Currently image resources, as opposed to e.g. font resources, are handled exclusively on a page-specific basis. Generally speaking this makes sense, since pages are separate from each other, however there's PDF documents where many (or even all) pages actually references exactly the same image resources (through the XRef table). Hence, in some cases, we're decoding the *same* images over and over for every page which is obviously slow and wasting both CPU and memory resources better used elsewhere.[1]
Obviously we cannot simply treat all image resources as-if they're used throughout the entire PDF document, since that would end up increasing memory usage too much.[2]
However, by introducing a `GlobalImageCache` in the worker we can track image resources that appear on more than one page. Hence we can switch image resources from being page-specific to being document-specific, once the image resource has been seen on more than a certain number of pages.
In many cases, such as e.g. the referenced issue, this patch will thus lead to reduced memory usage for image resources. Scrolling through all pages of the document, there's now only a few main-thread copies of the same image data, as opposed to one for each rendered page (i.e. there could theoretically be *twenty* copies of the image data).
While this obviously benefit both CPU and memory usage in this case, for *very* large image data this patch *may* possibly increase persistent main-thread memory usage a tiny bit. Thus to avoid negatively affecting memory usage too much in general, particularly on the main-thread, the `GlobalImageCache` will *only* cache a certain number of image resources at the document level and simply fallback to the default behaviour.
Unfortunately the asynchronous nature of the code, with ranged/streamed loading of data, actually makes all of this much more complicated than if all data could be assumed to be immediately available.[3]
*Please note:* The patch will lead to *small* movement in some existing test-cases, since we're now using the built-in PDF.js JPEG decoder more. This was done in order to simplify the overall implementation, especially on the main-thread, by limiting it to only the `OPS.paintImageXObject` operator.
---
[1] There's e.g. PDF documents that use the same image as background on all pages.
[2] Given that data stored in the `commonObjs`, on the main-thread, are only cleared manually through `PDFDocumentProxy.cleanup`. This as opposed to data stored in the `objs` of each page, which is automatically removed when the page is cleaned-up e.g. by being evicted from the cache in the default viewer.
[3] If the latter case were true, we could simply check for repeat images *before* parsing started and thus avoid handling *any* duplicate image resources.
With these changes SystemJS is now only used, during development, on the worker-thread and in the unit/font-tests, since Firefox is currently missing support for worker modules; please see https://bugzilla.mozilla.org/show_bug.cgi?id=1247687
Hence all the JavaScript files in the `web/` and `src/display/` folders are now loaded *natively* by the browser (during development) using standard `import` statements/calls, thanks to a nice `import-maps` polyfill.
*Please note:* As soon as https://bugzilla.mozilla.org/show_bug.cgi?id=1247687 is fixed in Firefox, we should be able to remove all traces of SystemJS and thus finally be able to use every possible modern JavaScript feature.
As part of reducing our reliance on SystemJS in the development viewer, this patch replaces usage of `require` statements with modern standards `import`/`export` statements instead.
If we want to try and move forward with reducing usage of SystemJS, we don't have much choice but to make these kind changes (despite what prior test-results showed, however I'm no longer able to reproduce the issues locally).
As part of trying to reduce the usage of SystemJS in the development viewer, this patch is a necessary step that will allow removal of some `require` statements.
Currently this uses `SystemJS.import` in non-PRODUCTION mode, but it should be possible to replace those with standard *dynamic* `import` calls in the future.
Given that the `getPDFFileNameFromURL` helper function has a specific code-path for handling non-string inputs, this empty string fallback really isn't necessary at the call-site in `web/pdf_document_properties.js`.
Apparently the old link format used in MOZCENTRAL-builds, with the blob URL separated from the filename with a `?` character violates the specification; see https://bugzilla.mozilla.org/show_bug.cgi?id=1632644#c5
Obviously just removing the `?`-part of the URL would have worked, but that would also have meant that we'd no longer be able to provide the correct filename when the user attempts to download the opened PDF attachment.
To fix this we'll instead append the filename in the hash-part of the URL, which however required using a *custom* hash-parameter to avoid triggering the fallback "named destination" code-paths in the viewer.
Note that only changing the `web/pdf_attachment_viewer.js` file wasn't sufficient to fix the bug, and we also need to tweak the `webViewerInitialized` function in `web/app.js` since MOZCENTRAL-builds used to ignore *everything* in the URL hash.
This particular code is very old, but changing it *should* be completely safe given that the `PDFViewerApplication.setTitleUsingUrl` method since some time now stores both the original URL (in `this.url`) as well as one without the hash (in `this.baseUrl`). The latter one is already used everywhere where it matters, so this change seem fine to me.
This patch thus restores the original behaviour for PDF attachments in the MOZCENTRAL-build, by once again allowing them to be opened *directly* in the browser without downloading. (The fallback added in PR 11845 is obviously kept, since it seems generally useful to have.)
Originally the `default_preferences.json` file was checked into the repository, and we thus needed to load it in non-PRODUCTION mode (which was originally done asynchronously using `XMLHttpRequest`). Over the years a lot has changed and the `default_preferences.json` file is now built, by the `gulp default_preferences` task, from the `web/app_options.js` file. Hence it's no longer necessary, in non-PRODUCTION mode, to use SystemJS here since we can simply use a standard `import` statement instead.
Note how e.g. `web/app.js` already imports from `web/app_options.js` in the same exact way that `web/preferences.js` now does, hence this patch will *not* result in any significant changes in the built/bundled viewer file.
This is another (small) part in trying to reduce usage of SystemJS, with the goal of hopefully getting rid of it completely. (I've started working on this, and doing so has identified a number of problem areas; this patch addresses one of them.)
This replaces some additional `require`/`exports` usage with standard `import`/`export` statements instead.
Hence another, small, part in the effort to reduce the reliance on SystemJS-specific functionality in the development viewer.
By re-factoring the existing gulp tasks, most of the code can be re-used for both the existing `gulp minified` as well as the new `gulp minified-es5` task.
While working on PR 11872, it occurred to me that it probably wouldn't be a bad idea to change the `_parsedAnnotations` getter to handle errors individually for each annotation. This way, one broken/corrupt annotation won't prevent the rest of them from being e.g. fetched through the API.
Given that the `PDFLinkService.setHash` method itself if completely synchronous, moving the handling of "nameddest" to occur last *shouldn't* cause any problems (famous last words).
This way the destination will still override any previous parameter, such as e.g. the "page", as expected. Furthermore, given that the `PDFLinkService.navigateTo` method is asynchronous that should provide additional guarantees that the "nameddest" parameter is always respected.
As sort-of expected, this fairly innocent looking change also required some tweaks in the `PDFHistory` to prevent dummy history entires upon document load (only an issue when both "page" *and* "nameddest" parameters are provided in the hash).
This rule complements the existing `accessor-pairs` nicely, and ensures that a getter/setter pair is always consistently ordered.
Please find additional details about this rule at https://eslint.org/docs/rules/grouped-accessor-pairs
Having `assert` calls without a message string isn't very helpful when debugging, and it turns out that it's easy enough to make use of ESLint to enforce better `assert` call-sites.
In a couple of cases the `assert` calls were changed to "regular" throwing of errors instead, since that seemed more appropriate.
Please find additional details about the ESLint rule at https://eslint.org/docs/rules/no-restricted-syntax
This should ensure that a page will always render successfully, even if there's errors during the Annotation fetching/parsing.
Additionally the `OperatorList.addOpList` method is also adjusted to ignore invalid data, to make it slightly more robust.
All of the other *similar* helper functions only target one file per function, and there's no particular reason for this one to be different.
This patch will simplify future changes, e.g. experimenting with using `gulp watch` instead of SystemJS for the development viewer.
Not only is there no code depending on it now, the actual task itself doesn't even make sense as-is. Note that it uses the default `DEFINES` configuration *unaltered*, which is neither useful nor correct since the resulting build thus won't make sense without an actual built target set.
Given that none of the PDF.js contributors know React, maintaining and/or providing supporting for the example isn't really feasible unfortunately.
Even something as simple as running/testing the example becomes difficult for anyone completely unfamiliar with React, and furthermore:
- It's very difficult to tell if the example demonstrates React best-practices, since the PDF.js contributors don't know React.
- We also have no reasonable way of keeping the example up-to-date with changes in React.
- The React example, in its current form, is even *hard-coding* the PDF.js version to a now unsupported version.
- The example is currently triggering "fake worker" usage, see issue 11729, which is really *really* bad. Note that the "fake worker" functionality is *only* intended as a fallback, and it should absolutely *not* under any circumstances be advertised and certainly shouldn't be triggered in official PDF.js examples.
By preserving the exception type, more fine-grained error handling can be performed via client-side logic (e.g. redirect to a search page if a PDF is not found, or to a ticket system in case of invalid PDF files).
The original exception is now re-thrown.
Fixes#11658
We already rendered the name for radio buttons, but it was missing for
all other interactive form elements. This commit adds that so that
values entered in form elements can be read based on the element name.
To save time or resources during development it can be useful to run
tests only in Firefox. Previously this could be done by editing the
browser manifest file, but since that file is no longer used for
Puppeteer, this command line flag replaces it. For example, executing
`gulp unittest --noChrome` will only run the unit tests in Firefox.
