{"id":52,"date":"2016-09-29T16:04:51","date_gmt":"2016-09-29T16:04:51","guid":{"rendered":"http:\/\/blogs.igalia.com\/tonikitoo\/?p=52"},"modified":"2016-09-29T16:44:02","modified_gmt":"2016-09-29T16:44:02","slug":"making-chromiums-pdfium-greater","status":"publish","type":"post","link":"https:\/\/blogs.igalia.com\/tonikitoo\/2016\/09\/29\/making-chromiums-pdfium-greater\/","title":{"rendered":"Making Chromium’s PDFium greater"},"content":{"rendered":"

One of the coolest things about working at Igalia<\/a>, is that we have the chance to work on real world problems from day 1, either by improving an existing solution or creating new ones, based on open source technologies.<\/p>\n

During the past month, I worked on a specific aspect of Chromium’s<\/a> PDF engine, PDFium<\/a>:\u00a0extending PDFium’s capability so that Annotations can be manipulated by Acrobat JS<\/a> scripting.<\/p>\n

In practice, the goal was to make the specific scenario described on bug 492<\/a> to behave\u00a0similarly in Chromium\/PDFium, when compared to IE\/Acrobat.<\/p>\n

“Mission given, is mission accomplished”<\/h4>\n

The following Acrobat JS APIs were implemented:<\/p>\n

– Document::URL<\/a>
\n– Document::gotoNamedDest<\/a>
\n– Document::syncAnnotScan<\/a> (only stubbed out)
\n– Document::getAnnot<\/a>
\n– Document::getAnnots<\/a><\/p>\n

… and Acrobat JS scripts like the snippet below can now be successfully executed by PDFium, making 200k+ PDF files to behave in Chromium\/PDFium similarly to IE\/Acrobat –
\nwith regards to the way Annotations are manipulated.<\/p>\n

1 this.disclosed = true;\r\n2 this.syncAnnotScan();\r\n3 var u = this.URL;\r\n4 var args = u.split('calcrtid=');\r\n5 var tags;\r\n6 if (args.length > 1) {\r\n7   tags = args[1].split('&');\r\n8   this.gotoNamedDest(tags[0]);\r\n9   var a = this.getAnnot(this.pageNum, tags[0]);\r\n10   if (a != null) {\r\n11     a.hidden = false;\r\n12     var rect = a.rect;\r\n13     if (rect != null) {\r\n14       this.scroll(rect[2] + 100, rect[3] + 100);\r\n15     }\r\n16   }\r\n17 }\r\n<\/pre>\n
Also as part of this project, it was possible to\u00a0fix two discrepancies in PDFium, again if compared to Acrobat: 1) Hidden annotations were shown<\/a>; 2)\u00a0PDFium’s positioning Text Markup annotations<\/a>.<\/p>\n
About (2) specifically,\u00a0on Acrobat if an “\/AP” clause is present as part of an “\/Annot” definition, the coordinates used to draw annotations come from “\/Rect”\u00a0values. On the other hand, when “\/AP” is not defined, “\/QuadPoints” values are used to grab the annotation coordinates from.\u00a0The divergence was coming\u00a0from the fact that PDFium uses “\/Rect” regardless\u00a0the presence or absence of “\/AP”.\u00a0This CL<\/a> fixed it.<\/p>\n
To wrap up the work, I was also able to extend PDFium’s main testing tool (pdfium_test<\/a>) to actually run Acrobat JS tests<\/a>\u00a0(follow up<\/a>), as<\/a> well<\/a> as<\/a> make<\/a> lots<\/a> of<\/a> driven-by<\/a>\u00a0clean ups, and change PDFium’s main client<\/a> (Chromium) to deal better when its async nature<\/a>.<\/p>\n
The PDFium community<\/h4>\n
Throughout the project, interactions with the PDFium community were\u00a0smooth:<\/p>\n