View names can be 255 characters long but this breaks stuff

Looking pretty good.

1. +++ b/core/modules/views/lib/Drupal/views/Plugin/views/display/DisplayPluginBase.php
   @@ -1124,7 +1124,7 @@ public function optionsSummary(&$categories, &$options) {
   +      'value' => views_ui_truncate($title, 32),
   

   Not the fault of this patch. We should move this somewhere more OO in a follow up. Isn't this a more general helper that we could just yoink out and put in out of our component classes?

2. +++ b/core/modules/views_ui/js/views-admin.js
   @@ -132,7 +132,7 @@
   -      return from + this.suffix;
   

   This is handled below now?

Yep that is a helper function for the views_ui module

Yes, that is what I was saying :) We should consider moving it from Views UI module....

+1 for the issue in general.

1: 2220765.1.patch queued for re-testing.

Tested manually and confirmed the patch resolves the issue. Giving the view a very long label and saving now works fine, with no errors about view displays the user hasn't even added. Adding (e.g.) a page display also now works without error.

Why isn't the bot picking up the test-only patch?

Oh, I see why maybe. There's some funky at the top of the patch.

xjm: alexpott: This test is hosed in the test database ... I'll have to sort out why and manually clear it up.

K ... should be unblocked.

All good. I even read the JavaScript.

I don't really think this is major, though, so downgrading.

The code looks good! (beside of the usual unimportant nitpicks).

Now I am wondering what @dawehner saw that I didn't. :)

Re-uploading the patch for clarity.

1. +++ b/core/modules/views_ui/lib/Drupal/views_ui/Tests/WizardTest.php
   @@ -0,0 +1,65 @@
   +/**
   + * Tests the wizard.
   + */
   +class WizardTest extends WizardTestBase {
   

   I really like to point to the actual tested class using @see

2. +++ b/core/modules/views_ui/lib/Drupal/views_ui/Tests/WizardTest.php
   @@ -0,0 +1,65 @@
   +
   +  public static function getInfo() {
   

   MIssing @inheritdoc

3. +++ b/core/modules/views_ui/lib/Drupal/views_ui/Tests/WizardTest.php
   @@ -0,0 +1,65 @@
   +    $this->drupalPostForm('admin/structure/views/add', $view, t('Save and edit'));
   ...
   +    $this->assertText('Machine-readable name cannot be longer than 128 characters but is currently 129 characters long.');
   +    $this->assertText('Path cannot be longer than 254 characters but is currently 255 characters long.');
   +    $this->assertText('Page title cannot be longer than 255 characters but is currently 256 characters long.');
   +    $this->assertText('View name cannot be longer than 255 characters but is currently 256 characters long.');
   +    $this->assertText('Feed path cannot be longer than 254 characters but is currently 255 characters long.');
   +    $this->assertText('Block title cannot be longer than 255 characters but is currently 256 characters long.');
   

   Is there a clear distinction between using t and not for checking stuff happened on the page?

We don't do @inheritdoc for getInfo().

For #3, our old rules would say to use t(). But since it doesn't actually matter in practice here, I'd be inclined to start leaving it out to get rid of unnecessary coupling unless we are actually testing translation. What do you think?

+1 for the first point though, though it needn't block commit.

2. We do @inheritdoc for getInfo() and setUp() now. See https://drupal.org/node/325974 and #338403: Use {@inheritdoc} on all class methods (including tests). The precedence argument works both ways :)

Thank you, I am fine with not using t(), but even in this new file we are already inconsistent.

That seems an unfortunate number of places to have to manually set that property, but I guess it can't really be helped.

However, shouldn't this actually be like 166 or whatever, pursuant to #2220757: Limit length of Config::$id to something <= 166 characters?

No, this is an end-user-facing label. Not a machine-name-ID-thing.

Edit: Or rather, various user-facing labels and other things like path aliases, none of which are config IDs. :)

Oops, I thought I had already committed this, so it totally dropped off my radar. :( Sorry about that!

Committed and pushed to 8.x. Thanks!