The viewBox attribute of the <svg> element allows both whitespace and
commas to be used as field separators.
https://www.w3.org/TR/SVG/coords.html#ViewBoxAttribute
Bug: T194192
Change-Id: Iae9be3e4fad3a8ffa411d7a76eee2f20cc39b718
One of the skip-over loops was missing an feof() check and could
cause infinite loops.
Includes test file created by truncating a tiny tiny .jpeg at
the right place...
With the fix, it doesn't loop but dies on an exception, which
is good!
Bug: T184048
Change-Id: Ica13d6b68c3c12f7ce414edd081bf0886714e465
Test case for Opus file check so we don't regress!
Confirms the fix for T151352 that Opus files return
type 'audio' and not 'multimedia'.
Bug: T151352
Change-Id: Ib868793d90b60609ccec25726e6fbaa28bffebaf
Existing srgb.jpg & tinyrgb.jpg have been replaced to be able to
easily compare a "fixed" missingprofile.jpg to tinyrgb.jpg.
With the existing files, when the tinyrgb profile was added to
missingprofile.jpg, it would end up basically the same as tinyrgb.jpg,
except that not all the exif data would be in the exact same order.
I've rebuilt srgb.jpg & tinyrgb.jpg by first removing their profile
(which is what missingprofile.jpg is), and then copying it over again:
exiftool -tagsfromfile srgb.jpg -ICC_Profile new_srgb.jpg
Meanwhile also moved the profile-swapping code to JpegHandler, as it
was jpeg-specific.
Bug: T134498
Change-Id: I722dd6f66f6007182ad9a215e5eb382776983c05
* Add $wgJpegPixelFormat, default to 'yuv420'
* Implemented for ImageMagick via CLI and extension
* Currently ignored for other scaler backends
* Added test case to run when using ImageMagick
4:2:0 subsampling can save an average of 17% bandwidth
over 4:4:4 subsampling, at the cost of some artifacting
at sharp red or blue edges. This is usually not noticeable
in photographic images.
To restore the previous behavior, set to false:
$wgJpegPixelFormat = false;
which will maintain the original file's pixel subsampling
settings in the thumbnail.
Can set explicitly to one of:
'yuv444' - never subsample
'yuv422' - subsample 2x horizontally, not vert
'yuv420' - subsample 2x in both dimensions
Bug: T129128
Change-Id: Ib9cb36c3a7e6a69d66c11150ef4a1d02dbac2df5
Adds basic image size detection for WebP and support in the
MediaHandler. Currently renders WebP files as PNGs, because that
handles transparency.
Bug: T50519
Change-Id: I3c00653a8a034efc3f6b60fe62b7ac2e5391f921
TinyRGB is an ICC profile released by Facebook under CC0.
It is designed to be fully compatible with sRGB.
It offers the vast advantages of being much smaller than sRGB,
as well as being free as in freedom (the sRGB profile found in the
majority of JPGs is copyrighted).
This change aims to provide the ability to swap sRGB for TinyRGB at
the time thumbnails are generated.
JPGs that use another ICC profile than sRGB or no profile at all are
unaffected.
Bug: T100976
Change-Id: I2ae35ddad4e8a82db8b9541974367dc76c884e7a
Caused the animation warning to not be present on
image description pages
Also add unit tests. Since I'm adding an animated svg file for
this unit test, also add a metadata extraction test for animated
svgs, since we're missing one.
Change-Id: Id03e1a0e1c151d3c575a695a42c54b709187d10a
follow-up: 876128f8c9
Thumbnails for portrait-orientation images have always been "too big",
especially when displayed in a gallery. The 'upright' option did not
completely fix the issue. Using a square bounding box for thumbnails
(and 'framed' images) without an explicit size specifiction provides
a better default appearance.
This also provides a clean syntax for content authored using
Parsoid/Visual Editor, which prefers square bounding boxes.
See:
https://www.mediawiki.org/wiki/Requests_for_comment/Square_bounding_boxes
Bug: 63903
Change-Id: I665d8945843d3b5437a74e376b63c44965590116
SVG files support having different renderings based on language.
We support changing the rendering language, but we offer no
discoverability of what languages are available.
Long term Jarry's TranslateSVG will hopefully be used and have
all the associated awesomeness. In the mean time, we should
probably have some indication on the image page what languages
are available for the SVG. This is step 1 (extract the language).
I expect to follow this up with a commit adding some sort of
simple selector on the image page.
As an aside, it should be noted, this only detects languages
if its in the first $wgSVGMetadataCutoff (256 kb). This is a
compromise to avoid OOM on huge SVG files.
Change-Id: I2a462270fe63eb3e3023419eddc8b06f5a617ab5
Currently file metadata is handler dependant. However they usually
end up extracting the same type of data (author, date, etc) plus
one or two handler specific things. This adds a handler independent
interface for getting metadata that is likely to be common for all
types of file (At the moment, this is the exif/iptc/xmp information)
This commit used to also contain stuff adding parser functions,
which is now split to its own commit. This commit is needed
by a bunch of other commits, in particular I0d957891e0.
Change-Id: I43d9252f69dc5b8ba0b848cf40aa1b97329c85ae
In EXIF, GPSAltitude is stored as a fraction string like "1/2".
For values below sea level we were negating this value, in order
to represent the sign and the magnitude in the same value. However,
I forgot to convert that to an integer before negating it. PHP was nice
enough to do a best effort conversion of the string to an integer.
This resulted in altitudes below sea level being taken as just the
numerator of the altitude, which gives results that can be significantly
off.
Also add unit tests for the GPS related image metadata stuff. Change the
existing GPS test to use a fractional altitude (Since this issue isn't
appearent if the denominator is 1). Add tests for XMP as well, since
XMP had same issue, and has to do same processing as EXIF stuff does.
In some future time, may want to consider just converting all exif rational values
to real numbers during the extraction process for generally better sanity.
Patchset 2: rebase
Change-Id: I49032b52a4c840b28e667a6a2b8ae23c508df247
was starting to get crowded with supporting image files.
This changes the following things
*Moves the media files to a different directory
*Changes the references to the files to reference there new locations
2 semi-unrelated changes
*Modify the README file that was basically giving credit for the images where due to explain what the images are for.
*set svn:mime-type to image/svg+xml for the svg files.
