2918 Commits

Author SHA1 Message Date
Brendan Dahl
0bef50d56d Fix two cmap related issues.
In issue , there's a char code mapped to a non-
existing glyph which shouldn't be drawn. However, we
saw it was missing and tried to then use the post table and
end up mapping it incorrectly.

This illuminated a problem with issue  and bug
893730 where glyphs disappeared after above fix.  This was
from the cmap returning the wrong glyph id. Which in turn was
caused because the font had multiple of the same type of cmap
table and we were choosing the last one. Now, we instead
default to the first one. I'm unsure if we should instead be
merging the multiple cmaps, but using only the first one works.
2017-08-03 22:19:36 -07:00
Yury Delendik
a1dfbec532 Properly cancel streams and guard at getTextContent. 2017-08-03 16:36:46 -05:00
Jonas Jenwald
e20d4a9c21 Merge pull request from brendandahl/glyph-ids
Fix several issues with glyph id mappings (issue 8668, bug 1383504)
2017-08-03 14:25:34 +02:00
Brendan Dahl
5b7f712ca7 Merge pull request from yurydelendik/issue-8591
Fallback on font widths if CFF data is broken
2017-08-02 10:53:14 -07:00
Apoorv Mishra
a129de7bd1 Add unit-tests for colorspace.js
Added unit-tests for DeviceGray, DeviceRGB and DeviceCMYK

Added unit-tests for CalGray

Added unit-tests for CalRGB

Removed redundant code

Added unit-tests for LabCS

Added unit-tests for IndexedCS

Update comment

Change lookup to Uint8Array as mentioned in pdf specs(these tests will pass after PR  is merged).

Added unit-tests for AlternateCS

Resolved code-style issues

Fixed code-style issues

Addressed issues pointed out in https://github.com/mozilla/pdf.js/pull/8611#pullrequestreview-52865469
2017-07-28 14:24:56 +05:30
Yury Delendik
343b4dc2b6 Merge pull request from mukulmishra18/network-streaming
Adds Streams API support for networking task of PDF.js project.
2017-07-27 16:15:06 -05:00
Mukul Mishra
109106794d Adds Streams API support for networking task of PDF.js project.
network.js file moved to main thread and `PDFNetworkStream` implemented
at worker thread, that is used to ask for data whenever worker needs.
2017-07-28 02:32:30 +05:30
Brendan Dahl
ac33358e1f Fix several issues with glyph id mappings.
The initial issue with  was I added a missing glyphs
check to adjustMapping, but this caused us to skip re-mapping
a glyph if the fontCharCode was a missingGlyph which in turn
caused us to overwrite a valid glyph id with an invalid one. While
fixing this, I also added a warning if the private use area is full since
this also accidentally happened when I made a different mistake.

This brought to light a number of issues where we map
missing glyphs to notdef, but often the notdef is actually defined
and then ends up being drawn. Now the glyphs don't get
mapped in toFontChar and so they are not drawn by the canvas.

Fixing the above brought up another issue though in bug1050040.pdf.
In this PDF, the font fails to load by the browser and before we were still
drawing the glyphs because it looked like the font had them, but with the fixes
above the glyphs showed up as missing so we didn't attempt draw them. To
fix this, I now throw an error when the loca table is in really bad shape and
we fall back to trying to use a system font. We now also use this fall back if
there are any format errors during converting fonts.
2017-07-26 13:00:55 -07:00
Tim van der Meij
37ac8f8623 Merge pull request from Snuffleupagus/issue-8697
Add a fallback for non-embedded SegoeUISymbol font (issue 8697)
2017-07-25 22:35:52 +02:00
Tim van der Meij
44a5cec25e Merge pull request from apoorv-mishra/fix-colorspace
Fix TypeError that occurs in colorspace.js on accidentally passing an 'Array' instead of 'TypedArray'
2017-07-25 22:13:20 +02:00
Yury Delendik
c830021b07 Fixes CFF data glyph widths 2017-07-25 12:29:51 -05:00
Jonas Jenwald
23ec6b16ca Add a fallback for non-embedded SegoeUISymbol font (issue 8697)
The PDF file uses a non-embedded SegoeUISymbol font, which is *not* a standard font (and is mainly used by Microsoft, see https://en.wikipedia.org/wiki/Segoe).

Fixes 8697.
2017-07-25 12:45:11 +02:00
Tim van der Meij
af71ea7a7d Merge pull request from Snuffleupagus/api-pageMode
[api-minor] Add support for PageMode in the API and viewer (issue 8657)
2017-07-23 13:17:07 +02:00
Tim van der Meij
e7cddcce28 Merge pull request from Snuffleupagus/rm-assert
Remove most `assert()` calls (issue 8506)
2017-07-22 19:42:24 +02:00
Tim van der Meij
7ded895d0c Merge pull request from Snuffleupagus/issue-4926-built-in-jpg
In `src/core/jpg.js`, ensure that the Adobe JPEG marker always takes precedence, even when the color transform code is zero
2017-07-22 17:25:09 +02:00
Jonas Jenwald
814fa1dee3 Remove most assert() calls (issue 8506)
This replaces `assert` calls with `throw new FormatError()`/`throw new Error()`.
In a few places, throwing an `Error` (which is what `assert` meant) isn't correct since the enclosing function is supposed to return a `Promise`, hence some cases were changed to `Promise.reject(...)` and similarily for `createPromiseCapability` instances.
2017-07-21 18:51:02 +02:00
Apoorv Mishra
d14956d4b8 Fix TypeError that occurs in colorspace.js on accidentally passing an 'Array' instead of 'TypedArray'
Fix TypeError that occurs in colorspace.js on accidentally passing an 'Array' instead of 'TypedArray'

Changed getRgbItem(...) to getRgbBuffer(...) since this.lookup has values in range[0, 255] whereas getRgbItem(...) expects those to be in range [0, 1]

Revert changes for IE9 compatibility
2017-07-21 01:15:05 +05:30
Jonas Jenwald
15f0963f51 Fix a typo, in the Catalog.numPages getter, than prevents shadowing from working correctly
Looking at the blame, it seems that this typo was present even before PR 700 (almost six years ago).
The result of using `'num'`, rather than the *correct* `'numPages'` string, is that the `Catalog.numPages` getter isn't actually being shadowed.
2017-07-20 12:35:09 +02:00
Jonas Jenwald
16c5d41c5b [api-minor] Add support for PageMode in the API (issue 8657)
Please refer to https://wwwimages2.adobe.com/content/dam/Adobe/en/devnet/pdf/pdfs/PDF32000_2008.pdf#page=82.
2017-07-19 16:40:03 +02:00
Jonas Jenwald
e2ea9b693c In src/core/jpg.js, ensure that the Adobe JPEG marker always takes precedence, even when the color transform code is zero
According to the PDF specification, please see http://www.adobe.com/content/dam/Adobe/en/devnet/acrobat/pdfs/PDF32000_2008.pdf#G6.2394361, if an Adobe JPEG marker is present it should always take precedence. This even seem to be consistent with the existing comment that is present in the code.
Hence it seems reasonable to interpret `transformCode === 0` as no color conversion being necessary.

Fixes the rendering of page 1 in `issue-4926` (from the test-suite), when the built-in `src/core/jpg.js` image decoder is used.
2017-07-11 17:08:30 +02:00
Yury Delendik
d028c26210 Removes error() 2017-07-07 09:40:24 -05:00
Jonas Jenwald
ea71d23f74 Fix a stupid spelling error in the ASCII85Decode name in Parser.makeInlineImage (issue 8613)
This is a trivial follow-up to PR 5383, and it's a bit strange that this has been wrong since late 2014 without anyone noticing (maybe because inline images aren't too common).
So, apparently code works better if you actually spell correctly, who knew ;-)

Fixes 8613.
2017-07-05 19:43:09 +02:00
Yury Delendik
b3bac5100c Merge pull request from mukulmishra18/proper-read-result
Fixes wrong structure of fullReader.read() result.
2017-07-05 09:03:57 -05:00
Jonas Jenwald
eff257b820 Merge pull request from brendandahl/missing-glyf
Fix how we detect and handle missing glyph data.
2017-07-04 12:16:07 +02:00
Brendan Dahl
9f5c1550ed Merge pull request from brendandahl/cmap-3-0
Only mask char codes of (3, 0) cmap tables in the range of 0xF000 to 0…
2017-07-03 17:58:28 -07:00
Brendan Dahl
efbbd8533f Only mask char codes of (3, 0) cmap tables in the range of 0xF000 to 0xF0FF. 2017-07-03 13:13:46 -07:00
Brendan Dahl
6d4f748fb1 Fix how we detect and handle missing glyph data. 2017-07-03 13:06:06 -07:00
Mukul Mishra
308a83e5ca Fixes wrong structure of fullReader.read() result. 2017-07-01 15:52:47 +05:30
Jonas Jenwald
de0e7a9a68 Check that the MessageHandler isn't already terminated in the onFailure handler in src/core/worker.js (issue 8584)
All other code-paths already checks that the `MessageHandler` isn't terminated, but apparently `onFailure` was missing that check (compare e.g. with the `onSuccess` function).
From what I can tell, this is only an issue if workers are *disabled*, hence why I didn't bother adding a unit-test.

Fixes 8584.
2017-06-30 10:11:13 +02:00
Brendan Dahl
a8a8909d2d Fix missing notdef in expert encoding. 2017-06-29 12:12:39 -07:00
Brendan Dahl
f1f9d98519 Merge pull request from Snuffleupagus/issue-8480
Only special-case OpenType fonts with `CFF` data if it's both a composite (i.e. Type0) font and also has a non-default CID to GID map (issue 8480)
2017-06-23 13:36:58 -07:00
Yury Delendik
e2ca894fec Merge pull request from mukulmishra18/streams-getTextContent
Streams get text content
2017-06-23 12:52:13 -05:00
Jonas Jenwald
73234577e1 Rename map to _map inside of Dict, to make it clearer that it should be regarded as a "private" property 2017-06-17 17:32:00 +02:00
Mukul Mishra
0c13d0ff46 Adds Streams API in getTextContent to stream data.
This patch adds Streams API support in getTextContent
so that we can stream data in chunks instead of fetching
whole data from worker thread to main thread. This patch
supports Streams API without changing the core functionality
of getTextContent.

Enqueue textContent directly at getTextContent in partialEvaluator.

Adds desiredSize and ready property in streamSink.
2017-06-17 20:03:27 +05:30
Jonas Jenwald
3a20fd165f Refactor ObjectLoader to use Dicts correctly, rather than abusing their internal properties
The `ObjectLoader` currently takes an Object as input, despite actually working with `Dict`s internally. This means that at the (two) existing call-sites, we're passing in the "private" `Dict.map` property directly.

Doing this seems like an anti-pattern, and we could (and even should) simply provide the actual `Dict` when creating an `ObjectLoader` instance.
Accessing properties stored in the `Dict` is now done using the intended methods instead, in particular `getRaw` which (as the name suggests) doesn't do any de-referencing, thus maintaining the current functionality of the code.

The only functional change in this patch is that `ObjectLoader.load` will now ignore empty nodes, such that `ObjectLoader._walk` only needs to deal with nodes that are known to contain data. (This lets us skip, among other checks, meaningless `addChildren` function calls.)
2017-06-16 22:59:32 +02:00
Jonas Jenwald
f2fc9ee281 Slightly refactor and ES6-ify the code in ObjectLoader
This patch changes all `var` to `let`, and caches the array lengths in all loops. Also removes two unnecessary temporary variable assignments.
2017-06-16 22:59:32 +02:00
Jonas Jenwald
e589834f13 Ensure that TilingPatterns have valid (non-zero) /BBox arrays (issue 8330)
Fixes 8330.
2017-06-09 21:41:48 +02:00
Jonas Jenwald
8b4a42e5b8 Only special-case OpenType fonts with CFF data if it's both a composite (i.e. Type0) font and also has a non-default CID to GID map (issue 8480)
*As mentioned the last time that I touched this particular part of the font code, I'm sincerely hope that this doesn't cause any regressions!*

However, the patch passes all tests added in PRs 5770, 6270, and 7904 (and obviously all other tests as well). Furthermore, I've manually checked all the issues/bugs referenced in those PRs without finding any issues.

Fixes 8480.
2017-06-09 21:15:39 +02:00
Jonas Jenwald
999e30723d Reduce the duplication slightly when detecting an OpenType font (in the Font constructor) 2017-06-09 18:26:57 +02:00
Jonas Jenwald
a8c87f8019 Fix inconsistent spacing and trailing commas in objects in src/core/ files, so we can enable the comma-dangle and object-curly-spacing ESLint rules later on
*Unfortunately this patch is fairly big, even though it only covers the `src/core` folder, but splitting it even further seemed difficult.*

http://eslint.org/docs/rules/comma-dangle
http://eslint.org/docs/rules/object-curly-spacing

Given that we currently have quite inconsistent object formatting, fixing this in *one* big patch probably wouldn't be feasible (since I cannot imagine anyone wanting to review that); hence I've opted to try and do this piecewise instead.

Please note: This patch was created automatically, using the ESLint --fix command line option. In a couple of places this caused lines to become too long, and I've fixed those manually; please refer to the interdiff below for the only hand-edits in this patch.

```diff
diff --git a/src/core/evaluator.js b/src/core/evaluator.js
index abab9027..dcd3594b 100644
--- a/src/core/evaluator.js
+++ b/src/core/evaluator.js
@@ -2785,7 +2785,8 @@ var EvaluatorPreprocessor = (function EvaluatorPreprocessorClosure() {
     t['Tz'] = { id: OPS.setHScale, numArgs: 1, variableArgs: false, };
     t['TL'] = { id: OPS.setLeading, numArgs: 1, variableArgs: false, };
     t['Tf'] = { id: OPS.setFont, numArgs: 2, variableArgs: false, };
-    t['Tr'] = { id: OPS.setTextRenderingMode, numArgs: 1, variableArgs: false, };
+    t['Tr'] = { id: OPS.setTextRenderingMode, numArgs: 1,
+                variableArgs: false, };
     t['Ts'] = { id: OPS.setTextRise, numArgs: 1, variableArgs: false, };
     t['Td'] = { id: OPS.moveText, numArgs: 2, variableArgs: false, };
     t['TD'] = { id: OPS.setLeadingMoveText, numArgs: 2, variableArgs: false, };
diff --git a/src/core/jbig2.js b/src/core/jbig2.js
index 5a17d482..71671541 100644
--- a/src/core/jbig2.js
+++ b/src/core/jbig2.js
@@ -123,19 +123,22 @@ var Jbig2Image = (function Jbig2ImageClosure() {
      { x: -1, y: -1, }, { x: 0, y: -1, }, { x: 1, y: -1, }, { x: -2, y: 0, },
      { x: -1, y: 0, }],
     [{ x: -3, y: -1, }, { x: -2, y: -1, }, { x: -1, y: -1, }, { x: 0, y: -1, },
-     { x: 1, y: -1, }, { x: -4, y: 0, }, { x: -3, y: 0, }, { x: -2, y: 0, }, { x: -1, y: 0, }]
+     { x: 1, y: -1, }, { x: -4, y: 0, }, { x: -3, y: 0, }, { x: -2, y: 0, },
+     { x: -1, y: 0, }]
   ];

   var RefinementTemplates = [
     {
       coding: [{ x: 0, y: -1, }, { x: 1, y: -1, }, { x: -1, y: 0, }],
-      reference: [{ x: 0, y: -1, }, { x: 1, y: -1, }, { x: -1, y: 0, }, { x: 0, y: 0, },
-                  { x: 1, y: 0, }, { x: -1, y: 1, }, { x: 0, y: 1, }, { x: 1, y: 1, }],
+      reference: [{ x: 0, y: -1, }, { x: 1, y: -1, }, { x: -1, y: 0, },
+                  { x: 0, y: 0, }, { x: 1, y: 0, }, { x: -1, y: 1, },
+                  { x: 0, y: 1, }, { x: 1, y: 1, }],
     },
     {
-      coding: [{ x: -1, y: -1, }, { x: 0, y: -1, }, { x: 1, y: -1, }, { x: -1, y: 0, }],
-      reference: [{ x: 0, y: -1, }, { x: -1, y: 0, }, { x: 0, y: 0, }, { x: 1, y: 0, },
-                  { x: 0, y: 1, }, { x: 1, y: 1, }],
+      coding: [{ x: -1, y: -1, }, { x: 0, y: -1, }, { x: 1, y: -1, },
+               { x: -1, y: 0, }],
+      reference: [{ x: 0, y: -1, }, { x: -1, y: 0, }, { x: 0, y: 0, },
+                  { x: 1, y: 0, }, { x: 0, y: 1, }, { x: 1, y: 1, }],
     }
   ];
```
2017-06-02 11:20:19 +02:00
Jonas Jenwald
982b6aa65b Convert the files in the /src/core folder to ES6 modules
Please note that the `glyphlist.js` and `unicode.js` files are converted to CommonJS modules instead, since Babel cannot handle files that large and they are thus excluded from transpilation.
2017-05-30 22:06:21 +02:00
Jonas Jenwald
4ce5e520fb Add different code-paths to {CMap, ToUnicodeMap}.charCodeOf depending on length, since Array.prototype.indexOf can be extremely inefficient for very large arrays (issue 8372)
Fixes 8372.
2017-05-24 19:47:04 +02:00
Jonas Jenwald
31c24ed631 Don't map glyphs to the HANGUL FILLER (0x3164) Unicode location (issue 8424)
*This patch follows a similar pattern as previous ones, by skipping certain problematic Unicode locations.*

According to http://searchfox.org/mozilla-central/rev/6c2dbacbba1d58b8679cee700fd0a54189e0cf1b/gfx/harfbuzz/src/hb-unicode-private.hh#136, it seems that the HANGUL FILLER (0x3164) location is "special".

Fixes 8424.
2017-05-23 16:12:45 +02:00
Jonas Jenwald
0ddf52aca5 Remove the special handling for nameddests that look like standard pageNumbers
PR 7341 added special handling for `nameddest`s that look like pageNumbers, to prevent issues since we previously *incorrectly* supported specifying a pageNumber directly in the hash; i.e. `#10` versus the correct `#page=10` format.

Since this behaviour wasn't correct, PR 7757 fixed and deprecated the old format, which means that we no longer need to maintain the `nameddest` hack in multiple files.
2017-05-20 11:29:29 +02:00
Yury Delendik
5dc8dcdc0f Merge pull request from Snuffleupagus/issue-8380
Cache JPEG images, just as we do for other image formats, in `evaluator.js` (issue 8380)
2017-05-17 17:25:51 -05:00
巴里切罗
8d5d97264e fix(svg) adjust strategy for decoding JPEG images 2017-05-08 11:32:44 +08:00
Jonas Jenwald
0c2ebda31c Cache JPEG images, just as we do for other image formats, in evaluator.js (issue 8380)
For some reason, we're putting all kind of images *except* JPEG into the `imageCache` in `evaluator.js`.[1]
This means that in the PDF file in issue 8380, we'll keep sending the *same* two small images[2] to the main-thread and decoding them over and over. This is obviously hugely inefficient!

As can be seen from the discussion in the issue, the performance becomes *extremely* bad if the user has the addon "Adblock Plus" installed. However, even in a clean Firefox profile, the performance isn't that great.

This patch not only addresses the performance implications of the "Adblock Plus" addon together with that particular PDF file, but it *also* improves the rendering times considerably for *all* users.
Locally, with a clean profile, the rendering times are reduced from `~2000 ms` to `~500 ms` for my setup!

Obviously, the general structure of the PDF file and its operator sequence is still hugely inefficient, however I'd say that the performance with this patch is good enough to consider the issue (as it stands) resolved.[3]

Fixes 8380.

---
[1] Not technically true, since inline images are cached from `parser.js`, but whatever :-)

[2] The two JPEG images have dimensions 1x2, respectively 4x2.

[3] To make this even more efficient, a new state would have to be added to the `QueueOptimizer`. Given that PDF files this stupid fortunately aren't too common, I'm not convinced that it's worth doing.
2017-05-07 13:07:41 +02:00
Yury Delendik
3adda80f97 Merge pull request from Snuffleupagus/PartialEvaluator-method-signatures
Change the signatures of the `PartialEvaluator` "constructor" and its `getOperatorList`/`getTextContent` methods to take parameter objects
2017-05-04 08:10:30 -05:00
Yury Delendik
74ba3033e8 Merge pull request from Snuffleupagus/Lexer-getNumber-ignore-line-breaks
Ignore line-breaks between operator and digit in `Lexer.getNumber`
2017-05-03 09:43:59 -05:00
Jonas Jenwald
3e20d30afc Change the signatures of the PartialEvaluator "constructor" and its getOperatorList/getTextContent methods to take parameter objects
Currently these methods accept a large number of parameters, which creates quite unwieldy call-sites. When invoking them, you have to remember not only what arguments to supply, but also the correct order, to avoid runtime errors.
Furthermore, since some of the parameters are optional, you also have to remember to pass e.g. `null` or `undefined` for those ones.
Also, adding new parameters to these methods (which happens occasionally), often becomes unnecessarily tedious (based on personal experience).

Please note that I do *not* think that we need/should convert *every* single method in `evaluator.js` (or elsewhere in `/core` files) to take parameter objects. However, in my opinion, once a method starts relying on approximately five parameter (or even more), passing them in individually becomes quite cumbersome.

With these changes, I obviously needed to update the `evaluator_spec.js` unit-tests. The main change there, except the new method signatures[1], is that it's now re-using *one* `PartialEvalutor` instance, since I couldn't see any compelling reason for creating a new one in every single test.

*Note:* If this patch is accepted, my intention is to (time permitting) see if it makes sense to convert additional methods in `evaluator.js` (and other `/core` files) in a similar fashion, but I figured that it'd be a good idea to limit the initial scope somewhat.

---

[1] A fun fact here, note how the `PartialEvaluator` signature used in `evaluator_spec.js` wasn't even correct in the current `master`.
2017-05-03 12:10:20 +02:00