This patch is essentially *another* continuation of PR 11263, which tried to improve loading/initialization performance of *very* large/long documents.
For most documents, unless they're *very* long, we'll eagerly initialize all of the pages in the viewer. For shorter documents having all pages loaded/initialized early provides overall better performance/UX in the viewer, however there's cases where it can instead *hurt* performance.
For documents with a couple of thousand pages[1], the parsing and pre-rendering of the *second* page of the document can be delayed (quite a bit). The reason for this is that we trigger `PDFDocumentProxy.getPage` for *all pages* early during the viewer initialization, which causes the worker-thread to be swamped with handling (potentially) thousands of `getPage`-calls and leaving very little time for other parsing (such as e.g. of operatorLists).
To address this situation, this patch thus proposes temporarily "pausing" the eager `PDFDocumentProxy.getPage`-calls once a threshold has been reached, to give the worker-thread a change to handle other requests.[2]
Obviously this may *slightly* delay the "pagesloaded" event in longer documents, but considering that it's already the result of asynchronous parsing that'll hopefully not be seen as a blocker for these changes.[3]
---
[1] A particularly problematic example is https://github.com/mozilla/pdf.js/files/876321/kjv.pdf (16 MB large), which is a document with 2236 pages and a /Pages-tree that's only *one* level deep.
[2] Please note that I initially considered simply chaining the `PDFDocumentProxy.getPage`-calls, however that'd slowed things down for all documents which didn't seem appropriate.
[3] This patch will *hopefully* also make it possible to re-visit PR 11312, since it seems that changing `Catalog.getPageDict` to an `async` method wasn't the problem in itself. Rather it appears that it leads to slightly different timings, thus exacerbating the already existing issues with the worker-thread being overloaded by `getPage`-calls.
Having recently worked with that method, there's a couple of (very old) issues that I'd also like to address and having `Catalog.getPageDict` be `async` would simplify things a great deal.
These changes improves the consistency ever so slightly in the `PDFOutlineViewer._dispatchEvent` method, by making sure that we can tell the following two cases apart:
- The "pagesloaded" event has *not yet* been fired.
- The "pagesloaded" event has been fired, but no pages were available.
*This patch can be tested e.g. with the `poppler-85140-0.pdf` document from the test-suite.*
For some sufficiently corrupt documents the `getDocument` call will succeed, but fetching even the very first page fails. Currently we only print error messages (in the console) from the `{BaseViewer, PDFThumbnailViewer}.setDocument` methods, but don't actually provide these errors to allow the viewer to handle them properly.
In practice this means that the GENERIC viewer won't display the `errorWrapper`, and in the MOZCENTRAL viewer the *browser* loading indicator is never hidden (since we never unblock the "load" event).
This patch is essentially a continuation of PR 11263, which tried to improve loading/initialization performance of *very* large/long documents.
Note that browsers, in general, don't handle a huge amount of DOM-elements very well, with really poor (e.g. sluggish scrolling) performance once the number gets "large". Furthermore, at least in Firefox, it seems that DOM-elements towards the bottom of a HTML-page can effectively be ignored; for the PDF.js viewer that means that pages at the end of the document can become impossible to access.
Hence, in order to improve things for these *very* large/long documents, this patch will now enforce usage of the (recently added) PAGE-scrolling mode for these documents. As implemented, this will only happen once the number of pages *exceed* 15000 (which is hopefully rare in practice).
While this might feel a bit jarring to users being *forced* to use PAGE-scrolling, it seems all things considered like a better idea to ensure that the entire document actually remains accessible and with (hopefully) more acceptable performance.
Fixes [bug 1588435](https://bugzilla.mozilla.org/show_bug.cgi?id=1588435), to the extent that doing so is possible since the document contains 25560 pages (and is 197 MB large).
This was added on the assumption that the viewer would (eventually) start using the `PDFSinglePageViewer` for e.g. PAGE-scrolling mode and PresentationMode. However, having both a `PDFViewer` and a `PDFSinglePageViewer` side-by-side in the viewer would've been tricky to implement well, which is why PR 14112 implemented PAGE-scrolling for the general `BaseViewer` instead.
Given that the default viewer is no longer (potentially) going to use `PDFSinglePageViewer`, there's code in the `SecondaryToolbar` (and related CSS rules) which is now unnecessary.
*This patch can be tested e.g. with the `sizes.pdf` document in the test-suite.*
While this patch isn't necessarily the best solution, e.g. it might be possible to solve this with *only* CSS, it's what I was able to come up with to address an old issue.
The solution here re-uses the `spread`-class in PresentationMode, since that one already takes care of centering pages *vertically*, together with a dummy-page that takes up the entire height of the window.
Finally, some PresentationMode-related CSS-rules are also simplified slightly, since the changes in PR 14112 (using Page-scrolling) allows some clean-up here.
The issue that this patch fixes has existed ever since the viewer was first re-factored into components, however it only really affects the `disableAutoFetch = true` mode.
By default we're fetching all pages in `BaseViewer.setDocument`, and as part of the parsing/initialization we're also populating the `PDFLinkService`-pagesRefCache. The purpose of that cache is to make navigating to any internal destinations faster, by not having to (asynchronously) lookup the pageNumber via the API when handling the destination.
In comparison, when the `disableAutoFetch = true` mode is being used we're instead *lazily* initializing the pages in the `BaseViewer.#ensurePdfPageLoaded`-method. For some reason, that I can only assume is a simple oversight, we're not attempting to update the `PDFLinkService`-pagesRefCache in that case.
In the `BaseViewer` this cache is mostly relevant in the `disableAutoFetch = true` mode, since the pages are being initialized *lazily* in that case.
In the `PDFThumbnailViewer` this cache is mostly used for thumbnails that are actually being rendered, as opposed to those created directly from the "regular" pages.
Please note that I'm not suggesting that we remove these caches because they're only used in some situations, but rather because they're for all intents and purposes actually *redundant*. In the API itself, we're already caching both the page-promises and the actual pages themselves on the `WorkerTransport`-instance.
Hence these viewer-caches aren't really necessary in practice, and adds what to me mostly seems like an unnecessary level of indirection.[1]
Given that the viewer now relies on caching in the API itself, this patch also adds a new unit-test to ensure that page-caching works (and keep working) as expected.
---
[1] In the `WorkerTransport.getPage`-method the parameter is being validated on every call, but that's hardly enough code to warrant keeping the "duplicate" caches in the viewer in my opinion.
*Please note:* These changes will primarily benefit longer documents, somewhat at the expense of e.g. one-page documents.
The existing `PDFDocumentProxy.getStats` function, which in the default viewer is called for each rendered page, requires a round-trip to the worker-thread in order to obtain the current document stats. In the default viewer, we currently make one such API-call for *every rendered* page.
This patch proposes replacing that method with a *synchronous* `PDFDocumentProxy.stats` getter instead, combined with re-factoring the worker-thread code by adding a `DocStats`-class to track Stream/Font-types and *only send* them to the main-thread *the first time* that a type is encountered.
Note that in practice most PDF documents only use a fairly limited number of Stream/Font-types, which means that in longer documents most of the `PDFDocumentProxy.getStats`-calls will return the same data.[1]
This re-factoring will obviously benefit longer document the most[2], and could actually be seen as a regression for one-page documents, since in practice there'll usually be a couple of "DocStats" messages sent during the parsing of the first page. However, if the user zooms/rotates the document (which causes re-rendering), note that even a one-page document would start to benefit from these changes.
Another benefit of having the data available/cached in the API is that unless the document stats change during parsing, repeated `PDFDocumentProxy.stats`-calls will return *the same identical* object.
This is something that we can easily take advantage of in the default viewer, by now *only* reporting "documentStats" telemetry[3] when the data actually have changed rather than once per rendered page (again beneficial in longer documents).
---
[1] Furthermore, the maximium number of `StreamType`/`FontType` are `10` respectively `12`, which means that regardless of the complexity and page count in a PDF document there'll never be more than twenty-two "DocStats" messages sent; see 41ac3f0c07/src/shared/util.js (L206-L232)
[2] One example is the `pdf.pdf` document in the test-suite, where rendering all of its 1310 pages only result in a total of seven "DocStats" messages being sent from the worker-thread.
[3] Reporting telemetry, in Firefox, includes using `JSON.stringify` on the data and then sending an event to the `PdfStreamConverter.jsm`-code.
In that code the event is handled and `JSON.parse` is used to retrieve the data, and in the "documentStats"-case we'll then iterate through the data to avoid double-reporting telemetry; see https://searchfox.org/mozilla-central/rev/8f4c180b87e52f3345ef8a3432d6e54bd1eb18dc/toolkit/components/pdfjs/content/PdfStreamConverter.jsm#515-549
It shouldn't be necessary to iterate through *all* pages when using a non-default `spreadMode`, since we already know which page(s) should become visible.
This code is a left-over from the initial (local) implementation that resulted in PR 14112, however I forgot to clean-up some things such as e.g. this loop.
Also fixes an outdated comment, see PR 14204 which removed the mentioned data-structure.
This patch preserves the old behaviour of appending a `loadingIcon`-div to all pages that are not yet loaded/rendered. However, the actual `loadingIcon`-spinner (i.e. the `loading-icon.gif` image) will only be displayed on *visible* pages to improve performance.
To avoid having to iterate through all pages in the document, which doesn't seem like a good idea for a PDF document with thousands of pages, we use a combination of the currently visible *and* cached pages to toggle the `loadingIcon`-spinner.
This code is the last piece[1] of the viewer that's not using standard `class`es, and by converting this code we get rid of some now unneeded boilerplate code (slightly reducing the size of the *built* `web/viewer.js` file).
Note that while this code was originally imported from a separate repository, it was last sync-ed with upstream *five years* ago which is why this re-factoring should be OK as far as I'm concerned (and we've done some other clean-up since then as well).
---
[1] Technically the `web/debugger.js` file is left as well, however that code is first of all not bundled in the *built* `web/viewer.js` file and secondly it's not even loaded by default either.
- First step to fix https://bugzilla.mozilla.org/show_bug.cgi?id=1737260;
- several interactive pdfs use the possibility to hide/show buttons to show different icons;
- render pushbuttons on their own canvas and then insert it the annotation_layer;
- update test/driver.js in order to convert canvases for pushbuttons into images.
Reporting telemetry, in Firefox, includes using `JSON.stringify` on the data and then sending an event to the `PdfStreamConverter.jsm`-code.
In that code the event is handled and `JSON.parse` is used to retrieve the data, and in the "pageInfo"-case we'll then proceed to ignore everything except *the first* such event; see https://searchfox.org/mozilla-central/rev/24fac1ad31fb9c6e9c4c767c6a7ff45d226078f3/toolkit/components/pdfjs/content/PdfStreamConverter.jsm#509-514
All-in-all, sending the "pageInfo" telemetry for each rendered page is thus unnecessary and this patch makes the viewer send it only *once* instead.
*This is a tentative patch, since we unfortunately cannot easily test it (as far as I can tell).*
In Firefox this (obviously) works as-is, but in Google Chrome the `markedContent` spans are inserted within the regular text-content (in the DOM) and with non-zero heights.
*This is a tentative patch, since I don't have the necessary hardware to test it.*
See https://developer.mozilla.org/en-US/docs/Web/CSS/text-size-adjust, which is currently ignored in Firefox.
It seems overall safer, and more future-proof, to simply add this to the *entire* `textLayer` rather than its individual elements.
With the previous patch, this helper function is no longer used and keeping it around will simply increase the size of the builds.
This removal is purposely done *separately*, to make it easy to revert the patch in the future if this helper function would become useful again.
This relies on the fact that `Set`s preserve the insertion order[1], which means that we can utilize an iterator to access the *first* stored view.
Note that in the `resize`-method, we can now move the visible pages to the back of the buffer using a single loop (hence we don't need to use the `moveToEndOfArray` helper function any more).
---
[1] This applies to `Map`s as well, although that's not entirely relevant here.
The `PDFPageViewBuffer`-code is very important for the correct function of the viewer, but it's currently not tested at all.
While the `PDFPageViewBuffer` is obviously intended to be used with `PDFPageView`-instances, it only accesses a couple of `PDFPageView` properties/methods and consequently it's fairly easy to unit-test this code with dummy-data.
These unit-tests should help improve our confidence in this code, and will also come in handy with other changes that I'm working on (regarding modernizing and re-factoring the `PDFPageViewBuffer`-code).
The way that we're currently handling the last-`id` is very old, and there's no longer any good reason to special-case things when only one thumbnail is visible.
Furthermore, we can also modernize the loop slightly by using `for...of` instead of `Array.prototype.some()` when checking for fully visible thumbnails.
Note how in `PDFPageViewBuffer.resize` we're manually iterating through the visible pages in order to build a Set of the visible page `id`s. By instead moving the building of this Set into the `getVisibleElements` helper function, as part of the existing parsing, this code becomes *ever so slightly* more efficient.
Furthermore, more direct access to the visible page `id`s also come in handy in other parts of the viewer as well.
In the `BaseViewer.isPageVisible` method we no longer need to loop through the visible pages, but can instead directly check if the pageNumber is visible.
In the `PDFRenderingQueue.getHighestPriority` method, when checking for "holes" in the page layout, we can also avoid some unnecessary look-ups this way.
*Sometimes I'll hopefully learn to optimize my code directly when writing it, rather than having to do multiple clean-up passes; sorry about the churn here!*
For most page layouts there won't be any "holes" in the visible pages (or thumbnails), and in those cases it'd obviously be preferable not having to repeat any checks of already rendered pages.
Rather than only checking the "distance" between the first/last pages, we can instead compare the theoretical number of pages (between first/last) with the actually visible number of pages instead. This way, we're able to better detect the "holes"-case and can skip unnecessary parsing in the common case.
I missed this one spot in PR 12870, when converting the other cases in the "keydown" event handler. However, given that it only matters in PresentationMode and/or when "page-fit" zooming is enabled, this oversight shouldn't have had any user-observable impact (but we should fix it nonetheless).
This is a follow-up to PR 10675, since there I completely overlooked that we also need to handle the case where a PDF document has *failed* to load when the "supportsRangedLoading" message is sent to the viewer.
There were some links not working in some XFA files,I realized that the anchor tag that contains the link has an inline display and couldn't receive any height, solved this by adding a "position: absolute". Tested with two different files in Firefox Nightly and Chrome and now all links are working perfectly fine. Added reftest to avoid future regressions
The only reason for using a `DocumentFragment` in the first place, originally added in PR 8724, was to prevent errors in the `PDFPageView`-constructor. However, we should be able to simply make its `container`-option *optional* instead, since it's not being used for anything else in the class.
Note that pre-rendering still works correctly in my testing, and given that the `BaseViewer` keeps references to all `PDFPageView`-instances (via its `_pages` Array) it also shouldn't be possible to "lose" any pages/canvases this way.
Unfortunately there exist PDF documents where all pageLabels are empty strings, see e.g. http://www.cs.cornell.edu/~ragarwal/pubs/blk-switch.pdf (taken from an old issue), which result in the pageNumber-input being completely blank. That doesn't seem very helpful, and this patch simply extends the approach used to ignore pageLabels that are identical to standard page numbering.
In the GENERIC viewer, e.g. when dragging-and-dropping a new PDF document which automatically opens the outline, there can now be breaking errors in the `{BaseViewer, PDFThumbnailViewer}.#getScrollAhead` methods since there's no visible pages/thumbs during loading; sorry about the breakage!
This was a stupid oversight on my part, since the first/last visible pages have obviously already been rendered at the point when we're checking for any potential "holes" in the page layout.
While this will obviously not have any measurable effect on performance, we should nonetheless avoid doing completely unnecessary checks here.
This is a very old "issue", which has existed since essentially forever, and it affects all of the available scrollModes. However, in the recently added Page-mode it's particularily noticeable since we use a *simulated* scroll direction there.
When deciding what page(s) to pre-render, we only consider the current scroll direction. This works well in most cases, but can break down at the start/end of the document by trying to pre-render a page *outside* of the existing ones. To improve this, we'll thus *force* the scroll direction at the start/end of the document.
*Steps to reproduce:*
0. Open the viewer, e.g. https://mozilla.github.io/pdf.js/web/viewer.html
1. Enable vertical scrolling.
2. Press the <kbd>End</kbd> key.
3. Open the devtools and, using the DOM Inspector, notice how page 13 is *not* being pre-rendered.
*Please note:* This is a tentative patch, since I don't have the necessary a11y-software to actually test it.
To avoid having to add a new API-method just for a single string, I figured that adding the new property to the existing `documentInfo`-data (accessed via `PDFDocumentProxy.getMetadata` in the API) will hopefully be deemed acceptable.
Looking at the code, I do have to agree with the point made in issue 12731 about it being unexpected/unhelpful that the `PDFFindController.executeCommand`-method isn't directly usable with the "find"-event.
The reason for it being this way is, as so often, for historical reasons: The `executeCommand`-method was added (just) prior to the introduction of the `EventBus` in the viewer.
Obviously we cannot simply change the existing `PDFFindController.executeCommand`-method, since that'd be a breaking change in code which has existed for over five years.
Initially I figured that we could simply add a new method in `PDFFindController` that'd accept the state from the "find"-event, however after thinking about this and looking through the use-cases in the default viewer I settled on a slightly different approach: Let the `PDFFindController` just listen for the "find"-event (on the `EventBus`-instance) directly instead, which also removes one level of (unneeded) indirection during searching in the default viewer.
For GENERIC builds of the PDF.js library, the old `PDFFindController.executeCommand`-method is still available with a deprecation warning.
Many years ago now there were some `Promise` implementations that had issues resolving with an *implicitly* `undefined` value. That should no longer be the case, and we've not been using the `Promise.resolve(undefined)` format for a long time, hence this patch fixes the few remaining cases.
Having recently worked with this code, in PR 14096 (and indirectly in PR 14112), I happened to notice a pre-existing issue with spreadModes at higher zoom levels.
The `PDFRenderingQueue` code was written back when the viewer only supported "normal" vertical scrolling, and some edge-cases related to spreadModes are thus not perfectly supported. Depending on the zoom level, it's possible that there are "holes" in the currently visible page layout, and those pages will not be pre-rendered as you'd expect.
*Steps to reproduce:*
0. Open the viewer, e.g. https://mozilla.github.io/pdf.js/web/viewer.html
1. Enable vertical scrolling.
2. Enable the ODD spreadMode.
3. Scroll down, such that both pages 1 and 3 are visible.
4. Zoom-in until *only* page 1 and 3 are visible.
5. Open the devtools and, using the DOM Inspector, notice how page 2 is *not* being pre-rendered despite all surrounding pages being rendered.
With the previous commit, both of the `PDFViewer` and `PDFSinglePageViewer` clases are now small/simple enough that it no longer seems necessary to keep them in separate files.
This implements a new Page scrolling mode, essentially bringing (and extending) the functionality from `PDFSinglePageViewer` into the regular `PDFViewer`-class. Compared to `PDFSinglePageViewer`, which as its name suggests will only display one page at a time, in the `PDFViewer`-implementation this new Page scrolling mode also support spreadModes properly (somewhat similar to e.g. Adobe Reader).
Given the size and scope of these changes, I've tried to focus on implementing the basic functionality. Hence there's room for further clean-up and/or improvements, including e.g. simplifying the CSS/JS related to PresentationMode and implementing easier page-switching with the mouse-wheel/arrow-keys.
Note that PR 14049 removed this, since https://github.com/mozilla/pdf.js/pull/14049#discussion_r716245518 claimed that it's not necessary anymore. Unfortunately that, in my testing on Windows, actually re-introduced exactly the issue described in the comment; more specifically once the *last* character has been entered in the comb-field it's now again incorrectly scrolled (with the first character being invisible) until focus is lost.
This can be tested with e.g. `f1040.pdf`, see page 2, from the test-suite.
This patch helps reduce some duplication, given that we now have a few essentially identical `addLinkAttributes` call-sites in the code-base.
To prevent runtime errors in the Annotation/XFA-layer code, we'll warn if a custom/incomplete `PDFLinkService` is being used (limited to GENERIC builds).
Please note that we (obviously) don't want to unconditionally pre-render more than one page all the time, since that could very easily lead to overall worse performance in some documents.[1]
However, when spreadModes are enabled it does make sense to attempt to pre-render both of the pages of the next/previous spread.
---
[1] Since it may cause pre-rendering to unnecessarily compete for parsing resources, on the worker-thread, with "regular" rendering.
Note how both the annotationLayer and the document outline will apply various URL-related options when creating the link-elements.
For consistency the `xfaLayer`-rendering should obviously use the same options, to ensure that the existing options are indeed applied to all URLs regardless of where they originate.
- it aims to fix https://bugzilla.mozilla.org/show_bug.cgi?id=1716758;
- some buttons have a JS action with the pattern `app.launchURL(...)` (or similar) so extract when it's possible the url and generate a <a> element with the href equals to the found url;
- pdf.js already had some code to handle that so this patch slightly refactor that.
This replaces direct `document.getElementsByName` lookups with a helper method which:
- Lets the AnnotationLayer use the data returned by the `PDFDocumentProxy.getFieldObjects` API-method, such that we can directly lookup only the necessary DOM elements.
- Fallback to using `document.getElementsByName` as before, such that e.g. the standalone viewer components still work.
Finally, to fix the problems reported in issue 14003, regardless of the code-path we now also enforce that the DOM elements found were actually created by the AnnotationLayer code.
With these changes we'll thus be able to update form elements on all visible pages just as before, but we'll additionally update the AnnotationStorage for not-yet-rendered elements thus fixing a pre-existing bug.
Currently any AppOptions set using e.g. the "webviewerloaded" event listener can/will by default be overridden when the Preferences are read.
To avoid that happening the "disablePreferences"-option can be used, however unless it's been explicitly set all non-default AppOptions will be silently ignored. This patch thus attempts to improve the current situation somewhat, for third-party implementations, by logging a warning in the console when this happens.
Rather than re-computing this value in a number of different places throughout the code-base[1], we can expose this in the API via the existing `PixelsPerInch`-structure instead.
There's also been feature requests asking for the old `CSS_UNITS` viewer constant to be made accessible, such that it could be used in third-party implementations.
I suppose that it could be argued that it's somewhat confusing to place a unitless property in `PixelsPerInch`, however given that the `PDF_TO_CSS_UNITS`-property is defined strictly in terms of the existing properties this is hopefully deemed reasonable.
---
[1] These include:
- The viewer, with the `CSS_UNITS` name.
- The reference-tests.
- The display-layer, when rendering images; see PR 13991.
Given the simplicity of this functionality, we can move it from the default viewer and into the `BaseViewer` class instead. This way, it's possible to support more scripting functionality in the standalone viewer components; please see PR 14038.
Please note that I purposely went with `increaseScale`/`decreaseScale`-method names, rather than using "zoom", to better match the existing `currentScale`/`currentScaleValue` getters/setters that's being used in the `BaseViewer` class.
Rather than forcing the "regular" `EventBus` to check and handle `isInAutomation` for every `dispatch` call, we can take advantage of subclassing instead.
Hence this PR introduces a new `AutomationEventBus` class, which extends `EventBus`, and is used by the default viewer when `isInAutomation === true`.
Originally the library/viewer didn't support forms, and the hiding of the Download-buttons (when new documents are opened) didn't really matter all that much. Hence the simplest solution, at the time, was to hide the Download-buttons since we use the URL as a fallback when downloading data in the GENERIC `DownloadManager`.
Nowadays we obviously want to support saving of forms in the GENERIC viewer, regardless of how the document was opened, which could thus *potentially* lead to the fallback download-URL being wrong.
In order to be able to show the Download-buttons unconditionally, this patch slightly re-factors the viewer to track the download-URL *separately* to prevent any issues there.
*Please note:* As mentioned in the issue, the ViewBookmark-buttons are specific to the initial URL when the viewer is first opened. Hence they (still) don't make sense when a new document has been opened, since it's then impossible to obtain a usable link to the *currently* active document.
By using CSS variables to set the width of the zoom dropdown, we can simplify both the relevant CSS and JS code. This will not only improve overall maintainability of this code, but should also make it (slightly) easier for third-party users that want to customize the width.
Note in particular that by having the code in `Toolbar._adjustScaleWidth` lookup the values of the CSS variables, we no longer need to worry about keeping hard-coded values up-to-date with the CSS rules.
While some of the output looks worse to my eye, this behavior more
closely matches what I see when I open the PDFs in Adobe acrobat.
Fixes: #4706, #9713, #8245, #1344
Similar to other viewer components, e.g. the `PDFFindBar` and `PDFPresentationMode`, there's no need to create a `PDFHistory`-instance when it's not going to be used.
The old `update`-signature started to annoy me back when I added optional content support to the viewer, since we're (often) forced to pass in a bunch of arguments that we don't care about whenever these methods are called.
This is tagged `api-minor` since `PDFPageView` is being used in the `pageviewer` component example, and it's thus possible that these changes could affect some users; the next commit adds fallback handling for the old format.
While e.g. the `simpleviewer` and `singlepageviewer` examples work, since they're based on the `BaseViewer`-class, the standalone `pageviewer` example currently doesn't support either XFA- or StructTree-layers. This seems like an obvious oversight, which can be easily addressed simply by exporting the necessary functionality through `pdf_viewer.component.js`, similar to the existing Text/Annotation-layers.
While working on, and testing, these changes I happened to notice a number of smaller things that's also fixed in this patch:
- Ensure that `XfaLayerBuilder.render` always have a *consistent* return type, to prevent possible run-time failures in `PDFPageView`; PR 13908 follow-up.
- Change the order of the options in the `XfaLayerBuilder`-constructor to agree with the parameter order in the `DefaultXfaLayerFactory.createXfaLayerBuilder`-method.
- Add a missing `textHighlighterFactory`-option, in the JSDocs for the `PDFPageView`-class.
- A couple of small tweaks in the `TextLayerBuilder.render`-method: Re-use an existing Array rather than creating a new one, and replace an `if` with optional chaining instead.
*Please note:* For now XFA-support is currently disabled by default, similar to the regular viewer.
A lot of the code in this getter has existed ever since the initial PresentationMode-implementation was first added all the way back in PR 1938 (which is nine years ago now).
At this point in time however, there's now a simpler way detect if a browser supports the FullScreen API and we should thus be able to simplify this getter; please refer to https://developer.mozilla.org/en-US/docs/Web/API/Document/fullscreenEnabled#browser_compatibility
The Viewer API definitions do not compile because of missing imports and
anonymous objects are typed as `Object`. These issues were not caught
during CI because the test project was not compiling anything from the
Viewer API.
As an example of the first problem:
```
/**
* @implements MyInterface
*/
export class MyClass {
...
}
```
will generate a broken definition that doesn’t import MyInterface:
```
/**
* @implements MyInterface
*/
export class MyClass implements MyInterface {
...
}
```
This can be fixed by adding a typedef jsdoc to specify the import:
```
/** @typedef {import("./otherFile").MyInterface} MyInterface */
```
See https://github.com/jsdoc/jsdoc/issues/1537 and
https://github.com/microsoft/TypeScript/issues/22160 for more details.
As an example of the second problem:
```
/**
* Gets the size of the specified page, converted from PDF units to inches.
* @param {Object} An Object containing the properties: {Array} `view`,
* {number} `userUnit`, and {number} `rotate`.
*/
function getPageSizeInches({ view, userUnit, rotate }) {
...
}
```
generates the broken definition:
```
function getPageSizeInches({ view, userUnit, rotate }: Object) {
...
}
```
The jsdoc should specify the type of each nested property:
```
/**
* Gets the size of the specified page, converted from PDF units to inches.
* @param {Object} options An object containing the properties: {Array} `view`,
* {number} `userUnit`, and {number} `rotate`.
* @param {number[]} options.view
* @param {number} options.userUnit
* @param {number} options.rotate
*/
```
- Use `Node.TEXT_NODE` rather than a magical constant, in `TextHighlighter._convertMatches`, to improve readability. According to MDN, this has been supported since "forever": https://developer.mozilla.org/en-US/docs/Web/API/Node/nodeType#browser_compatibility
- Remove the `pageIdx`-property, on `TextLayerBuilder`-instances, since the re-factoring in PR 13908 meant that it's now unused.
- Remove the `matches`-property, on `TextLayerBuilder`-instances, since the re-factoring in PR 13908 meant that it's now unused.
Generalizing, and documenting, the `PDFHistory`-implementation as part of the web-interfaces doesn't seem entirely necessary and in hindsight I'm not entirely sure why we need it since:
- The `PDFHistory` implementation is/was written specifically for the default viewer use-case, which is why e.g. the `simpleviewer` component example isn't using it. (While it *could* be used there, it'd need to be manually created/initialized correctly.)
- There's only *one* `PDFHistory`-implementation present (and no other viewer-component will fail without it being available), as opposed to the other web-interfaces documented in this file.
- The `PDFHistory` implementation is not even usable with e.g. the `pageviewer` component example, since it (obviously) requires a complete viewer to work. (This is in contrast to e.g. `IPDFTextLayerFactory` and `IPDFAnnotationLayerFactory`.)
*This is a follow-up to PRs 13867 and 13899.*
This patch is tagged `api-minor` for the following reasons:
- It replaces the `renderInteractiveForms`/`includeAnnotationStorage`-options, in the `PDFPageProxy.render`-method, with the single `annotationMode`-option that controls which annotations are being rendered and how. Note that the old options were mutually exclusive, and setting both to `true` would result in undefined behaviour.
- For improved consistency in the API, the `annotationMode`-option will also work together with the `PDFPageProxy.getOperatorList`-method.
- It's now also possible to disable *all* annotation rendering in both the API and the Viewer, since the other changes meant that this could now be supported with a single added line on the worker-thread[1]; fixes 7282.
---
[1] Please note that in order to simplify the overall implementation, we'll purposely only support disabling of *all* annotations and that the option is being shared between the API and the Viewer. For any more "specialized" use-cases, where e.g. only some annotation-types are being rendered and/or the API and Viewer render different sets of annotations, that'll have to be handled in third-party implementations/forks of the PDF.js code-base.
Moves the logic out of TextLayerBuilder to handle
highlighting matches into a new separate class `TextHighlighter`
that can be used with regular PDFs and XFA PDFs.
To mimic the current find functionality in XFA, two arrays
from the XFA rendering are created to get the text content
and map those to DOM nodes.
Fixes#13878
The `loadAndEnablePDFBug` helper function, in `web/app.js`, can be simplified a little bit by making it `async`. Furthermore, given how `PDFBug` is being used, we can also (slightly) re-factor `PDFBug.init` such that the `PDFBug.enable`-call is done internally rather than having to handle that manually at the call-site.
(Finally, utilize `await` more in the `loadFakeWorker` helper function.)
The original idea behind including the class name, when logging errors, was to improve things in the *hypothetical case* where `PDFViewer`- and `PDFSinglePageViewer`-instances would be used side-by-side.
Given that all of the relevant methods are synchronous this seem unlikely to really be necessary, and furthermore it's probably best to avoid using `this.constructor.name` since that's not guaranteed to do what you intend (we've seen repeated issues with minifiers mangling function/class names).
This patch removes the only remaining closure in the `src/display/api.js` file, utilizing a similar approach as used in lots of other parts of the code-base, which results in a small decrease in the size of the *build* `pdf.js` file.
Given that `PDFWorker` is exposed through the *public* API, this complicates things somewhat since there's a couple of worker-related properties that really should stay *private*. Initially, while working on PR 13813, I believed that we'd need support for private (static) class fields in order to get rid of this closure, however I've managed to come up with what's hopefully deemed an acceptable work-around here.
Furthermore, some helper functions were simply moved into the `PDFWorker` class as static methods, thus simplifying the overall implementation (e.g. we don't need to manually cache the Promise in the `PDFWorker._setupFakeWorkerGlobal`-method).
Finally, as part of this re-factoring a number of missing JSDoc-comments were added which *together* with the removal of the closure significantly improves the `gulp jsdoc` output for the `PDFWorker` class.
*Please note:* This patch is tagged with `api-minor` since it deprecates `PDFWorker.getWorkerSrc()` in favor of the shorter `PDFWorker.workerSrc`, with the fallback limited to `GENERIC` builds.
Without this patch, when using `PDFPageView` directly[1] this CSS variable won't be updated and consequently things won't work as intended.
This is purposely implemented such that when a `PDFPageView`-instance is part of a viewer, we don't repeatedly set the CSS variable for every single page.
---
[1] See e.g. the "pageviewer" example in the `examples/components/` folder.
Even though the code as-is *should* be safe, given that we're using an Object with a `null` prototype, it cannot hurt to change this to a Map to prevent any issues (since we're parsing unknown and potentially unsafe data).
Overall I also think that these changes improve the `parseQueryString` call-sites, since we now have a proper way of checking for the existence of a particular key (and don't have to use `in` which stringifies the keys in the Object).
This patch also changes the default, when no `value` exists, from `null` to an empty string since the use of `decodeURIComponent` currently can modify the value in a somewhat surprising way (at least to me).
Note how `decodeURIComponent(null) === "null"` which is unlikely to be what you actually want, whereas `decodeURIComponent("") === ""` which seems much more helpful.
Prior to PR 13042, when scripting wasn't really possible to use outside of the full viewer, the `enableScripting` option made sense.
However, at this point in time having to both pass in a `PDFScriptingManager`-instance *and* set the `enableScripting`-boolean when creating a `BaseViewer`-instance feels redundant and (mostly) annoying. Hence this patch, which removes the *separate* boolean and always enables scripting when `scriptingManager` is provided.
The relevant "viewer component" examples are also updated (with a comment), but in such a way that scripting support won't just break when used with the current PDF.js releases.
- All the scale factors in for the substitution font were wrong because of different glyph positions between Liberation and the other ones:
- regenerate all the factors
- Text may have polish chars for example and in this case the glyph widths were wrong:
- treat substitution font as a composite one
- add a map glyphIndex to unicode for Liberation in order to generate width array for cid font
- In order to better compute text fields size, use line height with no gaps (and consequently guessed height for text are slightly better in general).
- Fix default background color in fields.
Given that we've over time been reducing the number of `compatibilityParams` in use, there's now few enough left that I think it makes sense to simply inline them directly in the `web/app_options.js` file.
Note that we recently inlined/removed the separate `src/display/api_compatibility.js` file, see PR 13525, and that it (in my opinion) thus makes sense to do the same in the `web/`-folder. This patch will also slightly reduce the size of *built* `web/viewer.js` file, which cannot hurt.
There's no good reason, as far as I can tell, to have `PDFPageView.reset` attempt to cancel `annotationLayer`/`xfaLayer` rendering in one special-case (this is mostly a leftover from older code). Previously cancelling was moved into the separate `PDFPageView.cancelRendering`-method, and by slightly tweaking the conditions there we're able to remove a bit of now unnecessary code from the `PDFPageView.reset`-method.
Originally the `xfaLayer` wasn't implemented in such a way that it supported being removed from the DOM when pages were evicted from the cache, however this limitation was lifted in PR 13427 and the `xfaLayer` should thus be handled similar to e.g. the `annotationLayer`.
In addition to removing the `xfaLayer` from the DOM, this patch *also* implements proper rendering/hiding-handling for it (mirroring the `annotationLayer`-code).
*Please note:* This patch is tagged API-minor just in case[1], since it changes the signatures of a couple of `PDFPageView`-methods to improve readability of the code.
---
[1] Although users are *hopefully* not directly accessing any of the affected methods, and are rather using e.g. `PDFViewer` in which case none of these changes will matter.
- a rectangle can be defined after a field and so from a z-index pov, it's on top of the field, which means that the rectangle avoid to have some mouse events in the fields;
- so this patch just disable pointer events for all elements under svg (included).
Given that Internet Explorer is, since some time now, no longer supported in the PDF.js library this `<meta>` tag can also be removed (added in PR 6374).
- font line height is taken into account by acrobat when it isn't with masterpdfeditor: I extracted a font from a pdf, modified some ascent/descent properties thanks to ttx and the reinjected the font in the pdf: only Acrobat is taken it into account. So in this patch, line heights for some substituted fonts are added.
- it seems that Acrobat is using a line height of 1.2 when the line height in the font is not enough (it's the only way I found to fix correctly bug 1718741).
- don't use flex in wrapper container (which was causing an horizontal overflow in the above bug).
- consequently, the above fixes introduced a lot of small regressions, so in order to see real improvements on reftests, I fixed the regressions in this patch:
- replace margin by padding in some case where padding is a part of a container dimensions;
- remove some flex display: some containers are wrongly sized when rendered;
- set letter-spacing to 0.01px: it helps to be sure that text is not broken because of not enough width in Firefox.
While I don't know if it's technically correct to even do this, it could provide a slightly better out-of-the-box behaviour in browsers that specify (from the PDF.js `l10n`-folder perspective) "incomplete" language codes.
Rather than immediately falling back to English, we'll use a white-list to try and re-write a "partial" language code to a (hopefully) suitable one that matches an existing `l10n`-folder. The disadvantage of this solution is that the list needs to be kept *manually* up-to-date with any changes in the `l10n`-folder, however new locales are added infrequently enough that this should be acceptable.
Fixes 13689 (assuming we actually want/care to do so, otherwise we should just WONTFIX the issue).
- Fix a typo in order to open the pdf in issue #13679
- After fixing the fill default color there wer some regressions because of z-index
and when fixing z-index there were some regressions because of borders
- So fix the borders rendering.
The PDF.js API has only ever supported accessing the original file ID, however the second one that (should) exist in *modified* documents have thus far been completely inaccessible through the API.
That seems like a simple oversight, caused e.g. by the viewer not needing it, since it really shouldn't hurt to provide API-users with the ability to check if a PDF document has been modified since its creation.[1]
Please refer to https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf#G13.2261661 for additional information.
For an example of how to update existing code to use the new API, please see the changes in the `web/app.js` file included in this patch.
*Please note:* While I'm not sure if we'll ever be able to remove the old `PDFDocumentProxy.fingerprint` getter, given that it's existed since "forever", that probably isn't a big deal given that it's now limited to only `GENERIC`-builds.
---
[1] Although this obviously depends on the PDF software following the specification, by updating the second file ID as intended.
Using `instanceof Object` is generally problematic, since it's not guaranteed to always do the right thing for all Objects.
(I stumbled upon this while working on another patch, when I noticed that the `outlineView` was broken with workers disabled.)
- support paragraph margins, line height, letter spacing, ...
- compute missing dimensions from fields based almost on the dimensions of caption contents.
- it aims to fix#13583;
- fix the switch to breakBefore target;
- force the layout of an unsplittable element on an empty page;
- don't fail when there is horizontal overflow (except in lr-tb);
- handle correctly overflow in the same content area (bug 1717805, bug 1717668);
- fix a typo in radial gradient first argument.
Given that this property is only used with password protected documents, and is consequently document-specific rather than viewer-specific, ensure that `IPDFLinkService.externalLinkEnabled` is actually being reset by `PDFViewerApplication.close`.
To make things less confusing/inconsistent, remove the *undocumented* `externalLinkEnabled` property from the `PDFLinkService` constructor and force it to always be manually set when needed.
- when the CSS line-height property is set to 'normal' then the value depends of the user agent. So use a line height based on the font itself and if for any reasons this value is not available use 1.2 as default.
- it's a partial fix for https://bugzilla.mozilla.org/show_bug.cgi?id=1717681.
This patch provides an overall simpler *and* more consistent way of handling the `viewport` parameter during printing of XFA forms, since it's now again guaranteed to always be an instance of `PageViewport`.
Furthermore, for anyone attempting to e.g. implement custom printing of XFA forms this probably cannot hurt either.
Given that the "print"-intent is special, and that we should always fallback to the "display"-intent, let's ensure that the code actually reflects that.
Also, ensure that the method always returns a `Promise` since that's what the documentation says.
The `web/ui_utils.js` file should be usable from basically anywhere in the `web/`-folder, hence it should ideally not have any dependecies on its own and particularily *not* onces that pull in entire (large) factories.
This extends the approach in PRs 12848 and 13606 to also apply to the `xfaLayer`, since otherwise XFA forms will be similarly broken in most non-default scroll/spread modes.
Note that as far as I can tell, this is *not* a regression but rather a bug which has existed since basically "forever".
**In order to reproduce this easily:**
- Open the viewer.
- Set the zoom level to `400%`,
- Search for "expression".
The problem here is that when scrolling matches into view, we're scrolling to the start of the *containing* `textLayer` element rather than the start of the highlighted match itself.[1] When the entire width (or at least most) of the page is visible in the viewer, that doesn't really matter though which is likely why this bug has gone unnoticed for so long.[2]
Given that the highlighted match can be placed anywhere, e.g. even at the very end, within its `textLayer` element it's quite easy to see why the current implementation becomes a problem at higher zoom levels. All of this is then *further* exacerbated by `PDFFindController.scrollMatchIntoView` using a negative left offset, to ensure that the current match has some (visible) context available once scrolled into view.
In order to address this long-standing bug, we'll determine the (left) offset of the `selected` match and use that to modify the final position scrolled to in `PDFFindController.scrollMatchIntoView` such that the match is visible regardless of zoom level.
---
[1] Unfortunately we cannot directly scroll to the `selected` match, since it's not absolutely positioned and changing that would cause other bugs/regressions (note recent patches in that area).
[2] I did actually stumble upon this problem a little while ago, while working on PR 13482, but forgot to look into this again until I saw the new issue.
With the introduction of `PDFScriptingManager._closeCapability` in PR 13074, the pre-existing `PDFScriptingManager._pageEventsReady` boolean essentially became redundant.
Given that you always want to avoid tracking closely related state *separately*, since it's easy to introduce subtle bugs that way, we should just remove `PDFScriptingManager._pageEventsReady` now.
Obviously I *should* have done this already back in PR 13074, sorry about the churn here!
- a checkbox or radio doesn't have to be rescaled when the container is large so give the extra space to the caption to avoid some word wrapping.
- when the caption is on the right, then put ui on the left as first element and so remove flex:row-reverse stuff.
- some elements weren't displayed because their rotation angle was not taken into account;
- fix box model (XFA concept):
- remove use of outline;
- position correctly border which isn't part of box dimensions;
- fix margins issues (see issue #13474).
- move border on button instead of having it on wrapping div;
This code was added in PR 3968, apparently in order to fix scrolling of search results in HiDPI-mode.
However, after PR 4570 nothing is setting these `dataset`-properties any more and this is thus dead code which should be removed. (If that change had broken scrolling of search results in HiDPI-mode, you'd really expect that it'd been reported and fixed a long time ago.)
I missed this during review, since some of the changes in `web/pdf_print_service.js` broke printing.
Also, as part of these changes the patch replaces what looks like unnecessary `setAttribute` usage with "regular" `className` assignment and finally updates a couple of the CSS-rules to be more consistent.
- I thought it was possible to rely on browser layout engine to handle layout stuff but it isn't possible
- mainly because when a contentArea overflows, we must continue to layout in the next contentArea
- when no more contentArea is available then we must go to the next page...
- we must handle breakBefore and breakAfter which allows to "break" the layout to go to the next container
- Sometimes some containers don't provide their dimensions so we must compute them in order to know where to put
them in their parents but to compute those dimensions we need to layout the container itself...
- See top of file layout.js for more explanations about layout.
- fix few bugs in other places I met during my work on layout.
This patch fixes the referenced bugs/issues, in a way that won't interfere with keyboard users, assuming that we actually want to fix these old bugs/issues. (If not, we should close them as WONTFIX.)
After PR 13117 it's now (finally) possible for *different* build targets to specify individual options/preferences, and we can utilize that to only expose the `renderer`-preference in builds where `SVGGraphics` is actually defined.
Note that for e.g. `MOZCENTRAL`-builds, trying to enable SVG-rendering will throw immediately and the preference thus doesn't make sense to include there.
Also, update the dummy `SVGGraphics` to use a class, tweak the `PDFJSDev`-check in `src/display/svg.js` to agree fully with the option/preference, and remove an unnecessary `eslint-disable`.
Reasons for the removal include:
- This functionality was always somewhat experimental and has never been enabled by default, partly because of worries about rendering bugs caused by e.g. bad/outdated graphics drivers.
- After the initial implementation, in PR 4286 (back in 2014), no additional functionality has been added to the WebGL implementation.
- The vast majority of all documents do not benefit from WebGL rendering, since only a couple of *specific* features are supported (e.g. some Soft Masks and Patterns).
- There is, and has always been, *zero* test-coverage for the WebGL implementation.
- Overall performance, in the PDF.js library, has improved since the experimental WebGL implementation was added.
Rather than shipping unused *and* untested code, it seems reasonable to simply remove the WebGL implementation for now; thanks to version control it's always possible to bring back the code should the need ever arise.
This functionality was originally implemented in PR 7029; however it's not, nor has it ever been, used as far as I can tell.[1]
Note in particular that the default viewer does not expose either a preference or even an option with which `disableCanvasToImageConversion` can be toggled, and source-code modification is thus required.
Furthermore, note also that we have multiple other instances of `canvas`-data accesses in both the `src/display/canvas.js` and `src/display/text_layer.js` files. If any of those are blocked, by e.g. browser settings, there will be outright rendering bugs and non-working thumbnails thus seem like a very small issue in the grand scheme of things; hence why I'm suggesting that we remove the unused `disableCanvasToImageConversion` functionality.
---
[1] For the Tor use-case mentioned in issue 7026, I *believe* that the solution was to white-list `canvas`-data accesses for its built-in PDF Viewer.
- app.alert and few other function can use an object as parameter ({cMsg: ...});
- support app.alert with a question and a yes/no answer;
- update field siblings when one is changed in an action;
- stop calculation if calculate is set to false in the middle of calculations;
- get a boolean for checkboxes when they've been set through annotationStorage instead of a string.
Also, removes a couple of unnecessary local variables in the `Stepper.breakIt` method.
Finally, this patch also disables the ESLint `no-var` rule, in preparation for the next patch, for a couple of data-structures that need to remain globally available.
Given that both the textLayer rendering *and* the structTree parsing is asynchronous, it's possible that we'll attempt to insert the structTree in a removed page. While there's thankfully no outright breakage caused by this, it will nonetheless lead to errors being printed in the console and we should obviously avoid this.
To reproduce this bug (without the patch), open http://localhost:8888/web/viewer.html?file=/test/pdfs/pdf.pdf#disableStream=true&disableAutoFetch=true and scroll *very quickly* through the document and notice the following error being (intermittently) printed in the console:
```
Uncaught (in promise) TypeError: can't access property "appendChild", this.canvas is undefined
```
Using `for...of` is a modern and generally much nicer pattern, since it gets rid of unnecessary callback-functions. (In a couple of spots, a "regular" `for` loop had to be used.)
According to a decision by UX and PM, please see https://bugzilla.mozilla.org/show_bug.cgi?id=1705060#c2 (and implemented in https://bugzilla.mozilla.org/show_bug.cgi?id=1705327), we no longer show the notification-bar in Firefox; hence the special `PDFViewerApplication._delayedFallback` functionality should no longer be necessary.
Furthermore, note that at this point in time *most* of the features which used the `PDFViewerApplication._delayedFallback` functionality is now enabled by default; hence that provides even less reason to keep this code around and existing calls are thus converted to "regular" `PDFViewerApplication.fallback` calls.
According to a decision by UX and PM, please see https://bugzilla.mozilla.org/show_bug.cgi?id=1705060#c2, in Firefox we should first of all *not* display the notification-bar for signatures. Secondly, as can also be seen there, we shouldn't display the notification-bar *at all* and it's thus disabled in https://bugzilla.mozilla.org/show_bug.cgi?id=1705327.
If we purposely don't display a notification, for documents with signatures, in the *built in* Firefox PDF Viewer then it cannot be necessary in the GENERIC viewer either.
To simplify the overall implementation, given that it only applies to the GENERIC-viewer, this patch purposely re-uses the existing `errorWrapper`-functionality to display the message.
While that one is mostly intended for actual *errors*, by re-using it here we considerably reduce the amount of code/complexity necessary for supporting this new warning. It's obviously possible to re-factor/improve this later on, but the patch should do just fine here since it'll indeed inform users (of the GENERIC-viewer) about unverified signatures.
Finally this patch also tweaks the background-color of the `errorWrapper`, making it 20 percent lighter respectively darker (depending on the theme) to make it "stand out" a little bit *less*.[1] While it may perhaps be useful to re-style/re-factor the `errorWrapper`, this patch probably isn't the right place for doing that.
---
[1] Note how in the MOZCENTRAL-viewer, which instead uses the browser notification-bar, we're purposely using a neutral colour to not draw too much attention to the notification-bar.
This is first of all consistent with existing API-methods, where we return `null` when the data in question doesn't exist. Secondly, it should also be (slightly) more efficient since there's less dummy-data that we need to transfer between threads.
Finally, this prevents us from adding an empty/unnecessary span to *every* single page even in documents without any structure tree data.
I (unsurprisingly) managed to forget about handling the case where a "pagesloaded" event arrives *before* the outline has been parsed, in which case we'd not actually enable the `currentOutlineButton` as intended.
Also, in the "pagesloaded" event handler, we should ensure that there's actually any pages loaded since otherwise the "find current outlineItem"-feature doesn't make any sense.
- but don't validate them for now;
- Firefox will display a bar to warn that the signature validation is not supported (see https://bugzilla.mozilla.org/show_bug.cgi?id=854315)
- almost all (all ?) pdf readers display signatures;
- validation is done in edge but for now it's behind a pref.
It's obviously better and more correct to handle the "pagesloaded" case within `PDFOutlineViewer` *itself*, rather than essentially splitting the logic in two parts and forcing `PDFSidebar` to deal with what should've been handled internally in `PDFOutlineViewer`.
This is what I *should* have done in PR 12777, but for some reason didn't figure out how to implement it well enough back then; sorry about the churn here!
*This patch fixes some technical debt in the viewer.*
Given that most API methods are (purposely) asynchronous, there's always a risk that the viewer could have been `close`d before the requested data arrives.
Lately we've started to check this case before using the data, to prevent errors and/or inconsistent state, however the outline/attachments/layers fetching and rendering is old enough that it pre-dates those checks.
When a PDF is "marked" we now generate a separate DOM that represents
the structure tree from the PDF. This DOM is inserted into the <canvas>
element and allows screen readers to walk the tree and have more
information about headings, images, links, etc. To link the structure
tree DOM (which is empty) to the text layer aria-owns is used. This
required modifying the text layer creation so that marked items are
now tracked.
Scripting, as implemented, requires access to a complete document/viewer in order to work. Hence it doesn't really make sense to keep the `enableScripting`-option on `PDFPageView`-instances.[1]
---
[1] Note that there's the `PDFSinglePageViewer`, which can be used in cases where you want access to all features/functionality of the viewer but only display *one* page at a time.
Note how we purposely don't expose the `AnnotationStorage`-class directly in the official API (see `src/pdf.js`), since trying to use *multiple* ones simultaneously doesn't really make sense (e.g. in the viewer).
Instead we lazily initialize, and cache, just *one* instance via `PDFDocumentProxy.annotationStorage` which should thus be available internally in the API itself without having to be manually passed to various methods.
To support these changes, the `AnnotationStorage`-instance initialization is moved into the `WorkerTransport`-class to allow both `PDFDocumentProxy` and `PDFPageProxy` to access it.
This patch implements the following simplifications:
- Remove the `annotationStorage`-parameter from `PDFDocumentProxy.saveDocument`, since it's already available internally.
Furthermore, while it's currently possible to call that method without an `AnnotationStorage`-instance, that really does *not* make any sense at all. In this case you're effectively reducing `PDFDocumentProxy.saveDocument` to a "regular" `PDFDocumentProxy.getData` call, but with *a lot* more overhead, which was obviously not the intention of the `PDFDocumentProxy.saveDocument`-method.
- Try to discourage third-party users from calling `PDFDocumentProxy.saveDocument` unconditionally, as a replacement for `PDFDocumentProxy.getData` (note the previous point).
- Replace the `annotationStorage`-parameter, in `PDFPageProxy.render`, with a boolean `includeAnnotationStorage`-parameter which simply indicates if the (internally available) `AnnotationStorage`-instance should be used during rendering (e.g. for printing).
- By removing the need to *manually* provide `annotationStorage`-parameters to various API-methods, using the API should become simpler (e.g. for third-parties) since you no longer need to worry about manually fetching and passing around this data.
The reason for the fairly large discrepancy, in the thumbnail quality, between the `draw`/`setImage`-methods is that in the former case we *directly* render the thumbnails at the final size that they'll appear at in the sidebar. In the latter case, we instead downsize the (generally) much larger "regular" pages.
To address this, I'm thus proposing that we let `PDFThumbnailView.draw` render thumbnails at *twice* their intended size and then downsize them to the final size.
Obviously this will increase *peak* memory usage during thumbnail rendering in `PDFThumbnailView.draw`, since doubling the width/height of a `canvas` will lead to its pixel-count increasing by a factor of `4`. Furthermore, since you need four components per pixel (given that it's RGBA-data), this will thus lead to the *temporary* thumbnail `canvas`-sizes increasing by a factor of `16` during rendering. Hence why rendering thumbnails at their "original" scale, i.e. using something like `PDFPageProxy.getViewport({ scale: 1 });`, would be an absolutely terrible idea!
To reduce the size and scope of these changes, I've tried to re-factor and re-use as much of the existing downsizing-implementation already present in `PDFThumbnailView` as possible.
While this will generally *not* make thumbnails rendered by `PDFThumbnailView.draw` look *identical* to those based on the rendered pages (via `PDFThumbnailView.setImage`), it's a considerable improvement as far as I'm concerned and enough to call the issue fixed.
*Please note:* This patch will not lead to *any* additional overhead, in either memory usage or parsing, for thumbnails which are based on the rendered pages.
A loop is less efficient than just overwriting the content, which is what we've generally been using (for years) in other parts of the code-base (see e.g. `BaseViewer` and `PDFThumbnailViewer`).
These properties are always updated/used together, and there's no other methods which depend on just one of them, hence they're changed into local variables instead.
Looking through the history of this code, it seems they were converted *from* local variables and to properties all the way back in PR 2914; however as far as I can tell from that diff it doesn't seem to have been necessary even back then!?
As discussed in the issue, this is a small/simple patch that should help to prevent *outright* data loss in forms when a new document is opened in the GENERIC viewer.
While the implementation is perhaps a bit "simplistic", it does seem to work and should be fine given that this is an edge-case only relevant for the GENERIC viewer.
In the next patch we'll need to be able to actually wait for saving to complete, hence it's necessary to slightly re-factor the `save`-method.
As part of these changes, we can reduce some duplication in the `save`-method and slightly improve the overall code. For consistency, the `download`-method is updated similarily to improve the code (this functionality is *very* old, even pre-dating the introduction of Promises in the code-base).
As mentioned in the JSDoc comment, this should not be used unless you know what you're doing, since it will lead to increased memory usage. However, in some situations (e.g. SVG-rendering), we still want to be able to run general clean-up on both the main/worker-thread while keeping loaded fonts attached to the DOM.[1]
As part of these changes, `WorkerTransport.startCleanup` is converted to an async method and we'll also skip clean-up when destruction has started (since it's redundant).
---
[1] The SVG-rendering mode is obviously not officially supported, since it's both rather incomplete and inherently slower. However with recent changes, whereby we cache repeated images on the document rather than the page level, memory usage can be *a lot* worse than before if we never attempt to release e.g. cached image-data when the viewer is in SVG-rendering mode.
* JS - Handle correctly hierarchy of fields
- it aims to fix#13132;
- annotations can inherit their actions from the parent field;
- there are some fields which act as a container for other fields:
- they can be access through js so need to add them with an empty type (nothing in the spec about that but checked in Acrobat);
- calculation order list (CO) can reference them so need make them through this.getField;
- getArray method must return kids.
- field values are number, string, ... depending of their type but nothing in the spec on how to know what's the type:
- according to the comment for Canonical Format: https://www.adobe.com/content/dam/acom/en/devnet/pdf/pdfs/PDF32000_2008.pdf#page=461
- it seems that this "type" can be guessed from js action Format (when setting a type in Acrobat DC, the only affected thing is this action).
- util.scand with an empty string returns the current date.
Given that *all* data has been loaded on the main-thread, and then transferred to the worker-thread, ever since PR 8617 (almost four years ago) it should no longer be necessary to keep this special-case around.
Given that the `webViewerOpenFileViaURL` helper function is being defined in *all* builds anyway, the current pre-processor usage doesn't really improve readability in my opinion.
Currently `destRef === null`, which will only happen in documents with corrupt destinations, will (unsurprisingly) throw when trying to lookup the pageNumber. To avoid this, we can simply use the same format as in 1a2cdaffc5/web/pdf_link_service.js (L128)
The rotation handling that's currently living in `PDFViewerApplication` is *very* old, and pre-dates the introduction of the viewer components by years.
As can be seen in the `BaseViewer.pagesRotation` setter, we're not actually normalizing the rotation as intended and instead rely on the caller to handle that correctly. This is first of all inconsistent, given how other setters are implemented, and secondly it could also lead to the rotation being set to a value outside of the `[0, 360)`-range.
Finally, for improved consistency the rotation handling in `PageViewport` is updated similarly. Please note that this case, it's *not* changing the pre-existing logic.
This improves and simplifies #13102 in order to make printing of test-cases
like the one in bug 1698414 (where the real page is bigger than the target
page) much better, see incoming screenshots.
The reason why we need to stop setting .style.width / .style.height is to get
the right auto-sizing behavior in both axes. This shouldn't change behavior as
long as the print resolution is >= the CSS resolution, which seems like a
reasonable assumption.
If you try to print with a lower resolution than CSS, then instead of an
stretched canvas, you'd get a centered CSS-quality canvas, which seems
sensible. This could maybe be fixed with some CSS hackery (some combination of
min / max and viewport units perhaps?), but I think it's more trouble than it's
worth.
- implement few positioning properties: position, width, height, anchor;
- implement font element;
- implement fill element (used by font) and its children (linear, radial, ...);
- font property is inherited from ancestor container (see https://www.pdfa.org/wp-content/uploads/2020/07/XFA-3_3.pdf#page=43) so let CSS handles that stuff;
- in order to reduce the number of properties to set, only set non default properties and put the default in CSS;
- set a background to some containers to be able to see them (will be removed in a future commit).
The intention, in PR 12493, was that the page we're adding to the browser history should behave as if it were a "regular" internal destination (to properly convey user intent).
Unfortunately, since I didn't consider all the edge-cases correctly, it ended up behaving like a URL-hash instead which obviously wasn't intended. Note that currently this isn't a problem, however it can become an issue (in some cases) with upcoming re-factoring around `PDFHistory` and OpenAction support[1].
---
[1] I've started working on fixing the following TODO, which will require a couple of smaller tweaks here and there: 9d0ce6e79f/web/app.js (L1680-L1681)
In the `getAll`-method, we can have just one *explicit* loop rather than two indirect ones via the old `Object.assign`-call.
Also, changes the `get`-method to be slightly more compact (while keeping the logic intact).
Looking at this now, I cannot understand why we'd need to initialize `this.prefs` with all of the values from `this.defaults`.
Not only does this *indirectly* require one extra loop, via the `Object.assign`-call, but it also means that in GENERIC-builds changes to default-preference values might not be picked-up unless the the existing user-prefs are cleared (if the user had *manually* set prefs previously).
Given that the `enableXfa` parameter must to be passed to the API/Worker, and thus included in the `getDocument` call, it's not necessary to include it when initializing the `PDFViewer`-instance used in the default viewer. (Also, in `AppOptions`, the parameter is clearly marked with `OptionKind.API`.)
Furthermore, we probably don't want to display the fallback bar (in Firefox) for XFA documents when `enableXfa = true` is set.
While it's still not entirely clear if this would've prevented the issue as reported, given that the particular use-case reported apparently no longer applies, this small change really cannot hurt in general *and* it won't effect "regular" viewer builds in any way.
Given how the compatibility-values are being handled, it's not actually possible to override a *truthy* default-value with a *falsy* compatibility-value.
This is a simple oversight on my part, and with modern ECMAScript features this is very easy to support.
With the changes made in the previous patch, we can now list "disableTelemetry" in the `AppOptions` only for the `CHROME`-builds and thus remove the special-casing in the `checkChromePreferencesFile` helper function.
Originally the default preferences where simply placed in a JSON-file, checked into the repository, which over time became impractical, annoying, and error-prone to maintain; please see PR 10548.
While that improved the overall situation a fair bit, it however inherited one quite unfortunate property of the old JSON-based solution[1]: It's still not possible for *different* build targets to specify their *own* default preference values.
With some preferences, such as e.g. `enableScripting`, it's not inconceivable that you'd want to (at least) support build-specific default preference values. Currently that's not really possible, which is why this PR re-factors the default preferences generation to support this.
---
[1] This fact isn't really clear from the `AppOptions` implementation, unless you're familiar with the `gulpfile.js` code, which could lead to some confusion for those new to this part of the code-base.
A number of the currently supported *scripting* events only make sense in the "normal" viewer mode, and not when PresentationMode is active. For example:
- Changing the zoom-level will outright break rendering in PresentationMode, since it relies on "page-fit" being used.
- Focusing a particular (AcroForm) element won't work, and could break keyboard navigation, since forms should not be editable in PresentationMode (see issue 12232).
While this will perhaps not be perfect for *every* PDF document with mixed page orientation, based on the large number of bugs/issues seen over the years I'm however pretty convinced that it'll be an overall improvement in a majority of cases.
In order to improve things further, we'd probably need Firefox to support e.g. `@page` such that the viewer can provide better information to the print engine.
Currently, with `enablePrintAutoRotate = true` set, we're forced to loop through all the pages *twice* when checking for any landscape pages.
This seems completely unnecessary now, and using only *one* loop should be marginally more efficient in general.
Currently landscape pages are rotated *clockwise*, which for most documents feel wrong since holding the printed pages at their *left* edge causes the landscape pages to be viewed "upside down".
In general, since most documents are LTR ones, it feels more appropriate to instead rotate landscape pages *counterclockwise* for printing.
- add an option to enable XFA rendering if any;
- for now, let the canvas layer: it could be useful to implement XFAF forms (embedded pdf in xml stream for the background and xfa form for the foreground);
- ui elements in template DOM are pretty close to their html counterpart so we generate a fake html DOM from template one:
- it makes easier to translate template properties to html ones;
- it makes faster the creation of the html element in the main thread.
Given that https://bugzilla.mozilla.org/show_bug.cgi?id=1699219 has enabled scripting for all Firefox-channels, it seems reasonable to simply set `enableScripting = true` unconditionally in the viewer preferences/options.
For now, this patch leaves the standalone viewer-components alone (such as e.g. `BaseViewer`), and if those are used scripting will thus have to be manually enabled (see e.g. the "simpleviewer"/"singlepageviewer" examples).
It seems reasonable to place this alongside the *similar* `getFilenameFromUrl` helper function. This way, with the changes in the next patch, we also avoid having to expose the `isDataScheme` function in the API itself and we instead expose `getPdfFilenameFromUrl` in the API (which feels overall more appropriate).
