XFA - Add the possibily to layout and measure text

- some containers doesn't always have their 2 dimensions and those dimensions re based on contents; - so in order to measure text, we must get the glyph widths (for the xfa fonts) before starting the layout; - implement a word-wrap algorithm; - handle font change during text layout.
mozilla · Jun 16, 2021 · 4e75b6f · 4e75b6f
1 parent f9a0568
commit 4e75b6f
Show file tree

Hide file tree

Showing 11 changed files with 397 additions and 90 deletions.
diff --git a/src/core/document.js b/src/core/document.js
@@ -857,6 +857,10 @@ class PDFDocument {
     return shadow(this, "xfaFaxtory", null);
   }
 
+  get isPureXfa() {
+    return this.xfaFactory && this.xfaFactory.isValid();
+  }
+
   get htmlForXfa() {
     if (this.xfaFactory) {
       return this.xfaFactory.getPages();
@@ -898,8 +902,14 @@ class PDFDocument {
       options,
     });
     const operatorList = new OperatorList();
+    const pdfFonts = [];
     const initialState = {
-      font: null,
+      get font() {
+        return pdfFonts[pdfFonts.length - 1];
+      },
+      set font(font) {
+        pdfFonts.push(font);
+      },
       clone() {
         return this;
       },
@@ -947,6 +957,7 @@ class PDFDocument {
       );
     }
     await Promise.all(promises);
+    this.xfaFactory.setFonts(pdfFonts);
   }
 
   get formInfo() {

diff --git a/src/core/fonts.js b/src/core/fonts.js
@@ -844,6 +844,7 @@ class Font {
     this.capHeight = properties.capHeight / PDF_GLYPH_SPACE_UNITS;
     this.ascent = properties.ascent / PDF_GLYPH_SPACE_UNITS;
     this.descent = properties.descent / PDF_GLYPH_SPACE_UNITS;
+    this.lineHeight = this.ascent - this.descent;
     this.fontMatrix = properties.fontMatrix;
     this.bbox = properties.bbox;
     this.defaultEncoding = properties.defaultEncoding;
@@ -2466,13 +2467,16 @@ class Font {
       unitsPerEm: int16(tables.head.data[18], tables.head.data[19]),
       yMax: int16(tables.head.data[42], tables.head.data[43]),
       yMin: signedInt16(tables.head.data[38], tables.head.data[39]),
-      ascent: int16(tables.hhea.data[4], tables.hhea.data[5]),
+      ascent: signedInt16(tables.hhea.data[4], tables.hhea.data[5]),
       descent: signedInt16(tables.hhea.data[6], tables.hhea.data[7]),
+      lineGap: signedInt16(tables.hhea.data[8], tables.hhea.data[9]),
     };
 
     // PDF FontDescriptor metrics lie -- using data from actual font.
     this.ascent = metricsOverride.ascent / metricsOverride.unitsPerEm;
     this.descent = metricsOverride.descent / metricsOverride.unitsPerEm;
+    this.lineGap = metricsOverride.lineGap / metricsOverride.unitsPerEm;
+    this.lineHeight = this.ascent - this.descent + this.lineGap;
 
     // The 'post' table has glyphs names.
     if (tables.post) {

diff --git a/src/core/worker.js b/src/core/worker.js
@@ -187,13 +187,8 @@ class WorkerMessageHandler {
         await pdfManager.ensureDoc("checkFirstPage");
       }
 
-      const [numPages, fingerprint, htmlForXfa] = await Promise.all([
-        pdfManager.ensureDoc("numPages"),
-        pdfManager.ensureDoc("fingerprint"),
-        pdfManager.ensureDoc("htmlForXfa"),
-      ]);
-
-      if (htmlForXfa) {
+      const isPureXfa = await pdfManager.ensureDoc("isPureXfa");
+      if (isPureXfa) {
         const task = new WorkerTask("loadXfaFonts");
         startWorkerTask(task);
         await pdfManager
@@ -203,6 +198,17 @@ class WorkerMessageHandler {
           })
           .then(() => finishWorkerTask(task));
       }
+
+      const [numPages, fingerprint] = await Promise.all([
+        pdfManager.ensureDoc("numPages"),
+        pdfManager.ensureDoc("fingerprint"),
+      ]);
+
+      // Get htmlForXfa after numPages to avoid to create HTML 2 times.
+      const htmlForXfa = isPureXfa
+        ? await pdfManager.ensureDoc("htmlForXfa")
+        : null;
+
       return { numPages, fingerprint, htmlForXfa };
     }
 

diff --git a/src/core/xfa/factory.js b/src/core/xfa/factory.js
@@ -13,37 +13,71 @@
  * limitations under the License.
  */
 
-import { $toHTML } from "./xfa_object.js";
+import { $fonts, $toHTML } from "./xfa_object.js";
 import { Binder } from "./bind.js";
+import { warn } from "../../shared/util.js";
 import { XFAParser } from "./parser.js";
 
 class XFAFactory {
   constructor(data) {
     try {
       this.root = new XFAParser().parse(XFAFactory._createDocument(data));
       this.form = new Binder(this.root).bind();
-      this._createPages();
     } catch (e) {
-      console.log(e);
+      warn(`XFA - an error occured during parsing and binding: ${e}`);
     }
   }
 
+  isValid() {
+    return this.root && this.form;
+  }
+
   _createPages() {
-    this.pages = this.form[$toHTML]();
-    this.dims = this.pages.children.map(c => {
-      const { width, height } = c.attributes.style;
-      return [0, 0, parseInt(width), parseInt(height)];
-    });
+    try {
+      this.pages = this.form[$toHTML]();
+      this.dims = this.pages.children.map(c => {
+        const { width, height } = c.attributes.style;
+        return [0, 0, parseInt(width), parseInt(height)];
+      });
+    } catch (e) {
+      warn(`XFA - an error occured during layout: ${e}`);
+    }
   }
 
   getBoundingBox(pageIndex) {
     return this.dims[pageIndex];
   }
 
   get numberPages() {
+    if (!this.pages) {
+      this._createPages();
+    }
     return this.dims.length;
   }
 
+  setFonts(fonts) {
+    this.form[$fonts] = Object.create(null);
+    for (const font of fonts) {
+      const cssFontInfo = font.cssFontInfo;
+      const name = cssFontInfo.fontFamily;
+      if (!this.form[$fonts][name]) {
+        this.form[$fonts][name] = Object.create(null);
+      }
+      let property = "regular";
+      if (cssFontInfo.italicAngle !== "0") {
+        if (parseFloat(cssFontInfo.fontWeight) >= 700) {
+          property = "bolditalic";
+        } else {
+          property = "italic";
+        }
+      } else if (parseFloat(cssFontInfo.fontWeight) >= 700) {
+        property = "bold";
+      }
+
+      this.form[$fonts][name][property] = font;
+    }
+  }
+
   getPages() {
     if (!this.pages) {
       this._createPages();

diff --git a/src/core/xfa/html_utils.js b/src/core/xfa/html_utils.js
@@ -18,18 +18,14 @@ import {
   $getParent,
   $getSubformParent,
   $nodeName,
+  $pushGlyphs,
   $toStyle,
   XFAObject,
 } from "./xfa_object.js";
 import { getMeasurement } from "./utils.js";
+import { TextMeasure } from "./text.js";
 import { warn } from "../../shared/util.js";
 
-const wordNonWordRegex = new RegExp(
-  "([\\p{N}\\p{L}\\p{M}]+)|([^\\p{N}\\p{L}\\p{M}]+)",
-  "gu"
-);
-const wordFirstRegex = new RegExp("^[\\p{N}\\p{L}\\p{M}]", "u");
-
 function measureToString(m) {
   if (typeof m === "string") {
     return "0px";
@@ -192,65 +188,15 @@ const converters = {
   },
 };
 
-function layoutText(text, fontSize, space) {
-  // Try to guess width and height for the given text in taking into
-  // account the space where the text should fit.
-  // The computed dimensions are just an overestimation.
-  // TODO: base this estimation on real metrics.
-  let width = 0;
-  let height = 0;
-  let totalWidth = 0;
-  const lineHeight = fontSize * 1.5;
-  const averageCharSize = fontSize * 0.4;
-  const maxCharOnLine = Math.floor(space.width / averageCharSize);
-  const chunks = text.match(wordNonWordRegex);
-  let treatedChars = 0;
-
-  let i = 0;
-  let chunk = chunks[0];
-  while (chunk) {
-    const w = chunk.length * averageCharSize;
-    if (width + w <= space.width) {
-      width += w;
-      treatedChars += chunk.length;
-      chunk = chunks[i++];
-      continue;
-    }
-
-    if (!wordFirstRegex.test(chunk) || chunk.length > maxCharOnLine) {
-      const numOfCharOnLine = Math.floor(
-        (space.width - width) / averageCharSize
-      );
-      chunk = chunk.slice(numOfCharOnLine);
-      treatedChars += numOfCharOnLine;
-      if (height + lineHeight > space.height) {
-        return { width: 0, height: 0, splitPos: treatedChars };
-      }
-      totalWidth = Math.max(width, totalWidth);
-      width = 0;
-      height += lineHeight;
-      continue;
-    }
-
-    if (height + lineHeight > space.height) {
-      return { width: 0, height: 0, splitPos: treatedChars };
-    }
-
-    totalWidth = Math.max(width, totalWidth);
-    width = w;
-    height += lineHeight;
-    chunk = chunks[i++];
-  }
-
-  if (totalWidth === 0) {
-    totalWidth = width;
-  }
-
-  if (totalWidth !== 0) {
-    height += lineHeight;
+function layoutText(text, xfaFont, fonts, width) {
+  const measure = new TextMeasure(xfaFont, fonts);
+  if (typeof text === "string") {
+    measure.addString(text);
+  } else {
+    text[$pushGlyphs](measure);
   }
 
-  return { width: totalWidth, height, splitPos: -1 };
+  return measure.compute(width);
 }
 
 function computeBbox(node, html, availableSpace) {

diff --git a/src/core/xfa/template.js b/src/core/xfa/template.js
@@ -22,6 +22,7 @@ import {
   $extra,
   $finalize,
   $flushHTML,
+  $fonts,
   $getAvailableSpace,
   $getChildren,
   $getContainedChildren,
@@ -1521,14 +1522,51 @@ class Draw extends XFAObject {
 
     fixDimensions(this);
 
-    if (this.w !== "" && this.h === "" && this.value) {
-      const text = this.value[$text]();
-      if (text) {
-        const { height } = layoutText(text, this.font.size, {
-          width: this.w,
-          height: Infinity,
-        });
-        this.h = height || "";
+    if ((this.w === "" || this.h === "") && this.value) {
+      const maxWidth = this.w === "" ? availableSpace.width : this.w;
+      const fonts = getRoot(this)[$fonts];
+      let font = this.font;
+      if (!font) {
+        let parent = this[$getParent]();
+        while (!(parent instanceof Template)) {
+          if (parent.font) {
+            font = parent.font;
+            break;
+          }
+          parent = parent[$getParent]();
+        }
+      }
+
+      let height = null;
+      let width = null;
+      if (
+        this.value.exData &&
+        this.value.exData[$content] &&
+        this.value.exData.contentType === "text/html"
+      ) {
+        const res = layoutText(
+          this.value.exData[$content],
+          font,
+          fonts,
+          maxWidth
+        );
+        width = res.width;
+        height = res.height;
+      } else {
+        const text = this.value[$text]();
+        if (text) {
+          const res = layoutText(text, font, fonts, maxWidth);
+          width = res.width;
+          height = res.height;
+        }
+      }
+
+      if (width !== null && this.w === "") {
+        this.w = width;
+      }
+
+      if (height !== null && this.h === "") {
+        this.h = height;
       }
     }
 
@@ -2622,7 +2660,7 @@ class Font extends XFAObject {
     ]);
     this.posture = getStringOption(attributes.posture, ["normal", "italic"]);
     this.size = getMeasurement(attributes.size, "10pt");
-    this.typeface = attributes.typeface || "";
+    this.typeface = attributes.typeface || "Myriad Pro";
     this.underline = getInteger({
       data: attributes.underline,
       defaultValue: 0,
@@ -4483,7 +4521,6 @@ class Template extends XFAObject {
         children: [],
       });
     }
-
     this[$extra] = {
       overflowNode: null,
       pageNumber: 1,