This should *hopefully* fix 17228, by tweaking the build scripts to give the GENERIC viewer something to await to avoid breaking third-party users of the standalone viewer components.
This button is *only* used in the GENERIC viewer, and will currently be visible either in the main or secondary toolbars (depending on the viewer width).
To simplify upcoming changes, and to avoid then having to complicate the relevant CSS rules unnecessarily, let's place the "Open file"-button permanently in the secondary toolbar instead.
(Note that the GENERIC viewer also, since five years, supports drag-and-drop in order to open local files.)
The `viewerCssTheme`-implementation has always been somewhat hacky, and now it's also *partially* broken ever since we've started using CSS nesting.
Trying to support nested media queries would thus require a lot more parsing of the CSS rules, which seems inefficient and thus generally undesirable.[1]
As discussed on Matrix, let's try to remove the `viewerCssTheme`-option and see if there's any (significant) fallout from this.
---
[1] If this option is brought back, it seems to me that it (in Firefox) should probably be set through the platform-code that handles theming.
Depending on the structure of the outline we could potentially need to expand a few levels, especially in long PDF documents, hence it cannot hurt to pause translation in that case as well.
With the changes in PR 17208, where browser-preferences are now handled as "regular" viewer-options, we can tweak the definition of `canvasMaxAreaInBytes` to slightly simplify things in the `PDFViewerApplication.open` method.
Given that this branch is only necessary in development mode and *during* building, but is never actually used in the final viewer-bundles, we can utilize the pre-processor to ignore this code.
Currently we *synchronously* fetch a number of browser preferences/options, from the platform code, during the viewer respectively PDF document initialization paths.
This seems unnecessary, and we can re-factor the code to instead include the relevant data when fetching the regular viewer preferences.
Note that we must append the textLayer to the DOM *before* enabling the `highlighter` and `accessibilityManager`, to avoid breaking e.g. a pending searching operation.
The least invasive solution, that I was able to come up with, is to introduce a new `TextLayerBuilder` callback-function for this purpose.
After PR 17177 the interface of `XfaLayerBuilder` is now inconsistent, since whether or not we directly append the xfaLayer to the DOM now depends on the rendering intent.
I forgot to include `web/l10n_utils.js` in PR 17161, which currently breaks `ConstL10n` since there's no longer a method called `setL10n`; sorry about that!
Most of the strings shouldn't contain special chars (<= 0x1F) so we can
have a fast path which just checks if the string contains at least one such
a char.
To prevent the *standalone* viewer-components from breaking, we need to ensure that the `NullL10n`-interface won't accidentally diverge from the actual `L10n`-implementations.
Looking at the `PDFThumbnailView.setPageLabel` method you'll see that we update e.g. the "aria-label" of the thumbnail-image for documents that contain (valid) pageLabels.
This isn't done in `PDFPageView`, which seems inconsistent, hence this patch.
This patch changes almost all viewer-components[1] to use "data-l10n-id"/"data-l10n-args" for localization, which means that in many cases we no longer need to pass around the `L10n`-instance any more.
One part of the code-base where the `L10n`-instance is still being used "directly" is the AnnotationEditors, however while it might be possible to convert (most of) that code as well that's not attempted in this patch.
---
[1] The one exception is the `PDFDocumentProperties` dialog, since the way it's currently implemented makes that less straightforward to fix without a lot of code changes.
*Please note:* In the Firefox PDF Viewer this findbar is only used for PDF documents placed in e.g. `<iframe>` elements.
By registering a `ResizeObserver` when the `PDFFindBar` is open we slightly unify and simplify how the findbar layout (row vs column) is handled.
This will be especially helpful with upcoming changes, where we'll make use of "data-l10n-id"/"data-l10n-args" to trigger translation in the viewer.
- The old translation engine handled language code casing slightly differently, hence we need to tweak the non-metric locale check in `PDFDocumentProperties` to account for that.
- Use only lowercase names for the pre-defined page names, to improve overall consistency.
*Please note:* These changes only affect the GENERIC build, since `NullL10n` is only a stub elsewhere (see PR 17135).
After the changes in PR 17115, which modernized and improved l10n-handling, the `NullL10n`-implementation is no longer a good fallback for the "proper" `L10n`-classes.
To improve this situation, especially for the *standalone* viewer-components, this patch makes the following changes:
- Let the `NullL10n`-implementation extend an actual `L10n`-class, which is constant and lazily initialized, to ensure that it works *exactly* like the "proper" ones.
- Automatically bundle the "en-US" l10n-strings in the build, via the pre-processor, such that we don't need to remember to manually update them.
- Ensure that the *standalone* viewer-components register their DOM-elements for translation, similar to the default viewer, since this will allow future code improvements by using "data-l10n-id"/"data-l10n-args" in most (if not all) parts of the viewer.
- Remove the `NullL10n` from the `AnnotationLayer`, to avoid affecting bundle size too much.
For third-party users that access the `AnnotationLayer`, as exposed in the main PDF.js library, they'll now need to *manually* register it for translation. (However, the *standalone* viewer-components still works given the point above.)
Given that there's now a bit more asynchronicity in the l10n-initialization in the Firefox PDF Viewer, after PR 17115, try to limit the impact of that by moving it to occur a tiny bit earlier in the default viewer initialization.
In Firefox debug builds, there is an assertion to check that we don't connect
a subelement of an already connected root. Thanks to this assertion, we can see
that the root has already been added to Fluent, hence we don't need to do it
a second time.
We don't need to await anymore on the translation in order to update the
toolbar: it'll be done by Fluent, so we can safely remove the "localized"
event and avoid to wait for it.
*Please note:* This patch contains a couple of micro-optimizations, hence I understand if it's deemed unnecessary.
Move the `AppOptions` initialization into the `Preferences` constructor, since that allows us to remove a couple of function calls, a bit of asynchronicity and one loop that's currently happening in the early stages of the default viewer initialization.
Finally, move the `Preferences` initialization to occur a *tiny* bit earlier since that cannot hurt given that the entire viewer initialization depends on it being available.
Note that CSS-features such as e.g. `flex` didn't exist, or had poor cross-browser support, back when the JavaScript-based solution was initially implemented.
- For the generic viewer we use @fluent/dom and @fluent/bundle
- For the builtin pdf viewer in Firefox, we set a localization url
and then we rely on document.l10n which is a DOMLocalization object.
Given that we only use standard `import`/`export` statements now, after recent PRs, the "exports" global is unused.
Instead we add "__non_webpack_import__" to the `globals` to avoid having to sprinkle disable statements throughout the code.
Finally, the way that `globals` are defined has changed in ESLint and we should thus explicitly specify them as "readonly"; please find additional details at https://eslint.org/docs/latest/use/configure/language-options#specifying-globals
When pdfBug is true, the substitution font is used in the text layer in order
to be able to know what is the font really used thanks to the devtools.
And to be sure that fonts are loaded, the font cache isn't cleaned up when
the debugger is active.
It's been loaded as a JavaScript module for a long time, and given that the file is bundled as-is (without building) it seems reasonable to just change the file extension now.
At this point in time all browsers, and also Node.js, support standard `import`/`export` statements and we can now finally consider outputting modern JavaScript modules in the builds.[1]
In order for this to work we can *only* use proper `import`/`export` statements throughout the main code-base, and (as expected) our Node.js support made this much more complicated since both the official builds and the GitHub Actions-based tests must keep working.[2]
One remaining issue is that the `pdf.scripting.js` file cannot be built as a JavaScript module, since doing so breaks PDF scripting.
Note that my initial goal was to try and split these changes into a couple of commits, however that unfortunately didn't really work since it turned out to be difficult for smaller patches to work correctly and pass (all) tests that way.[3]
This is a classic case of every change requiring a couple of other changes, with each of those changes requiring further changes in turn and the size/scope quickly increasing as a result.
One possible "issue" with these changes is that we'll now only output JavaScript modules in the builds, which could perhaps be a problem with older tools. However it unfortunately seems far too complicated/time-consuming for us to attempt to support both the old and modern module formats, hence the alternative would be to do "nothing" here and just keep our "old" builds.[4]
---
[1] The final blocker was module support in workers in Firefox, which was implemented in Firefox 114; please see https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Statements/import#browser_compatibility
[2] It's probably possible to further improve/simplify especially the Node.js-specific code, but it does appear to work as-is.
[3] Having partially "broken" patches, that fail tests, as part of the commit history is *really not* a good idea in general.
[4] Outputting JavaScript modules was first requested almost five years ago, see issue 10317, and nowadays there *should* be much better support for JavaScript modules in various tools.
The minified default viewer has never been distributed in either official releases or through pdfjs-dist, which means that it's most likely unused, and it's has never been tested nor actively maintained.
Setting the alpha-value explicitly to `1` in `rgb` colors is unnecessary, since that's the default value, and this way we ever so slightly reduce the size of our CSS files.
Unfortunately I've not found a Stylelint rule to enforce this automatically, and the patch was generated using search and replace.
When an editing button is disabled, focused and the user press Enter (or space), an
editor is automatically added at the center of the current page.
Next creations can be done in using the same keys within the focused page.
This has been deprecated since version `2.15.349`, which is a year ago.
Removing this will also simplify some upcoming changes, specifically outputting of JavaScript modules in the builds.
This removes the only remaining old and non-standard handling of exports in the `web/`-folder, since some initial attempts at outputting JavaScript modules in the builds have identified this file as a potential problem.
While this uses a hard-coded list, for overall simplicity, I don't believe that that's a big problem since:
- Generating this file automatically would require a bunch more parsing *every single time* that the library is built.
- The official API-surface doesn't change often enough for this to really impede development in any significant way.
- The added unit-test helps ensure that this list cannot accidentally become outdated.
Given that this is accessed multiple times per page in the viewer, that leads to a number of (strictly speaking unneeded) function calls and allocated Objects for each invocation. By converting `layerProperties` to a, lazily initialized, Object we can avoid this.
When PR 17015 removed the `disabled` handling for the "Save"-button it left a bunch of now unused CSS rules behind, which seems like a simply oversight.
Rather than shipping "dead" CSS rules, let's remove those until such a time that they're actually needed.
*For many non-English locales the translated strings will be longer, which is easy to forget about during development/review.*
Note how for some locales (e.g. Swedish) the altText-button end up looking horizontally "cramped", hence it seems reasonable to add a bit of inline padding to improve this.
but keep it for the text area.
Disable pointerdown on the alt-text button to disable dragging the editor
when the button is clicked (especially when slightly moving the mouse
between the down and the up).
The dialog element handles closing with <kbd>Esc</kbd> automatically, however we're not reporting telemetry in that case.
In order to fix that the easiest solution, as far as I'm concerned, seem to be moving the telemetry reporting into the dialog-close handler since it's always invoked.
This patch addresses an edge-case that'll probably never happen, but it nonetheless seems like something that we want to fix.
Note how we're using the `#currentEditor`-field to prevent opening the dialog when it's already active, and it being reset once the dialog has been closed.
By also resetting the `#currentEditor`-field during destruction, instead of waiting until the dialog has actually closed (assuming it's currently open), there's a *tiny* window of time[1] during which we could theoretically allow to (incorrectly) re-open the dialog and thus getting out-of-sync state in the viewer-component.
---
[1] Since the "close" event, on a dialog-element, is dispatched asynchronously by the browser.
When the user edit an existing alt-text and remove it, we want to be able
to save this state and consequently remove the done state from the
alt-text button.
Remove the button from its parent when the editor is removed: it should
help to save few Kb of memory.
Radio-buttons can also be toggled by clicking on their associated `label`-elements, and not only the `input`-elements itself, however it seems that "pointerdown" event listeners don't cover that case.
Hence it's possible that telemetry could miss certain cases of a mouse being used, and the easiest solution seem to be to instead use "click" event listeners and just ignore keyboard-based events.
Given the limitations of the old pre-processor that's used for CSS/HTML files, this unfortunately isn't as "easy" to implement as it is for JavaScript code.
Since this is the first case where we've wanted to do conditional CSS imports, rather than trying to completely re-write the pre-processor, this patch settles for handling it explicitly in the `expandCssImports` function.
Looking at the save-telemetry values they're all boolean *except* for "alt_text_edit" in one instance, since `this.#previousAltText` may be an empty string (looking at the `editAltText` method) and this value may thus become an empty string as well.
When closing a document in the viewer, e.g. by running `PDFViewerApplication.close()` in the console, the `AltTextManager.#finish` method currently throws *unless* the `altText` dialog is actually open.
Similar to e.g. the PasswordPrompt, we should thus only attempt to close the `altText` dialog when it's open.
It's a part of the UX specifications. There's a drawing issue in Firefox
(see bug https://bugzilla.mozilla.org/1853288) but setting the
background-clip property to content-box seems to be a good workaround.
The goal is to always have something which is focusable to let the user select
it with the keyboard.
It fixes the mentioned bug because, the annotation layer will now have a container
to attach the canvas for annotations having their own canvas.
`.grab-to-pan-grab:active` is `#viewerContainer` when the mouse is
pressed down. It is supposed to have a `cursor: grabbing` appearance
immediately on mousedown,
`.grab-to-pan-grabbing` is the overlay that is supposed to cover
everything, and also has the `cursor: grabbing` appearance. The "cover
everything" result is achieved through `position:fixed`, `inset:0`, etc.
The block with these CSS properties for "cover everything" is currently
shared by `.grab-to-pan-grab:active` and `.grab-to-pan-grabbing`, but
only "cursor" need to be shared. The original JS and CSS code at
https://github.com/Rob--W/grab-to-pan.js shows that these were supposed
to be associated with the overlay only.
The PR that added this to PDF.js also shows that the "cover everything"
CSS properties were supposed to be limited to the overlay only:
https://github.com/mozilla/pdf.js/pull/4209#discussion-diff-9285917
But the final version of the PR mistakenly merged them together.
This patch rectifies that mistake.
Using `removeNullCharacters` on the URL should be completely redundant, given the kind of data that we're passing to the `addLinkAttributes` helper function. Note that whenever we're handling a URL, originating in the worker-thread, in the viewer that helper function is always being used.
Furthermore, on the worker-thread all URLs are parsed with the `createValidAbsoluteUrl` helper function, which uses `new URL()` to ensure that a valid URL is obtained. Note that the `URL` constructor will either throw, or in some cases just ignore them, when encountering `\u0000`-characters during parsing.
Hence it should be *impossible* for a valid URL to contain `\u0000`-characters and we can thus simplify the viewer-code a tiny bit. The use of `removeNullCharacters` is most likely a left-over from back when `new URL()` wasn't generally available in browsers.
Testing the `tagged_stamp.pdf` document locally in the viewer, I noticed that e.g. the /Alt entry for the StampAnnotation contains "Secondary text for stamp\u0000".
Elsewhere in the viewer we're skipping null-chars and it's easy enough to do that in the `StructTreeLayerBuilder` class as well. (Note that we generally let the API itself return the data as-is.)
This fixes invalid type references (either due to invalid paths for the
import or missing imports) in the JS doc, as well as some missing or
invalid parameter names for @param annotations.
The main stamp button will be used to just enter in a add/edit image mode:
- the user can add a new image in using the new button.
- the user can edit an image in resizing, moving it.
In image mode, when the user clicks outside on the page but not on an editor,
then all the selected editors will be unselected.
After the `src/core/`-changes in PR 16779 the `PDFDocumentProxy.getJSActions` method should no longer be able to return *empty* entries, which means that we can simplify the "JavaScript support is not enabled"-warning in the viewer.
Furthermore, improve the auto-printing hack used when scripting is disabled.
There are 2 rotation we've to deal with: the viewer one and the editor one.
The previous implementation was a bit complex and having to deal with these
rotation would have potentially increase it.
So this patch aims to simplify the implementation and deal with all the possible
cases.
The main idea is to transform the mouse deltas according to the rotations and then
apply the resizing in the page coordinates system.
When resizing an editor we're currently using unidirectional cursors, please refer to https://developer.mozilla.org/en-US/docs/Web/CSS/cursor
Given that editors can (generally) be resized to become either smaller or larger, it seems overall more appropriate to use bidirectional cursors to make this clearer to the user.
Note that as mentioned in the MDN article some environments, which seems to apply to e.g. Windows 11, doesn't differentiate between the two cursor formats and simply use bidirectional ones unconditionally.
One additional benefit of these changes is that the relevant CSS rules become slightly more compact.
