SVG data:uri images are not handled properly

• Component: SVG
• Status: RESOLVED DUPLICATE

[Philip Rogers]

At the moment SVG data uri images (bitmap or vector) are not allowed in SVG images. This is a huge bummer and prevents WebKit from showing many Adobe Illustrator SVG files where Illustrator filters are flattened into images.

For example:
<svg width="100" height="100" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
<!-- base64 encoded image: <svg width='100' height='100' xmlns='http://www.w3.org/2000/svg'><rect fill='green' width='100' height='100'></rect></svg> -->
<image width="100" height="100" xlink:href="data:image/svg+xml;base64,PHN2ZyB3aWR0aD0nMTAwJyBoZWlnaHQ9JzEwMCcgeG1sbnM9J2h0dHA6Ly93d3cudzMub3JnLzIwMDAvc3ZnJz48cmVjdCBmaWxsPSdncmVlbicgd2lkdGg9JzEwMCcgaGVpZ2h0PScxMDAnPjwvcmVjdD48L3N2Zz4="/>
</svg>

A full testcase is available at crbug.com/224317

This can be fixed by merging the following Blink patches:
1) https://codereview.chromium.org/16433002 - Prepopulate the memoryCache with data:uri images
(This is not a straight merge and requires plumbing the data uri parsing from somewhere, possible DataURI.h)
2) https://codereview.chromium.org/16922002 - Do not load deferred/disabled data uri images
3) https://codereview.chromium.org/17682003 - Do not fire load events from frames with scripting disabled

[Radar WebKit Bug Importer]

<rdar://problem/14281148>

[youenn fablet]

Why not simply updating CachedImage::load to enforce data:uri images loading even when network loading is disabled?

[Philip Rogers]

(In reply to comment #2)
> Why not simply updating CachedImage::load to enforce data:uri images loading even when network loading is disabled?

I'm afraid the loader code (in both Blink and WebKit) has diverged significantly since I wrote this patch. My best recollection is that doing this would result in a request being sent to the platform layer with a null networking context which could escape the SVG Image "sandbox" that prevents resource requests. This could allow an SVG image embedded in an email to work as a tracker, for example.

This may no longer be true (or may have only been true for Chromium) and I would encourage you to trace this code further to see what happens today.

[youenn fablet]

Moving data uri parsing into WebCore allows removing memory copy between WebCore and the network stack.
I did a test with run-perf-test and this makes a difference in terms of loading time for large data uri images.

I will shortly submit a patch along the lines of Blink patches.
The data URL parsing routine code originates from Source/WebCore/platform/network/DataURL.cpp.

[youenn fablet]

Created attachment 215923 [details]
Patch

[WebKit Commit Bot]

Attachment 215923 [details] did not pass style-queue:

Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'LayoutTests/ChangeLog', u'LayoutTests/loader/data-uri-images-load-synchronously-expected.txt', u'LayoutTests/loader/data-uri-images-load-synchronously.html', u'LayoutTests/loader/data-uri-images-reload-synchronously-expected.txt', u'LayoutTests/loader/data-uri-images-reload-synchronously.html', u'LayoutTests/svg/as-image/resources/svg-with-data-uri.svg', u'LayoutTests/svg/as-image/svg-as-image-with-data-uri-expected.txt', u'LayoutTests/svg/as-image/svg-as-image-with-data-uri.html', u'Source/WTF/ChangeLog', u'Source/WTF/wtf/text/Base64.cpp', u'Source/WTF/wtf/text/Base64.h', u'Source/WebCore/ChangeLog', u'Source/WebCore/dom/Document.cpp', u'Source/WebCore/loader/ImageLoader.cpp', u'Source/WebCore/loader/cache/CachedRawResource.cpp', u'Source/WebCore/loader/cache/CachedRawResource.h', u'Source/WebCore/loader/cache/CachedResourceLoader.cpp']" exit_code: 1
Source/WebCore/loader/cache/CachedResourceLoader.cpp:40:  Alphabetical sorting problem.  [build/include_order] [4]
Total errors found: 1 in 17 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Build Bot]

Comment on attachment 215923 [details]
Patch

Attachment 215923 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.appspot.com/results/19668774

New failing tests:
fast/forms/basic-buttons.html
compositing/tiling/huge-layer-img.html

[Build Bot]

Created attachment 215925 [details]
Archive of layout-test-results from webkit-ews-13 for mac-mountainlion-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: webkit-ews-13  Port: mac-mountainlion-wk2  Platform: Mac OS X 10.8.5

[Build Bot]

Comment on attachment 215923 [details]
Patch

Attachment 215923 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.appspot.com/results/19658832

New failing tests:
fast/forms/basic-buttons.html
compositing/tiling/huge-layer-img.html

[Build Bot]

Created attachment 215931 [details]
Archive of layout-test-results from webkit-ews-07 for mac-mountainlion

The attached test failures were seen while running run-webkit-tests on the mac-ews.
Bot: webkit-ews-07  Port: mac-mountainlion  Platform: Mac OS X 10.8.5

[Build Bot]

Comment on attachment 215923 [details]
Patch

Attachment 215923 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.appspot.com/results/19618811

New failing tests:
fast/forms/basic-buttons.html
compositing/tiling/huge-layer-img.html

[Build Bot]

Created attachment 215938 [details]
Archive of layout-test-results from webkit-ews-04 for mac-mountainlion

The attached test failures were seen while running run-webkit-tests on the mac-ews.
Bot: webkit-ews-04  Port: mac-mountainlion  Platform: Mac OS X 10.8.5

[Philip Rogers]

Comment on attachment 215923 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=215923&action=review

> Source/WebCore/ChangeLog:5
> +

Please add more information about what this patch does. It may be useful to list the specific patches being merged as well.

> Source/WebCore/loader/cache/CachedResourceLoader.cpp:164
> +static PassRefPtr<ResourceBuffer> parseDataURI(const URL &url, ResourceResponse &response)

This seems to be an exact copy of WebCore/platform/network/DataURL.cpp's handleDataURL. Can these be unified?

> Source/WebCore/loader/cache/CachedResourceLoader.cpp:246
> +            memoryCache()->add(resource);

I think this may cause a security bug in the xml parser, see https://chromiumcodereview.appspot.com/18226005

[youenn fablet]

Created attachment 216279 [details]
Patch

[EFL EWS Bot]

Comment on attachment 216279 [details]
Patch

Attachment 216279 [details] did not pass efl-ews (efl):
Output: http://webkit-queues.appspot.com/results/22508163

[EFL EWS Bot]

Comment on attachment 216279 [details]
Patch

Attachment 216279 [details] did not pass efl-wk2-ews (efl-wk2):
Output: http://webkit-queues.appspot.com/results/22698186

[kov's GTK+ EWS bot]

Comment on attachment 216279 [details]
Patch

Attachment 216279 [details] did not pass gtk-ews (gtk):
Output: http://webkit-queues.appspot.com/results/22748151

[Build Bot]

Comment on attachment 216279 [details]
Patch

Attachment 216279 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.appspot.com/results/22738141

[Build Bot]

Comment on attachment 216279 [details]
Patch

Attachment 216279 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.appspot.com/results/22778136

[Build Bot]

Comment on attachment 216279 [details]
Patch

Attachment 216279 [details] did not pass win-ews (win):
Output: http://webkit-queues.appspot.com/results/22368149

[youenn fablet]

Created attachment 216297 [details]
Patch

[Build Bot]

Comment on attachment 216297 [details]
Patch

Attachment 216297 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.appspot.com/results/22638205

[Build Bot]

Comment on attachment 216297 [details]
Patch

Attachment 216297 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.appspot.com/results/22538193

[youenn fablet]

Created attachment 216366 [details]
Patch

[Build Bot]

Comment on attachment 216366 [details]
Patch

Attachment 216366 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.appspot.com/results/22638447

[Build Bot]

Comment on attachment 216366 [details]
Patch

Attachment 216366 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.appspot.com/results/22508463

[youenn fablet]

Created attachment 217540 [details]
xcode proj fixed

[Build Bot]

Comment on attachment 217540 [details]
xcode proj fixed

Attachment 217540 [details] did not pass mac-wk2-ews (mac-wk2):
Output: http://webkit-queues.appspot.com/results/29928217

New failing tests:
fast/forms/basic-buttons.html
compositing/tiling/huge-layer-img.html

[Build Bot]

Created attachment 217545 [details]
Archive of layout-test-results from webkit-ews-13 for mac-mountainlion-wk2

The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews.
Bot: webkit-ews-13  Port: mac-mountainlion-wk2  Platform: Mac OS X 10.8.5

[Build Bot]

Comment on attachment 217540 [details]
xcode proj fixed

Attachment 217540 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.appspot.com/results/29858088

New failing tests:
fast/forms/basic-buttons.html
compositing/tiling/huge-layer-img.html

[Build Bot]

Created attachment 217548 [details]
Archive of layout-test-results from webkit-ews-07 for mac-mountainlion

The attached test failures were seen while running run-webkit-tests on the mac-ews.
Bot: webkit-ews-07  Port: mac-mountainlion  Platform: Mac OS X 10.8.5

[youenn fablet]

Created attachment 217874 [details]
updated mac baselines

failing mac tests results are slightly changing due to data urls loading being synchronous with the patch, hence updating the expected tests results.

[youenn fablet]

Patch should be simplified once data:// url handling is migrated into WebCore (cf. bug 109199).

[Steven Vachon]

http://www.svachon.com/webframes/examples.html

The non-vector examples do not work in WebKit/Safari, but they work everywhere else. Please fix this in time for Safari 8's release.

[Alexey Proskuryakov]

*** This bug has been marked as a duplicate of bug 99677 ***