tests
- BoF / Mid-way Demo!
Sprint 5: October 6 - October 17 - Entity/Field API (including contrib Entity API module?) #2342323: [meta] Entity Field API fun time!
- Features? CTools?
- State API?
- Finish up form API?
Sprint 6: October 20 - October 31 - *.libraries.yml
- Theme system
Sprint 7: November 3 - November 14
(BADCamp)
- Docs!
- Panic! :D
- Final show-off
Support from Acquia helps fund testing for Drupal Acquia logo

Comments

webchick’s picture

Issue summary: View changes

Linking some stuff.

webchick’s picture

Issue summary: View changes
webchick’s picture

Here's an update from our scrum call earlier today!

Adam's first task has been working on refactoring the module to make it both easier to contribute to and easier to maintain. He is keeping compatibility with PHPCS so we can remain in-sync with Coder Review and benefit from IDE integration, but are using new PHP OO hotness (traits, etc.) to break apart monolithic horror classes into nice bite-sized chunks. We figure that some upfront investment here will get us a lot better velocity going forward.

For example, instead of just a single InfoToYamlSniff, containing many turns and twisty passages, each logic check is broken into its own class like CoreVersionSniff and ObsoleteDependenciesSniff. Where there's common functionality that many sniffs might need, these have been pulled into traits like a ReaderTrait, WriterTrait, OncePerFileTrait, etc. This avoids a lot of copy/pasta code.

There's also now Drush integration which wraps phpcs / phpcbf with nice commands like drush dmu-analyze [modulename] and drush dmu-upgrade [modulename].

This work is all happening in the refactor branch. Testing and reviews welcome!

Between now and next week, Acquia has another Hackathon happening. Adam's going to join up with the team porting Acquia Connector module in order to get some "in the field" knowledge of what's needed by DMU, and Tim Plunkett is going to help us identify a suitable module from the Demo Framework distribution to use as a target module for DMU going forward. While Pants is cool and all, it doesn't really have much real-world value. ;)

Wim Leers’s picture

Hurray, that already looks much better! :)

japerry’s picture

Wow! Yay! the fun part about proof of concepts is the unholy things we do to make the concept look neat ;-)

phenaproxima’s picture

A note about the refactor branch: it's still not QUITE where I want it -- namely, phpcbf's behavior is unpredictable -- but it's coming along. I'm hoping to get at least the info file sniffs done and working by the end of today.

phenaproxima’s picture

An update...

After banging my head against PHP_CodeSniffer's limitations all day yesterday, I concluded that using it to cover extensive API changes is impractical. PHPCS was designed to detect violations of a coding standard, and to that end it excels at making minor, surgical (search-and-replace) fixes. Function name changes, arguments added or removed...these are what PHPCS is good for, and we should use it to make those kinds of changes. Its ability to generate reports, and its integration with Coder and IDEs, is also handy.

Major API changes -- the new info file format, for example, or the new routing system -- need to be converted in a different way. They need a lot more scope and power than PHPCS can provide. So I have been writing custom classes to perform those sorts of changes, and they have proven to be about a million times easier to write, to use, and to maintain. We will be able to cover many D8 APIs this way, not just the 10 or so most important ones.

We'll be able to (and we should) write sniffs to detect API changes, but we shouldn't try to wrangle PHPCS into doing things that are out of its reach. Instead, we should just make them non-fixable errors, and let the conversion classes do the heavy lifting via Drush. This way, we'll get the best of all worlds:

  • Developers will be able to run PHPCS and see what needs changing. They'll be able to use PHPCBF to automatically fix the low-hanging fruit.
  • They can use Drush to do heavy changes, or they can do them manually.
  • Drush can invoke PHPCS and PHPCBF in order to fix the low-hanging fruit before it does the harder stuff.

So with this approach, people will be able to convert their modules the easy way (let Drush and PHPCS do it all), or the hard way (let PHPCS do the easy stuff, then do the harder stuff manually).

Make sense?

Wim Leers’s picture

PHPCS was designed to detect violations of a coding standard, and to that end it excels at making minor, surgical (search-and-replace) fixes.

Major API changes -- the new info file format, for example, or the new routing system -- need to be converted in a different way. They need a lot more scope and power than PHPCS can provide.

That's what I've been sensing since the beginning, so this makes sense to me.

So I have been writing custom classes to perform those sorts of changes, and they have proven to be about a million times easier to write, to use, and to maintain.

This sounds sensible also, except… isn't this what https://github.com/grom358/pharborist (#2258215: Look into Pharborist and PHP-Parser) provides in the first place?

So: what's the value of sticking to PHPCS? It sounds like the value is that if it's all the same system, it's easier to maintain, and to provide both "the small stuff" and "the big stuff" in one consistent way. But at the same time it seems like Pharborist could be a much better tool for "the big stuff"?

Instead, we should just make them non-fixable errors, and let the conversion classes do the heavy lifting via Drush.

Here, I'm lost. Isn't Drush simply a way to not have to execute PHPCS itself directly, so that we can do state tracking etc? How does Drush help with converting "the big stuff"?

phenaproxima’s picture

The custom conversion classes use PHP Parser to manipulate the code, but they're still responsible for the higher-level stuff, like creating controller classes, moving functions around, and so forth. In other words, the conversion classes know what to do and how to do it, and they use PHP Parser to do the actual grunt work of altering the syntax tree.

The purpose of using Drush, apart from ease of installation and use, is that it provides us with a bootstrapped Drupal environment in which to run these conversion routines.

The value of supporting PHPCS is mainly to integrate with IDEs and Coder, and to do the really small-scale stuff that might be awkward with PHP Parser. It would also be of value to developers who want to know WHAT needs doing, but would prefer to do it themselves.

P.S. I'm sorry I'm not too good at explaining this stuff. Reading the code in the refactor branch might make things a bit clearer, starting from the drush.inc file.

webchick’s picture

That outline actually sounds really great to me. Very impressed with how fast this is coming together so far. :D

Wim Leers’s picture

Before reading the new comments here, I read #2258215: Look into Pharborist and PHP-Parser, where I saw "PHP Parser" being mentioned for the first time. You also mention it in #9 here. So you already are using something else than PHPCS for "the big stuff".

Sounds sensible :)

RE: explaining stuff. This is tricky stuff to explain, so no worries. But perhaps you could create a schematic (maybe using Google Drawing) to explain the high-level architecture as you currently see it? You could then attach it as an image or PDF to this issue, and whenever you make significant architectural changes, you could post an updated version? That will make it easier for anybody who's interested to follow along :)

(There are 15 followers on this issue, though "the original team" working on this was only 7 people!)

johnv’s picture

@Wim: Well, anonymous subscribing is one of the new shiny features of drupal.org! As a co-maintainer of several D7 modules, I am very interested in this topic. My plan was to try and port some of them to D8 this summer before DrupalCon arrives in my own country (yeah!). But now I am in doubt: "should I wait or should I port?".
BTW: my main tools are git and editpad. Drush sounds fine, but for parttime programmers, each tool is a tool too much. So, a UI like Coder has, would be a top prio for me!

phenaproxima’s picture

@johnv -- I definitely agree with you that adding tools indiscriminately would be bad...but I figured that Drush is a good bet because it's so commonly deployed and so very handy. :)

phenaproxima’s picture

The refactor branch has been merged into master, and removed. The following conversion routines are now (entirely or mostly) implemented: info files, hook_menu(), hook_init(), and hook_exit(). Documentation is forthcoming.

webchick’s picture

Report from my sync-up with Adam again today!

All the stuff from the refactor branch has now been merged into the master branch. It includes:

  • .info.yml generation, .routing.yml generation, event subscribers for hook_init/exit, and controller class generation.
  • Still leverages phpcs as the problem detection routines, to benefit from IDE integration/Coder Review UI, and to allow people to fix the problems manually if they'd like.
  • However, no longer using phpcbf to convert the code, but rather a Drush plugin. Run drush dmu-upgrade target_modulename from the Drupal 8 root (after enabling the module and running composer install). No output == it worked ok. :)
  • For conversions, we're using the pharborist library, which is both easier to read/write compared PHP Parser and especially raw PHP tokens, and also has an extremely active/enthusiastic maintainer (grom358). The library also handles whitespace, which will probably come in handy.
  • Conversions themselves are implemented as Drupal 8 plugins. This will hopefully make the project much easier for Drupalists to contribute to, since this is a common pattern used in Drupal 8. You can see some examples under the src/Plugin/DMU folder. Here's HookMenu.php so you can get an idea of the syntax:
    
    namespace Drupal\drupalmoduleupgrader\Plugin\DMU;
    
    use Drupal\drupalmoduleupgrader\Converter\HookConverterBase;
    
    /**
     * @Converter(
     *  id = "HookMenu",
     *  description = @Translation("Converts Drupal 7's hook_menu() to shiny new routes."),
     *  hook = "menu"
     * )
     */
    class HookMenu extends HookConverterBase {
    
      /**
       * {@inheritdoc}
       */
      public function execute() {
        $hook = $this->module->getHook('menu');
    
        if ($this->containsLogic($hook)) {
          throw new \LogicException('Cannot convert a hook that contains logic.');
        }
    
        // Get the plugin manager for route converters.
        $manager = \Drupal::service('plugin.manager.drupalmoduleupgrader.route_converter');
    
        // Execute hook_menu() to get the router items.
        $router_items = $this->module->execute($hook->getName()->__toString());
    
        foreach ($router_items as $path => $router_item) {
          foreach ($manager->getDefinitions() as $id => $definition) {
            $plugin = $manager->createInstance($id)->setModule($this->module);
    
            if ($plugin->isExecutable($router_item)) {
              $plugin->execute($path, $router_item);
            }
          }
        }
      }
    
    }
    
  • For classes and other things that will be written out multiple times, these are encapsulated in Twig in the templates folder. For example, Form.twig:
    
    namespace Drupal\{{ module }}\Form;
    
    use Drupal\Core\Form\FormBase;
    use Drupal\Core\Form\FormStateInterface;
    
    class {{ class }} extends FormBase {
    }
    
  • Dynamic elements are replaced during code write-out. Once again, this basic syntax should be familiar to Drupal 8 developers who want to contribute to the module and add other conversion routines.

---

Adam's now working on writing up some docs about the overall architecture which Wim asked for, then will try and get form generation working. Exciting!!!

Wim Leers’s picture

Wow, exciting progress! :D

larowlan’s picture

The design sounds awesome!

phenaproxima’s picture

I just added options to drush dmu-upgrade allowing you to specify which conversion routines to run or bypass. You can do:

drush dmu-upgrade module --only=HookMenu,InfoToYAML,...

drush dmu-upgrade module --skip=InfoToYAML,HookExit,...

Fixed some bugs as well, and worked with @grom358 to finish (more or less) the GenericRoute converter plugin. Which means that some callback functions are now moved into default controllers...we're getting there!!!

webchick’s picture

Here's today's update!

- Docs, docs, docs! Both the project page and README was updated for the latest version of the code to help people get the sucker up and running, and Adam's made a great start at developer documentation explaining how everything works. YEAH!

- Lots of more work on various conversions. Here's a sample diff generated by the script as of 5 minutes ago. Covers .info.yml, .routing.yml, basics of page/form controller generation, block plugins, etc. HUGE thanks to grom358 for all of his help with Pharborist, a highly usable PHP parser, which is really coming along!

- Next up is working on function replacement stuff which will allow us to hit variable_*, global $user, $user->uid, theme(), etc.

- If various of those function replacement changes are made by hand, you actually start to see stuff happening now! Very exciting. :) Here's a block from a D7 module that's been converted with DMU showing up on a D8 site:

Pants block showing up on Drupal 8 site

phenaproxima’s picture

As of today, DMU robustly handles the following function conversions:

* variable_get/set/del (plus extracting default configuration from variable_get)
* system_settings_form
* watchdog

webchick’s picture

"End of sprint 1" update!

Just officially closed out #2320075: [meta] Get DMU working on Pants module again. Obviously, the module is not "working" working yet, but quite far enough, and even further than this project did in February! :)

Here's the current diff generated: https://www.drupal.org/files/issues/current-conversion_7.patch

As you can see from the diff, the following things are covered:

- .info => .info.yml
- variable_* -> config API
- Route conversions, including access callbacks
- Settings form conversions
- Blocks -> plugins
- A variety of simple find/replace rules

If you turn Pants module on, you get no errors; you can place and view the change pants block; the configuration form is functioning. YEAH!

Config form converted.

Tagged a 0.1.0 release to refer back to, in order to free master up for moar refactoring!

Next up will be menu links, a smarter way to do entity property -> method conversions, and more! We'll also now start to test this against "real" module targets, the first one being Diff, on tim.plunkett's recommendation.

webchick’s picture

Issue summary: View changes

We demoed the project internally to some OCTO folk today to wrap things up for this sprint, and got some great feedback which we've captured in the following issues:

The biggest thing there is obviously ditching PHPCodeSniffer entirely in favor of a "dmu-analyze" Drush command. Our main concern here was one of usability/contribute-ability. PHPCS makes the installation instructions for this project much more onerous than they need to be, and requires contributors to the project to duplicate their logic checks in two places (one for the "finding" and one for the "fixing") using two incompatible systems (one being investigating individual tokens which is roughly like fighting with both hands tied behind your back and both legs chopped off). Additionally, of our target audience, many more will have Drush installed than PHPCS, so it lowers the barrier to using it as well.

We still do definitely want to keep the reporting aspect, however. There's tremendous value for people new to Drupal 8 to understand the nature of what code changes have happened that affect their modules, and how the new systems work. So since phenaproxima made so much progress on the project this sprint outside of what we originally "scoped," it's likely that we'll shift some energy into this over the next two weeks.

More on Monday!

webchick’s picture

Issue summary: View changes

Updating the battle plan!

So last sprint phenaproxima actually got WAY further than we originally suspected, so re-jiggering the plan for this sprint.

This sprint, the focus shifts from the silly demo module Pants to a project of actual community significance, Diff. While that module has a D8 port already in progress, Diff exercises enough of the main Drupal APIs that we should be able to find lots of juicy things to work on there. ;)

We took a spin through the module code. Last sprint, phenaproxima got a good chunk of hook_menu() and hook_block_*() working. This sprint, we're diving deeper into those APIs, trying to handle e.g. MENU_LOCAL_TASK and other variants, page callbacks in include flies, and configurable blocks.

And finally, the major architecture overhaul to remove PHPCS and replace it with the dmu-analyze is big another area of focus. This code already exists in master.

Should be fun! :)

webchick’s picture

Issue summary: View changes

Penciling in some thoughts for next sprint as well, though note that this is subject to change.

webchick’s picture

Oops, haven't updated this in awhile. :)

#2325587: Ditch PHPCodeSniffer, in favor of "dmu-analyze" report containing similar thing is complete, although the output is pretty stark and needs some feedback on how to improve. Fortunately, we have an issue for that very thing here: #2328647: dmu-analyze's output is hideous

#2193649: Upgrade hook_menu() is coming along. It's generating menu links, local tasks, and actions, although there are still a few snags that Adam is working through atm, chiefly #2328827: Local tasks not appearing on admin/config/content/diff and #2331863: "Some mandatory parameters are missing errors" with formerly optional route parameters.

As an added bonus, it's also handling moving classes to PSR-4, and a few new function replacements. :)

Latest diff of conversions against Diff module is here: https://www.drupal.org/files/issues/current-conversion_11.patch

webchick’s picture

YESSSS! Check this out!

Tabs with sub-tabs

DMU now converting:
- Routes
- Menu links
- Local tasks
- Action links
- Probably other things I forget now.

Additionally, lots of work done upstream on Pharborist to allow things like commenting out entire chunks of code. This will be super useful later on.

Things that are left to do before demo tomorrow: https://www.drupal.org/project/issues/search/drupalmoduleupgrader?projec...

- Fix dmu-analyze #2335915: dmu-analyze: Illegal string offset 'page callback' HookMenu.php:176
- Get Parametric rewriters working again. #2336299: Get parametric rewriters working again
- If possible, get Diff's "Entities" tab to show up (e.g. by commenting out code)
- #2331849: Convert drupal_set_title()
- #2336291: Convert module_invoke_all()
- Figure out what got committed to 8.x core between last night and this morning broke the plugin registry #2336289: Get DMU working again against the latest D8

Home stretch!! :D

pfrenssen’s picture

Some excellent work being done here, great job guys!

webchick’s picture

Issue summary: View changes

Sketching out some more battle plan ideas for the future, based on today's demo call. Will update this better on Monday.

webchick’s picture

Oh btw, in case you missed it on the Twitters. ;P Here is a ~5 min video of the most recent tagged release. https://www.youtube.com/watch?v=RbKjHbpdGYo Also adding to the project page momentarily.

webchick’s picture

Issue summary: View changes

Hello lovely people!

For this sprint, we chose to focus on prepping the module for Amsterdam, where it will be demoed at both a dedicated BoF session (https://amsterdam2014.drupal.org/bof/drupal-module-upgrader-help-collabo...) and incorporated briefly into the module upgrade lab at https://amsterdam2014.drupal.org/session/upgrade-your-module-drupal-8. That entails a couple of things:

- #2327335: [meta] Port automated tests to D8. This is very useful because if a module (such as Pants) took the time to write tests in advance of the port, by doing a few simple find/replace routines we're able to give module authors a complete list of what's left to fix in their module before tagging an 8.x release. Currently working for WebTestCase, UnitTestCase still a todo. Here's a screenshot of it mostly working:

Passes and fails
- #2328647: dmu-analyze's output is hideous. One of the key benefits of this module to people who are new to Drupal 8 is a curated intro to the list of change records specifically for your module. The command-line way was getting too onerous, and so we went with prettified HTML output, which currently looks something like this:

Warnings and errors

Feedback on the exact format warmly welcome!

- #2340435: Create docs for DMU contributors Since one of the goals of Amsterdam is to get some new contributors involved with writing routines, we want to make sure this project is as easy to contribute to as possible. If you have any specific items you feel need to be addressed or questions that need answers, please post 'em there!

Adam has also started digging in on config schema and Twig, though those were both slotted for later sprints. Such an over-achiever. :D

xjm’s picture

Issue summary: View changes
webchick’s picture

Version: » 8.x-1.x-dev
Issue summary: View changes

Greetings, faithful followers! :)

DMU passed the formidable QA hurdles for this sprint and is now ready for download as 8.x-0.4!

Now that we're using standard contrib numbering schemes, you can just drush dl drupalmoduleupgrader and it'll grab it. :D (Don't forget to run composer install afterwards though!)

Here are the release notes:
- Nice, HTML-ified "analyze" report with linkified links, files/line numbers, etc.
- Covers conversion of automated (web) tests, to assist in porting modules with test coverage. Unit tests are not converted yet.
- Moves/splits classes to proper PSR-4 namespace.
- Generates FIXME messages for some things it can't do yet, and comments out some invalid function calls that cannot be converted (again, that depends on plugin coverage).
- A first crack at some of the common Entity API functions: entity_load/multiple, entity_save/delete, and entity type-specific variants: node_load, user_delete, etc. Covers nodes, users, comments, and taxonomy terms.
- Modifies certain basic CTools functions. Calls to ctools_include() are commented out; the object caching functions are rewritten to use the TempStore API; in some circumstances, it will try to convert ctools_get_plugins() to an empty plugin service.
- Updated routing access checks for compatibility with 8.0.x (added simple AccessResultInterface::allowedIf() wrapper around boolean return values in custom access callbacks)
- Grep converter type now uses YAML instead of JSON for easier editing; it's also a config file.
- Non-destructive. Most old code is left in place instead of being removed.

On the topic of QA, the next sprint will focus on automated testing of the current code, so that future refactorings are less likely to introduce regressions. Hope to see some of you in Amsterdam next week at:

Woohoo!

tim.plunkett’s picture

Non-destructive. Most old code is left in place instead of being removed.

A cool feature might be an opt-in flag like --prune/--destroy/--remove that would kill the old code if we REALLY want the old stuff gone.

webchick’s picture

webchick’s picture

Crashing for now, but as an update, we had another demo call today with other OCTO folks, and here was the hit-list. Unfortunately I will not have a chance to go and issue-ize all of this, but so long as it's captured I figure that's better than nothing. :)

BEFORE AMSTERDAM:
----------------
- l()/url() use: how to match a D7 path to D8 path? can we use the routing system to resolve? if yes (and is literal), convert it, if no, @fixme.
  ** Internal URLs (admin/foo - convert these where possible)
  - Internal unrouted paths (robots.txt, cron.php) - router won't find it
  - External URLs (https://drupal.org/node/foo) - comment out
- access param should get AccountInterface $account variable in there somewhere. :)
- @fixme for unit tests
- Move any tests to the Tests namespace, even if we don't understand their base class.
- (During) Tag 0.5 after b1, if that happens during DrupalCon (we hope!)

IF POSSIBLE:
-----------
- changeAcces: instead of ->allowedIf() use ->allowedIfHasPermission()
- Fix AjaxTestCase and other base classes (create a map of names before/after?)
- --destroy flag (or --dry-run?)
Docblocks on generated methods
'HookMenuAlter' 'error'?
- Human readable name of test? How to get that instead of the class namespace?
- Tests change from ending in *Case to not. (PantsConfigurationTestCase => PantsConfigurationTest)
- $account / $user switched meanings between D7 => D8. ;( Can we switch them?

AFTER AMSTERDAM:
---------------
CTools plugins: See if there are patterns within popular contrib modules, try and do that use case. — start with plugins for ctools-dependent API modules (Views, Context, Entity Reference, Rules, Features?) - Ask eclipsegc for a recommemdation. (Drupal Commerce / Address Field)
- include file per plugin, starts with $plugin variable, move to annotation.
- functions that can be moved as methods to class
- class could be in that same file or could not be. :D

- What about Views?
  - Views handlers
  - Views plugins
  - hook_views_data()
  - BADCamp?
xjm to post a picture from a flip chart
webchick’s picture

Greetings! Long time no ping!

Back from DrupalCon Amsterdam, where our BoF was sparsely attended but our Upgrading Modules lab where we demoed DMU was *hugely* attended, so we managed to grow the number of participants in the issue queue by half a dozen or so! (Hello there, new folks! :D)

Over the last week and a half, phenaproxima has been working on refactoring the code to reduce stateful-ness and increase unit testability so that we could start adding some much-needed automated tests. However, this left a few bugs in the functionality which we are currently sorting out, with a tentative plan to release a 0.5 version on Wednesday or so.

Next up will be entity/field API stuff now that D8 is in beta and those APIs are more locked down. Adam also wants to embark on some serious refactoring to de-couple the code more now that we have a basic safety net of tests to inform us if things broke.

webchick’s picture

Ok, heeeeere's v0.6! https://www.drupal.org/node/2361937

By far the biggest change is a re-write on top of a new architecture which is hopefully much more approachable for new contributors. The DMU plugins are now split into four main types:

Analyzers
These plugins will add things to fix to the dmu-analyze report, along with links to docs on how to fix them.
Converters
These plugins actually convert the code from D7 to D8 (e.g. .info -> .info.yml).
Cleaners
After conversion, these plugins will clean out the old D7 code that's no longer needed (for example, delete .info files once .info.yml exists)
Indexers
These plugins (which are relatively few in number and only for modules that create new "architectural concepts") actually find the things in the code to analyze/convert (e.g. hook implementations, automated tests, etc.)

This separation is very nice because Analyzers are actually incredibly easy to write, and we can then start "crowd-sourcing" coverage of most of the changes in https://www.drupal.org/list-changes shortly, with Converter and Cleaner logic to be implemented later when/if someone has time.

Adam is in the process of writing extensive docs for this over at https://www.drupal.org/documentation/modules/drupalmoduleupgrader/contri.... Feedback wanted!

Another thing we could use feedback on is #2340935: Order analyze report into logical sections. Given that the dmu-analyze report will be indispensable for developers initially moving from D7 to D8, what's the most logical way in which to order the issues in the report?

Please give 0.6 a try and let us know what you think! :D

webchick’s picture

Status: Active » Fixed
FileSize
24.61 KB

Ok, 1.0 of Drupal Module Upgrader has been tagged!! :) Grab it here: https://www.drupal.org/node/2374897

1.0 marks a feature-complete (analyze / upgrade / clean) and API stable (for plugin implementers) version.

Adam's internship at Acquia also ends tomorrow. Job super well done!! Both of us definitely plan to stay involved in this project though, on a volunteer basis. :)

But, since we will no longer be using this battleplan, and instead just using the normal issue queue going forward, going to close this issue out.

And remember: it's not goodbye, it's....

See you later!

Status: Fixed » Closed (fixed)

Automatically closed - issue fixed for 2 weeks with no activity.

Shaun Holt’s picture

Thought I would test it out with Drupal RPG.... for the most part everything does well with the modules within.... but the main module in its root folder.... causes an error.... I am using the 8.1.dev version of dmu.

/opt/lampp/htdocs/social83/modules/drpg$ drush dmu-upgrade --path=modules/drpg drpg
Indexing...done.
Argument 2 passed to [error]
Drupal\drupalmoduleupgrader\Routing\LinkBinding\LinkBindingFactory::create()
must be an instance of
Drupal\drupalmoduleupgrader\Routing\Drupal8\RouteWrapper, null given,
called in
/opt/lampp/htdocs/social83/modules/drupalmoduleupgrader/src/Plugin/DMU/Converter/Links.php
on line 92 and defined LinkBindingFactory.php:34
E_RECOVERABLE_ERROR encountered; aborting. To ignore recoverable [error]
errors, run again with --no-halt-on-error
Drush command terminated abnormally due to an unrecoverable error. [error]

jurgenhaas’s picture

@Shaun Holt, I'm not a maintainer of this module, I just saw your comment.

I think it's better to open a new issue for your problem as this one has been closed 18 months ago and is about a different topic. Your chances to get feedback are much better with a new issue.

Shaun Holt’s picture

Thank you Jurgenhaas