Problem/Motivation

We got 50 test modules inside system module to test various core operation. However this modules written in D7 way and converting them as it is to D8 might not be very efficient.

Proposed resolution

Rewrite the modules so that it can be well unit-tested and stand as a reference modules for contrib module developers to port their modules into D8.

Remaining tasks

1. Need discussion on rewrite.

Modules

  1. ajax_forms_test
  2. ajax_test
  3. batch_test
  4. bundle_test
  5. cache_test
  6. common_test
  7. common_test_cron_helper
  8. condition_test
  9. config_upgrade
  10. database_test
  11. design_test
  12. drupal_system_listing_compatible_test
  13. drupal_system_listing_incompatible_test
  14. entity_cache_test
  15. entity_cache_test_dependency
  16. entity_crud_hook_test
  17. entity_test
  18. error_test
  19. form_test
  20. image_test
  21. menu_test
  22. module_autoload_test
  23. module_test
  24. paramconverter_test
  25. path_test
  26. plugin_test
  27. requirements1_test
  28. requirements2_test
  29. router_test
  30. serialization_test
  31. session_test
  32. system_dependencies_test
  33. system_incompatible_core_version_dependencies_test
  34. system_incompatible_core_version_test
  35. system_incompatible_module_version_dependencies_test
  36. system_incompatible_module_version_test
  37. system_mail_failure_test
  38. system_module_test
  39. system_test
  40. taxonomy_test
  41. test_page_test
  42. theme_page_test
  43. theme_test
  44. transliterate_test
  45. twig_theme_test
  46. update_script_test
  47. update_test_1
  48. update_test_2
  49. update_test_3
  50. url_alter_test

User interface changes

Hopefully not.

API changes

N/A

Original report by @dawehner

IRC chat:

[14:03] <dawehner> vijaycs85: hi
[14:03] <dawehner> vijaycs85: even more important: having proper unit tests would make things way faster to test
[14:04] <dawehner> vijaycs85: and the amount of custom test code would be reduced etc.
[14:04] <vijaycs85> dawehner: agreed... however we might need to start it as a meta :)
[14:05] <dawehner> vijaycs85: good point
[14:05] <vijaycs85> dawehner: oh god.. we got 48 MODULES!!!
[14:05] <dawehner> vijaycs85: you know, at the same time a lot of the custom functions should be maybe ported to a proper more d8 like api
[14:06] <dawehner> vijaycs85: so it's maybe hard to crowd-source these tasks?
[14:06] <dawehner> in the sense of: it's hard to describe generically what to do
[14:10] <vijaycs85> dawehner: I guess,  this modules are going to be first of its kind and it good as in  we will have something to refer in core and bad, it takes time as it is first
[14:11] <vijaycs85> dawehner: however keeping old module and trying to convert every piece of it also sounds not fun - may lead to bad example of D8 module
[14:11] <dawehner> vijaycs85: right
[14:12] <vijaycs85> dawehner: may be we can start a meta and try to see how long we can go? :)
[14:13] <dawehner> yeah,  a meta would certainly help
[14:13] <dawehner> vijaycs85: i guess it's also important to set these issues to postponed or something like that, so people don't waste their time
[14:14] <vijaycs85> dawehner: agreed.. I will update them...
[14:18] <dawehner> vijaycs85++
[14:30] <vijaycs85> dawehner: totally 83 defects (out of 148) are test modules :)
[14:30] == Artusamak has changed nick to Artusamak_afk
[14:31] <dawehner> vijaycs85: well test modules are for itself not bad, as they even provide some example code
[14:31] <dawehner> but yeah on the longrun they for sure have to be more like proper d8 code
[14:32] <vijaycs85> dawehner: okay

Comments

Issue summary:View changes

Updated issue summary.

Issue summary:View changes

Updated issue summary.

Issue summary:View changes

Updated issue summary.

Issue summary:View changes

Updated issue summary.

Issue summary:View changes

Updated issue summary.

Issue summary:View changes

Updated issue summary.

Issue summary:View changes

Updated issue summary.

Issue summary:View changes

Updated issue summary.

Issue tags:+WSCCI

Um. What's happening here?

While a lot of test modules ought not to be necessary, any test module that has a hook_menu must be converted! Why were those all closed? They shouldn't be closed. They need to be converted unless someone eliminates them first.

And I don't think we have time to rebuild all of them before we have to rip out the old routing system, which is starting to cause problems.

Yeah, while this is a really good idea in general, we cannot and should not block release-blocking conversions on it.

ok, We got the list of page controller & few form builder issues already (We can add missing form callback issues, once done with controllers). We may need to re-open them all and start progress on them.

Status:Active» Closed (won't fix)

As per #2, re-opened all test related issue. So closing this issue for now. if we find time to re-write test modules, we can re-visit this.

Issue summary:View changes

Summary update.