pdf.js.mirror/test/integration/text_layer_spec.mjs
Nicolò Ribaudo 886c90d1a5
Add support for right-clicking on images
This patch adds right-click support for images in the PDF, allowing
users to download them. To minimize memory consumption, we:
- Do not store the images separately, and instead crop them out of the
  PDF page canvas
- Only extract the images when needed (i.e. when the user right-clicks
  on them), rather than eagery having all of them available.

To do so, we layer one empty 0x0 canvas per image, stretched to cover
the whole image, and only populate its contents on right click.
These images need to be inside the text layer: they cannot be _behind_
it, otherwise they would be covered by the text layer's container and
not be clickable, and they cannot be in front of it, otherwise they
would make the text spans unselectable.

This feature is managed by a new preference, `imagesRightClickMinSize`:
- when it's set to `-1`, right-click support is disabled
- when set to `0`, all images are available for right click
- when set to a positive integer, only images whose width and height are
  greater than or equal to that value (in the PDF page frame of
  reference) are available for right click.

This features is disabled by default outside of MOZCENTRAL, as it
significantly degrades the text selection experience in non-Firefox
browsers.
2026-03-10 14:51:03 +01:00

681 lines
24 KiB
JavaScript

/* Copyright 2024 Mozilla Foundation
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
import {
closePages,
closeSinglePage,
getSpanRectFromText,
loadAndWait,
waitForEvent,
} from "./test_utils.mjs";
import { startBrowser } from "../test.mjs";
describe("Text layer", () => {
describe("Text selection", () => {
// page.mouse.move(x, y, { steps: ... }) doesn't work in Firefox, because
// puppeteer will send fractional intermediate positions and Firefox doesn't
// support them. Use this function to round each intermediate position to an
// integer.
async function moveInSteps(page, from, to, steps) {
const deltaX = to.x - from.x;
const deltaY = to.y - from.y;
for (let i = 0; i <= steps; i++) {
const x = Math.round(from.x + (deltaX * i) / steps);
const y = Math.round(from.y + (deltaY * i) / steps);
await page.mouse.move(x, y);
}
}
function middlePosition(rect) {
return {
x: Math.round(rect.x + rect.width / 2),
y: Math.round(rect.y + rect.height / 2),
};
}
function middleLeftPosition(rect) {
return {
x: Math.round(rect.x + 1),
y: Math.round(rect.y + rect.height / 2),
};
}
function belowEndPosition(rect) {
return {
x: Math.round(rect.x + rect.width),
y: Math.round(rect.y + rect.height * 1.5),
};
}
beforeEach(() => {
jasmine.addAsyncMatchers({
// Check that a page has a selection containing the given text, with
// some tolerance for extra characters before/after.
toHaveRoughlySelected({ pp }) {
return {
async compare(page, expected) {
const TOLERANCE = 10;
const actual = await page.evaluate(() =>
// We need to normalize EOL for Windows
window.getSelection().toString().replaceAll("\r\n", "\n")
);
let start, end;
if (expected instanceof RegExp) {
const match = expected.exec(actual);
start = -1;
if (match) {
start = match.index;
end = start + match[0].length;
}
} else {
start = actual.indexOf(expected);
if (start !== -1) {
end = start + expected.length;
}
}
const pass =
start !== -1 &&
start < TOLERANCE &&
end > actual.length - TOLERANCE;
return {
pass,
message: `Expected ${pp(
actual.length > 200
? actual.slice(0, 100) + "[...]" + actual.slice(-100)
: actual
)} to ${pass ? "not " : ""}roughly match ${pp(expected)}.`,
};
},
};
},
});
});
describe("using mouse", () => {
describe("doesn't jump when hovering on an empty area", () => {
let pages;
beforeEach(async () => {
pages = await loadAndWait(
"tracemonkey.pdf",
`.page[data-page-number = "1"] .endOfContent`,
undefined,
undefined,
(_page, browserName) => ({
// Enable images in Firefox, to ensure that they do not interfere
// with text selection. We do not test it in Chrome because we
// know that they do degrate the text selection experience there.
imagesRightClickMinSize: browserName === "firefox" ? 16 : -1,
})
);
});
afterEach(async () => {
await closePages(pages);
});
it("in a single page", async () => {
await Promise.all(
pages.map(async ([browserName, page]) => {
const [positionStart, positionEnd] = await Promise.all([
getSpanRectFromText(
page,
1,
"(frequently executed) bytecode sequences, records"
).then(middlePosition),
getSpanRectFromText(
page,
1,
"them, and compiles them to fast native code. We call such a se-"
).then(belowEndPosition),
]);
await page.mouse.move(positionStart.x, positionStart.y);
await page.mouse.down();
await moveInSteps(page, positionStart, positionEnd, 20);
await page.mouse.up();
await expectAsync(page)
.withContext(`In ${browserName}`)
.toHaveRoughlySelected(
"code sequences, records\n" +
"them, and compiles them to fast native code. We call suc"
);
})
);
});
it("across multiple pages", async () => {
await Promise.all(
pages.map(async ([browserName, page]) => {
const scrollTarget = await getSpanRectFromText(
page,
1,
"Unlike method-based dynamic compilers, our dynamic com-"
);
await page.evaluate(top => {
document.getElementById("viewerContainer").scrollTop = top;
}, scrollTarget.y - 50);
const [
positionStartPage1,
positionEndPage1,
positionStartPage2,
positionEndPage2,
] = await Promise.all([
getSpanRectFromText(
page,
1,
"Each compiled trace covers one path through the program with"
).then(middlePosition),
getSpanRectFromText(
page,
1,
"or that the same types will occur in subsequent loop iterations."
).then(middlePosition),
getSpanRectFromText(
page,
2,
"Hence, recording and compiling a trace"
).then(middlePosition),
getSpanRectFromText(
page,
2,
"cache. Alternatively, the VM could simply stop tracing, and give up"
).then(belowEndPosition),
]);
await page.mouse.move(positionStartPage1.x, positionStartPage1.y);
await page.mouse.down();
await moveInSteps(page, positionStartPage1, positionEndPage1, 20);
await moveInSteps(page, positionEndPage1, positionStartPage2, 20);
await expectAsync(page)
.withContext(`In ${browserName}, first selection`)
.toHaveRoughlySelected(
/path through the program .*Hence, recording a/s
);
await moveInSteps(page, positionStartPage2, positionEndPage2, 20);
await page.mouse.up();
await expectAsync(page)
.withContext(`In ${browserName}, second selection`)
.toHaveRoughlySelected(
/path through.*Hence, recording and .* tracing, and give/s
);
})
);
});
});
describe("doesn't jump when hovering on an empty area, with .markedContent", () => {
let pages;
beforeEach(async () => {
pages = await loadAndWait(
"chrome-text-selection-markedContent.pdf",
`.page[data-page-number = "1"] .endOfContent`,
undefined,
undefined,
(_page, browserName) => ({
// Enable images in Firefox, to ensure that they do not interfere
// with text selection. We do not test it in Chrome because we
// know that they do degrate the text selection experience there.
imagesRightClickMinSize: browserName === "firefox" ? 16 : -1,
})
);
});
afterEach(async () => {
await closePages(pages);
});
it("in per-character selection mode", async () => {
await Promise.all(
pages.map(async ([browserName, page]) => {
const [positionStart, positionEnd] = await Promise.all([
getSpanRectFromText(
page,
1,
"strengthen in the coming quarters as the railway projects under"
).then(middlePosition),
getSpanRectFromText(
page,
1,
"development enter the construction phase (estimated at around"
).then(belowEndPosition),
]);
await page.mouse.move(positionStart.x, positionStart.y);
await page.mouse.down();
await moveInSteps(page, positionStart, positionEnd, 20);
await page.mouse.up();
await expectAsync(page)
.withContext(`In ${browserName}`)
.toHaveRoughlySelected(
"rs as the railway projects under\n" +
"development enter the construction phase (estimated at "
);
})
);
});
it("in per-word selection mode", async () => {
await Promise.all(
pages.map(async ([browserName, page]) => {
const [positionStart, positionEnd] = await Promise.all([
getSpanRectFromText(
page,
1,
"strengthen in the coming quarters as the railway projects under"
).then(middlePosition),
getSpanRectFromText(
page,
1,
"development enter the construction phase (estimated at around"
).then(belowEndPosition),
]);
// Puppeteer doesn't support emulating "double click and hold" for
// WebDriver BiDi, so we must manually dispatch a protocol action
// (see https://github.com/puppeteer/puppeteer/issues/13745).
await page.mainFrame().browsingContext.performActions([
{
type: "pointer",
id: "__puppeteer_mouse",
actions: [
{ type: "pointerMove", ...positionStart },
{ type: "pointerDown", button: 0 },
{ type: "pointerUp", button: 0 },
{ type: "pointerDown", button: 0 },
],
},
]);
await moveInSteps(page, positionStart, positionEnd, 20);
await page.mouse.up();
await expectAsync(page)
.withContext(`In ${browserName}`)
.toHaveRoughlySelected(
"quarters as the railway projects under\n" +
"development enter the construction phase (estimated at around"
);
})
);
});
});
describe("when selecting over a link", () => {
let pages;
beforeEach(async () => {
pages = await loadAndWait(
"annotation-link-text-popup.pdf",
`.page[data-page-number = "1"] .endOfContent`,
undefined,
undefined,
(_page, browserName) => ({
// Enable images in Firefox, to ensure that they do not interfere
// with text selection. We do not test it in Chrome because we
// know that they do degrate the text selection experience there.
imagesRightClickMinSize: browserName === "firefox" ? 16 : -1,
})
);
});
afterEach(async () => {
await closePages(pages);
});
it("allows selecting within the link", async () => {
await Promise.all(
pages.map(async ([browserName, page]) => {
const [positionStart, positionEnd] = await Promise.all([
getSpanRectFromText(page, 1, "Link").then(middleLeftPosition),
getSpanRectFromText(page, 1, "mozilla.org").then(
middlePosition
),
]);
await page.mouse.move(positionStart.x, positionStart.y);
await page.mouse.down();
await moveInSteps(page, positionStart, positionEnd, 20);
await page.mouse.up();
await expectAsync(page)
.withContext(`In ${browserName}`)
.toHaveRoughlySelected("Link\nmozil");
})
);
});
it("allows selecting within the link when going backwards", async () => {
await Promise.all(
pages.map(async ([browserName, page]) => {
const [positionStart, positionEnd] = await Promise.all([
getSpanRectFromText(page, 1, "Text").then(middlePosition),
getSpanRectFromText(page, 1, "mozilla.org").then(
middlePosition
),
]);
await page.mouse.move(positionStart.x, positionStart.y);
await page.mouse.down();
await moveInSteps(page, positionStart, positionEnd, 20);
await page.mouse.up();
await expectAsync(page)
.withContext(`In ${browserName}`)
.toHaveRoughlySelected("a.org\nTe");
})
);
});
it("allows clicking the link after selecting", async () => {
await Promise.all(
pages.map(async ([browserName, page]) => {
const [positionStart, positionEnd] = await Promise.all([
getSpanRectFromText(page, 1, "Link").then(middleLeftPosition),
getSpanRectFromText(page, 1, "mozilla.org").then(
middlePosition
),
]);
await page.mouse.move(positionStart.x, positionStart.y);
await page.mouse.down();
await moveInSteps(page, positionStart, positionEnd, 20);
await page.mouse.up();
await waitForEvent({
page,
eventName: "click",
action: () => page.mouse.click(positionEnd.x, positionEnd.y),
selector: "#pdfjs_internal_id_8R",
validator: e => {
// Don't navigate to the link destination: the `click` event
// firing is enough validation that the link can be clicked.
e.preventDefault();
return true;
},
});
})
);
});
it("allows clicking the link after changing selection with the keyboard", async () => {
await Promise.all(
pages.map(async ([browserName, page]) => {
const [positionStart, positionEnd] = await Promise.all([
getSpanRectFromText(page, 1, "Link").then(middleLeftPosition),
getSpanRectFromText(page, 1, "mozilla.org").then(
middlePosition
),
]);
await page.mouse.move(positionStart.x, positionStart.y);
await page.mouse.down();
await moveInSteps(page, positionStart, positionEnd, 20);
await page.mouse.up();
await page.keyboard.down("Shift");
await page.keyboard.press("ArrowRight");
await page.keyboard.up("Shift");
await waitForEvent({
page,
eventName: "click",
action: () => page.mouse.click(positionEnd.x, positionEnd.y),
selector: "#pdfjs_internal_id_8R",
validator: e => {
// Don't navigate to the link destination: the `click` event
// firing is enough validation that the link can be clicked.
e.preventDefault();
return true;
},
});
})
);
});
});
describe("when selecting text with find highlights active", () => {
let pages;
beforeEach(async () => {
pages = await loadAndWait(
"find_all.pdf",
".textLayer",
100,
undefined,
(_page, browserName) => ({
// Enable images in Firefox, to ensure that they do not interfere
// with text selection. We do not test it in Chrome because we
// know that they do degrate the text selection experience there.
imagesRightClickMinSize: browserName === "firefox" ? 16 : -1,
})
);
});
afterEach(async () => {
await closePages(pages);
});
it("doesn't jump when selection anchor is inside a highlight element", async () => {
await Promise.all(
pages.map(async ([browserName, page]) => {
// Highlight all occurrences of the letter A (case insensitive).
await page.click("#viewFindButton");
await page.waitForSelector("#findInput", { visible: true });
await page.type("#findInput", "a");
await page.click("#findHighlightAll + label");
await page.waitForSelector(".textLayer .highlight");
// find_all.pdf contains 'AB BA' in a monospace font. These are
// the glyph metrics at 100% zoom, extracted from the PDF.
const glyphWidth = 15.98;
const expectedFirstAX = 30;
// Compute the drag coordinates to select exactly "AB". The
// horizontal positions use the page origin and PDF glyph
// metrics; the vertical center comes from the highlight.
const pageDiv = await page.$(".page canvas");
const pageBox = await pageDiv.boundingBox();
const firstHighlight = await page.$(".textLayer .highlight");
const highlightBox = await firstHighlight.boundingBox();
// Drag from beginning of first 'A' to end of second 'B'
const aStart = pageBox.x + expectedFirstAX;
const startY = Math.round(
highlightBox.y + highlightBox.height / 2
);
const bEnd = Math.round(aStart + glyphWidth * 2);
await page.mouse.move(aStart, startY);
await page.mouse.down();
await moveInSteps(
page,
{ x: aStart, y: startY },
{ x: bEnd, y: startY },
20
);
await page.mouse.up();
const selection = await page.evaluate(() =>
window.getSelection().toString()
);
expect(selection).withContext(`In ${browserName}`).toEqual("AB");
// The selectionchange handler in TextLayerBuilder walks up
// from .highlight to its parent span before placing
// endOfContent (see text_layer_builder.js). Without that
// fix, endOfContent would be inserted inside the text span
// (as a sibling of the .highlight) instead of as a direct
// child of .textLayer. Verify the correct DOM structure.
const endOfContentIsDirectChild = await page.evaluate(() => {
const eoc = document.querySelector(".textLayer .endOfContent");
return eoc?.parentElement?.classList.contains("textLayer");
});
expect(endOfContentIsDirectChild)
.withContext(`In ${browserName}`)
.toBeTrue();
})
);
});
});
});
describe("using selection carets", () => {
let browser;
let page;
beforeEach(async () => {
// Chrome does not support simulating caret-based selection, so this
// test only runs in Firefox.
browser = await startBrowser({
browserName: "firefox",
startUrl: "",
extraPrefsFirefox: {
"layout.accessiblecaret.enabled": true,
"layout.accessiblecaret.hide_carets_for_mouse_input": false,
},
});
page = await browser.newPage();
await page.goto(
`${global.integrationBaseUrl}?file=/test/pdfs/tracemonkey.pdf#zoom=page-fit`
);
await page.bringToFront();
await page.waitForSelector(
`.page[data-page-number = "1"] .endOfContent`,
{ timeout: 0 }
);
});
afterEach(async () => {
await closeSinglePage(page);
await browser.close();
});
it("doesn't jump when moving selection", async () => {
const [initialStart, initialEnd, finalEnd] = await Promise.all([
getSpanRectFromText(
page,
1,
"(frequently executed) bytecode sequences, records"
).then(middleLeftPosition),
getSpanRectFromText(
page,
1,
"(frequently executed) bytecode sequences, records"
).then(middlePosition),
getSpanRectFromText(
page,
1,
"them, and compiles them to fast native code. We call such a se-"
).then(belowEndPosition),
]);
await page.mouse.move(initialStart.x, initialStart.y);
await page.mouse.down();
await moveInSteps(page, initialStart, initialEnd, 20);
await page.mouse.up();
await expectAsync(page)
.withContext(`first selection`)
.toHaveRoughlySelected("frequently executed) byt");
const initialCaretPos = {
x: initialEnd.x,
y: initialEnd.y + 10,
};
const intermediateCaretPos = {
x: finalEnd.x,
y: finalEnd.y + 5,
};
const finalCaretPos = {
x: finalEnd.x + 20,
y: finalEnd.y + 5,
};
await page.mouse.move(initialCaretPos.x, initialCaretPos.y);
await page.mouse.down();
await moveInSteps(page, initialCaretPos, intermediateCaretPos, 20);
await page.mouse.up();
await expectAsync(page)
.withContext(`second selection`)
.toHaveRoughlySelected(/frequently .* We call such a s/s);
await page.mouse.down();
await moveInSteps(page, intermediateCaretPos, finalCaretPos, 20);
await page.mouse.up();
await expectAsync(page)
.withContext(`third selection`)
.toHaveRoughlySelected(/frequently .* We call such a s/s);
});
});
});
describe("when the browser enforces a minimum font size", () => {
let browser;
let page;
beforeEach(async () => {
// Only testing in Firefox because, while Chrome has a setting similar to
// font.minimum-size.x-western, it is not exposed through its API.
browser = await startBrowser({
browserName: "firefox",
startUrl: "",
extraPrefsFirefox: { "font.minimum-size.x-western": 40 },
});
page = await browser.newPage();
await page.goto(
`${global.integrationBaseUrl}?file=/test/pdfs/tracemonkey.pdf#zoom=100`
);
await page.bringToFront();
await page.waitForSelector(
`.page[data-page-number = "1"] .endOfContent`,
{ timeout: 0 }
);
});
afterEach(async () => {
await closeSinglePage(page);
await browser.close();
});
it("renders spans with the right size", async () => {
const rect = await getSpanRectFromText(
page,
1,
"Dynamic languages such as JavaScript are more difficult to com-"
);
// The difference between `a` and `b`, as a percentage of the lower one
const getPercentDiff = (a, b) => Math.max(a, b) / Math.min(a, b) - 1;
expect(getPercentDiff(rect.width, 315)).toBeLessThan(0.03);
expect(getPercentDiff(rect.height, 12)).toBeLessThan(0.03);
});
});
});