Largely this is written by an LLM after I provided the targets, and I've looked through the tests themselves to confirm they are accurate. There may be improvements in how to actually *set up* the tests, though, as I am less familiar with those.
This has two test suites: the first for html structure, the second for behavior. Each of them calls another script that provides helper functions for setting things up and taking them down. This keeps the test suites themselves fairly lean.
This fixes the odd alignments of menu-wrapper names by ensuring that both menuitems and menu-wrappers (actually their child triggers) both have space for icons.
Unless a menu item has the `no-dismiss` attribute (set to either `''` or `'true'`), clicking a menu item will dismiss the menu. Clicking a submenu trigger will not dismiss the menu.
The `no-dismiss` attribute must take a string, either empty or `'true'` because react doesn't allow boolean custom attributes.
Only 400 and "800" are used in the codebase for Open Sans. And, we've only ever had up to 700 as available weights, so everything that is set to "800" is actually getting weight 700 (which looks better).
Since the print cycle now loads all images not-yet-loaded (due to lazy loading), there can be a moment of time where it appears pressing Get PDF is doing nothing, depending on connection speed. To add a "loading" message, a custom event is fired at the start and end of the print cycle (before the print dialog comes up).
When initiating the print dialog, it first grabs all img elements with `loading="lazy"` attribute, flips that to `eager`, and then waits for every image to load before resolving a promise and opening the Print dialog.
Removes the additional line after a snippet, so it doesn't effectively create a `\n` at the end of the snippet. This allows snippets to be inline with text.
This simply moves the codeEditor folder to the components directory. `codeEditor.jsx` simply builds a component which is then used elsewhere, and isn't shared between client and server, so it should be in the client directory.
Arguably it could go in `client/homebrew/editor` but I think the `/homebrew/` folder should be eliminated down the line and it's contents just re-sorted.
When working on the editor right now it's a pain to switch between `shared` and `client` hunting for the right file.
Move the SVGs out of the `shared` directory, closer to where they are actually used. They are svg jsx files, setup as components, so i've moved them to the components directory.
Deleted `combat.svg.jsx` because we don't use it anywhere.
renamed the remaining two to be more descriptive of the image for additional clarity.
Moving this component to the component folder so it's closer to where it is actually used. Not moving to the `homebrew` folder because ultimately i think pretty much everything in that folder should move to the components as well (`homebrew` isn't a helpful folder distinction).
Deletes a directory, moves 2 files.
Making some changes to the reset.less so that some default UA button styling is removed.
Then, changing core.less so that the classic "HB" button styling is scoped to a certain class `.colorButton`. This will make it easier to use the button element in other places.
When the page is zoomed in very close, such that <30% of the page is in view, it doesn't register changes to the 'current page'. This fixes that, passing in the 'centerPage' if 'visiblePages' is empty.
I don't love this fix, i think the visiblePages should always have *something* in it, but I can't quite figure out how to set that (since the normal update to visiblePages is happening in an observer that doesn't fire if nothing is in view).
the `safe` keyword for `justify-content`, in combo with `center`, means that the content will be centered in the viewport unless there is not enough space for it. If there is not enough space, it aligns it to the *start*/left edge, rather than keeping it centered and clipping the left edge of the page.
The memoization of the renderPages() method prevents a re-render when something like pageShadows is updated, so displayOptions are added to the dependency array in the memo method.
Prior to fix, the "next page" button in the toolbar wouldn't work well if there were multiple pages in view that were in a single 'row'. This is because the logic is to take the pages that are "visible", take the max of those pages, and then scroll to that page. But the issue is that if the 'max' page is in the same row as other pages, the range of visible pages doesn't change....the max will always be the same.
So the change here basically runs the scroll function twice-- if the first run results in the same 'max' page as before the scroll, it runs it again but with the target page being "max + 1", which will bump the target to the next row.
Reduce the visual prominence of the page input by using a darker background and a text color that matches the rest of the toolbar icons. Darker background still indicates this is an interactive item (is an input), hopefully.
Bad commit here with too much stuff. I apologize.
This sets up two Intersection Observers: the first captures every page that is at least 30% visible inside the `.pages` container, and the second captures every page that has at least one pixel on the horizontal center line of `.pages`. Both can be arrays of integers (page index).
The "visiblePages" array is duplicated and formatted into a "formattedPages" state, which gets displayed in the toolbar.
The toolbar displays that, unless the user clicks into the page input and enters their own integer (only a single integer, no range), which can then jump the preview to that page on Enter or blur().
The Arrow 'change page' buttons jump the preview back and forth by a 'full set'.
If one page is viewed at a time, this is moved on page a time, and if 10 pages are viewed at a time it jumps the pages by 10.
Left to do: adapt the "jump editor to match preview" divider button to work with new "centerPage".
Gazook89