Skip to content

Conflict investigation - #230

Closed
matt-bernhardt wants to merge 44 commits into
masterfrom
conflict
Closed

matt-bernhardt wants to merge 44 commits into
masterfrom
conflict

Conversation

@matt-bernhardt

Copy link
Copy Markdown
Member

This is meant for an investigation into a merge conflict.

djanelle-mit and others added 30 commits August 31, 2026 16:10
This sets up a root-level package.json that will manage all the plugins' internal dependencies
via workspaces. It allows us to only have one package-lock.json file, rather than one per plugin,
which makes keeping dependencies in sync across plugins easier to achieve.

It also saves a bit of space in the respository, as these files can be tens of thousands of lines long and largely duplicative.
** Why are these changes being introduced:

We already have a plugin that implements the hours ecosystem, so it
makes sense to define the new hours block for the homepage within that
plugin, rather than in a separate plugin.

** How does this address that need:

This moves the content of the hours block into the existing hours
plugin. As part of this move, a few things get adjusted:

* Package.json at the repostitory root gets the new plugin path added to
its list of workspaces

* Because the hours plugin already uses the src/ folder to define PHP
classes, it doesn't make sense to re-use that location for the block
source files. We instead take a page from WordPress itself, and place
those materials in the blocks/ folder.

* Because the path to the block materials changes, we need to update the
package.json file in the plugin to pass the --source-path argument to
the build and start scripts.

* The PHP script at the plugin root needs to merge with the existing
plugin. The only unique part from the new plugin is the block
block registration function (which is hooked to the init action) - so we
use that and leave the rest of the boilerplate behind.

* The readme.txt file from the old plugin is no longer needed, as this
is a file we don't really use in this monorepo.

* We also re-run the build command to make sure that everything is wired
up correctly. This updates the hash and index.js file.

* All of this means we bump the plugin version number.

** Document any side effects to this change:

Hopefully none...
(Squashed to remove build artifacts for on block)
This moves the source materials for the hours block into the new mitlib-blocks plugin

- block.json
- edit.js
- editor.scss
- index.js
- save.js
- style.scss

It doesn't touch the built materials, that will happen in a future commit
Now that the hours block has been relocated into the new plugin, and
the naming convention for folders in that plugin have been udpated,
there are a few cleanup tasks that are necessary for everything to
function:

- package.json at the root level now only lists a single workspace, as
  the hours plugin no longer has one
- package.json within mitlib-blocks needs to be told where to find the
  block materials
- The hours plugin itself no longer needs a register_blocks function,
  nor does it need package.json or .gitignore
- The text-domain value for the hours block needs to be updated to use
  its new home
- The mitlib-blocks gitignore needs to ignore .map files

There is an additional side effect in this commit, which is not strictly
related but seems like a good idea at this point:

- The PHPCS configuration now ignores materials in a build/ folder,
  as these are not meant to be run through any sort of static code
  analysis.

Squash me
This also removes the deprecated build materials from the hours plugin
Move Hours block into new block plugin
djanelle-mit and others added 12 commits September 14, 2026 08:37
Let editor select a featured librarian from the block editor
Most of these blocks don't use this boilerplate file, which only had a
console.log statement. In those cases, we remove the file entirely, as
well as the entry in block.json which called for it to be loaded.
These blocks aren't part of a javascript application stack, so there is
no need to declare an entry point into their materials.
The boilerplace declares wordpress dependencies to _always_ use the
latest version, which causes problems in our CI workflow because a new
release may have become available between when the developer started a
branch and when the CI runs to evaluate it.

Keeping branches short-lived is only a partial solution. The more robust
option is to declare our dependencies using a predictable version. This
provides for consistency between environments (among developers, and
with our CI workflows).

The cost of this is that we need to manage dependency updates ourselves,
but this is already something we are doing.
The javascript community has a set of formatting rules which they've
adopted, and their tooling will automatically apply that formatting.

This commit applies the changes necessary for `npm run lint:js` to
return no output.

(We can tweak the rules that it applies as we go forward - some of these
seem pretty silly to me - but having those discussions will take time to
reflect and reach consensus, and we want to get this branch done quickly
for the moment.
This is part of the NPM linter, but there were so many of these changes
that it seemed better to split them out into their own commit.
Use of ' instead of ' causes complaints by the linter...
The boilerplate included an import of useBlockProps - but then did not
actually make use of it. Our choice is to either finish the
implementation, or to stop importing it in the first place.

I initially tried to actually implement this library for the blocks that
are not rendered via render.php, but this resulted in editor-facing
styles leaking out to the public UI.

Instead, I've opted to remove the import entirely. It does not appear to
impact how our defined options are updated via the editor interface, so
I don't think this is something we need at this time.
The featured collection never passes any output through a translation
function, so we do not need to import it in the first place.
This is the same fixup step for the PHP files as we previously took for
the JS files. Run the linter, note what it complains about, and then fix
those lines until the linter reports no output.
This keeps everything in sync with the latest releases of our tooling

@accesslint accesslint Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Found 1 issue across 1 rule.

Reviewed by AccessLint, which checks every pull request for accessibility issues.

?>

</div>
<span class="hero-image-credit">from the <a href="https://archivesspace.mit.edu/repositories/2/resources/603">Harold E. Edgerton papers</a></span>

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

WCAG 4.1.2: Focusable element inside a "img" role whose children are presentational.

Elements with a role that makes children presentational must not contain focusable content.

Details

Roles like button, checkbox, img, tab, and others make their children presentational — hidden from assistive technologies. If those children are focusable, keyboard users can reach elements that screen reader users cannot perceive. Move focusable content outside the parent or remove the focusability.

@matt-bernhardt

Copy link
Copy Markdown
Member Author

This has been superseded by other cleanup work.

@matt-bernhardt
matt-bernhardt deleted the conflict branch September 15, 2026 13:35
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants