Restructure the Dockerfile and the CSS/JS files so that we only rebuild
the CSS and JS when they change, not on every commit.
Before it took several minutes to rebuild the Docker image after every
commit, even when the JS/CSS files didn't change. This also made pulling
images slower.
This requires refactoring the CSS and JS to not use embedded Ruby (ERB)
templates, since this made the CSS and JS dependent on the Ruby
codebase, which is why we had to rebuild the assets after every Ruby
change.
Replace the old IQDB API client with a new client for the new forked
version of IQDB at https://github.com/danbooru/iqdb.
Changes:
* The /iqdb_queries endpoint now returns `hash` and `signature` fields.
The `signature` is the full decoded Haar signature, while the `hash`
is a encoded version of the signature.
* The /iqdb_queries endpoint no longer returns `width` and `height`
fields in the response (these were always 128x128).
* We no longer need the IQDBs frontend server, now we talk to the IQDB
instance directly.
* We no longer send add/remove image commands to IQDB through AWS SQS,
now we send them to IQDB directly. They are sent in a delayed job so
that if IQDB is down, uploading images is still possible, the add
image commands will just get queued up.
* Fix a bug where regenerating an image's thumbnails didn't regenerate
IQDB, because IQDB silently ignored add image commands when the image
already existed in the database.
Flash is dead. It's no longer supported by browsers, it's not
well-supported by emulators, and only two Flash posts were uploaded in
the last year anyway. Old Flash files will continue to exist, but new
Flash uploads will no longer be allowed.
Adjust input boxes to fit the screen better on mobile, and to use
consistent sizes throughout the site (no ad-hoc overrides to make
certain input boxes a bit bigger in random places).
On desktop, this makes most input boxes a bit taller and narrower. On
mobile, it makes input boxes the full width of the screen.
This most notably affects the tag edit box, the comment and forum
post box, the wiki page edit box, and the commentary box.
Regression caused by the upgrade to Webpacker 6.0 in 90cd3293e. This
caused various Javascript errors in old versions of Chrome, which
somehow resulted in the keyboard shortcut for visiting the next page
being triggered when you pressed any key.
Specifically, the mobx library (used by the TagCounter component) called
`Object.entries`, which isn't available in Chrome 49, and for some
unknown reason this triggered the buggy shortcut behavior.
`Object.entries` is supposed to be automatically polyfilled by Babel to
support old browsers, but something changed in Webpacker 6 that broke
this and I couldn't get it working again. The probable cause is that
Webpacker 6 no longer transpiles code inside ./node_modules by default,
which means that any libraries we use that use new Javascript features
won't get transpiled down to support old browsers, but even after fixing
that it still didn't work. The workaround is to just drop mobx and
preact entirely to avoid the issue.
Factor out FontAwesome icons into a set of helpers. This is so that it's
easier to keep track of which icons we're using and easier to change
icons globally.
Refactor the post preview html to use the ViewComponent framework. This
lets us encapsulate all the HTML, CSS, and helper methods for a UI
component in a single place.
See https://viewcomponent.org.
Disable the browser's native spellchecking ability on all form inputs,
except for DText inputs. We do this by setting `spellcheck="false"` on
the <body> tag, and `spellcheck="true"` on DText <input> tags.
This fixes browsers displaying a red wavy underline beneath tags in the
tag search box, among other places. We disable spellchecking globally
because most form inputs, except for DText inputs, aren't meant for
natural English language.
Rewrite the tag counter widget (the one above the tag edit box) to use
React. This is a trial run for using React elsewhere.
We actually use Preact instead of React because it's lighter weight.
We use Mobx for state management because it's cleaner than React's
setState or useState.
This incidentally fixes a couple bugs:
* The tag counter wasn't updated when deleting tags with backspace
* The tag counter wasn't updated when pasting in new tags.
* Show completed uploads to other users.
* Don't show failed or incomplete uploads to other users.
* Don't show tags to other users.
* Delete completed uploads after 1 hour.
* Delete incomplete uploads after 1 day.
* Delete failed uploads after 3 days.
Bug: the uploads page showed a remote size of 146 bytes for Pixiv uploads.
Cause: we didn't spoof the Referer header when making the HEAD request
for the image, causing Pixiv to return a 403 error.
Also fix the case where the Content-Length header is absent.
* Fix preview images not being resized to fit the screen when using the
bookmarklet.
* Fit images to both screen width and height by default. Previously we
fit to screen width only, so tall images were hard to see.
* Allow clicking on the image or pressing Z to toggle image size.
* Move size information to above the image and add resize links:
** small: fit width and height
** large: fit width
** full: fit none
* Bind the image error handler in an onerror attribute on the image
itself so that it will always fire. Before it wouldn't fire if the image
failed immediately on pageload before we could bind the error handler.
Also remove options to configure names of wiki notice pages. These names
generally don't need to be changed and we already hardcode links to wiki
pages in other places anyway.
Change calling convention to explicitly indicate whether the attributes
are for the <th> element or the <td> element. Fixes various cases where
the two were mixed up.
* Fix .col-expand classes not being set correctly on the /post_versions,
/pool_versions, and /notes pages.
* Fix .updater and .updated-at classes not being set correctly on the
/forum_topics page.
* Fix the name param being ignored (noticeable in the post count field
on the /tags page).
* Don't pass empty string when column has no name.
* Fix inconsistencies in how wiki pages were linked.
* Link directly to the wiki instead of to a title search that is expected
to redirect to the wiki.
Don't send the referer when hotlinking preview images on the
/uploads/new page. Improves user privacy and may bypass anti-hotlinking
protections in some cases.
* Standardize the upload widget background colors (light green for success,
light red for error).
* Move the progress bar to the bottom of the upload widget.
* Change the progress bar color to blue.
* Hide the "drag and drop a file here" message after a file has
already been uploaded.
Fineprint text was variously styled with `.info`, `.tn`, `.hint`, or
`.cost-footnote` css classes. Standardize on `.fineprint` instead. Use
`.hint` only for form hints and `.tn` only for <tn> tags in translation
notes.
Incidentally changes the font size of form hints to 0.8em (was 0.7em)
and the color of fineprint to #888 (was #AAA or #666).
Ensure dtext is always wrapped in a `<div class="prose">` or
`<span class="prose">` (for inline dtext) container so that dtext css is
properly applied.
