Sakurai/pdf.js
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Improve the space detection for the getTextContent function

Browse Source
This commit is contained in:
Julian Viereck 2012-09-14 10:53:06 -07:00
parent ee0c2e2ab7
commit 7432e596a1
1 changed files with 10 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

27
src/evaluator.js
View File

@ -533,36 +533,29 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
switch (cmd) { switch (cmd) {
// TODO: Add support for SAVE/RESTORE and XFORM here. // TODO: Add support for SAVE/RESTORE and XFORM here.
case 'Tf': case 'Tf':
font = handleSetFont(args[0].name).translated; font = handleSetFont(args[0].name);
break; break;
case 'TJ': case 'TJ':
var items = args[0]; var items = args[0];
for (var j = 0, jj = items.length; j < jj; j++) { for (var j = 0, jj = items.length; j < jj; j++) {
if (typeof items[j] === 'string') { if (typeof items[j] === 'string') {
chunk += fontCharsToUnicode(items[j], font); chunk += fontCharsToUnicode(items[j], font.translated);
} else if (items[j] < 0) { } else if (items[j] < 0 && font.spacedWidth > 0) {
// making all negative offsets a space - better to have var numFakeSpaces = Math.round(-e / font.spacedWidth);
// a space in incorrect place than not have them at all if (numFakeSpaces > 0) {
chunk += ' '; chunk += ' ';
// This is a better way to detect spacing in the future. }
// However, for now let's keep it simple (also, font.spacedWidth)
// is not available.
// } else if (items[j] < 0 && font.spacedWidth > 0) {
// var numFakeSpaces = Math.round(-e / font.spacedWidth);
// if (numFakeSpaces > 0) {
// chunk += ' ';
// }
} }
} }
break; break;
case 'Tj': case 'Tj':
chunk += fontCharsToUnicode(args[0], font); chunk += fontCharsToUnicode(args[0], font.translated);
break; break;
case "'": case "'":
chunk += fontCharsToUnicode(args[0], font) + ' '; chunk += fontCharsToUnicode(args[0], font.translated) + ' ';
break; break;
case '"': case '"':
chunk += fontCharsToUnicode(args[2], font) + ' '; chunk += fontCharsToUnicode(args[2], font.translated) + ' ';
break; break;
case 'Do': case 'Do':
// Set the chunk such that the following if won't add something // Set the chunk such that the following if won't add something