summaryrefslogtreecommitdiff
path: root/tests/fixtures
Commit message (Collapse)AuthorAgeFilesLines
* meta: Move lib and bins into separate cratesSophie Herold2023-03-311630-57928/+0
| | | | | | Closes #950 Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/822>
* Remove the test filter-morphology-from-reference-page.svgFederico Mena Quintero2023-03-302-37/+0
| | | | | | | | | | | | | This was added during the port of the filters code to Rust, but it is redundant with tests/fixtures/reftests/svg1.1/filters-morph-01-f.svg Also this file produces frequent failures in the test suite whenever the font rendering stack changes even just a bit. The point of the test is to test the feMorphology, not text rendering - so remove it in favor of the other test from the SVG1.1 test suite, which does erosion and dilation filters on paths, not on text. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/817>
* Add a test for xinclude recursion across two filesFederico Mena Quintero2023-03-142-0/+6
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/810>
* (#942): Fix crash when XML files get recursively included through XIncludeFederico Mena Quintero2023-03-141-0/+3
| | | | | | | | | | | | | | | | | | | | There is a new limit, MAX_XINCLUDE_DEPTH, which is a constant with the maximum level of nesting for XInclude. We keep a counter of the current nesting level in XmlStateInner, and check against the limit every time we need to xinclude another XML document. The sample file has <xi:include parse="xml" href=""/> which properly causes the *same* file to be included, per https://www.w3.org/TR/xinclude-11/#include_element The href attribute is optional; the absence of this attribute is the same as specifying href="", that is, the reference is to the same document. Fixes https://gitlab.gnome.org/GNOME/librsvg/-/issues/942 Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/810>
* (#930): Validate all clipPath and mask transformsFederico Mena Quintero2023-02-222-0/+13
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously, an invalid "transform" property specified for clipPath would bubble up as a non-recoverable "invalid matrix" error from Cairo, stopping rendering at that point. Instead, we want to validate all transforms and just not render the offending element. This commit has many changes: * Introduce a ValidTransform newtype, created with "impl TryFrom<Transform> for Transform". This sees if the transform is invertible before wrapping it. The idea is to do all transform operations first, then see if the result is valid by trying to convert to a ValidTransform. * There is no "impl From<Transform> for cairo::Matrix" anymore, but there is From<ValidTransform>. That way, we guarantee that an unvalidated transform cannot be easily set on a Cairo context. * DrawingCtx::get_transform() returns a ValidTransform, since a) it comes from the cr, and Cairo already validated that transform for us. Otherwise, the cr would have been in an error state. * There is a new variant RenderingError::InvalidTransform, which lets us use "?" everywhere as normal. This error case is caught in Node::draw() and the node is just not drawn. Eventually we can move to a scheme where we distinguish between fatal errors (e.g. RenderingError::LimitExceeded) and recoverable ones. Fixes https://gitlab.gnome.org/GNOME/librsvg/-/issues/930 Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/800>
* Add test for #932Federico Mena Quintero2023-02-141-0/+9
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/796>
* (#743): implement the feDropShadow elementFederico Mena Quintero2023-02-092-0/+27
| | | | | | Fixes https://gitlab.gnome.org/GNOME/librsvg/-/issues/743 Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/793>
* (#928): Fix panic when a feTile filter gets an empty source regionFederico Mena Quintero2023-01-241-0/+12
| | | | | | | | | | When the input_bounds from the filter source is empty, we tried to create a zero-sized image for tiling. This is not allowed. There is nothing to tile in that case, so just return the original image to filter. Fixes https://gitlab.gnome.org/GNOME/librsvg/-/issues/928 Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/787>
* text tests: Rename test files that start with a bug numberFederico Mena Quintero2022-11-284-0/+0
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/776>
* reference tests: Rename test files that start with a bug numberFederico Mena Quintero2022-11-2895-0/+0
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/776>
* errors tests: Rename test files that start with a bug numberFederico Mena Quintero2022-11-283-0/+0
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/776>
* dimensions tests: Rename test files that start with a bug numberFederico Mena Quintero2022-11-284-1/+1
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/776>
* crash tests: Rename test files that start with a bug numberFederico Mena Quintero2022-11-2813-0/+0
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/776>
* cmdline tests: Rename test files that start with a bug numberFederico Mena Quintero2022-11-283-0/+0
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/776>
* api tests: Rename test files that start with a bug numberFederico Mena Quintero2022-11-281-0/+0
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/776>
* render-crash tests: Rename test files that start with a bug numberFederico Mena Quintero2022-11-2824-0/+0
| | | | | | | From 123-blahblah.svg to bug123-blahblah.svg, so that it's easier to turn them into Rust symbols like bug123_blahblah later. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/776>
* Add a test for non-positive stdDeviation in feGaussianBlurFederico Mena Quintero2022-11-042-0/+26
| | | | | | | | | | | This is https://gitlab.gnome.org/GNOME/librsvg/-/issues/913 and it already works fine, but more or less by accident - the feGaussianBlur filter primitive gets marked as "in error" due to the negative stdDeviation, and the filters code skips it. This test is for future-proofing the refactoring that follows. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/763>
* Implement vector-effect="non-scaling-stroke"Michael Howell2022-09-052-0/+8
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/740>
* Apply width, height when `<use>`ing an `<svg>`Michael Howell2022-09-022-0/+13
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/739>
* (#92): In the use element, honor the overflow:hidden property of a ↵Federico Mena Quintero2022-08-262-0/+17
| | | | | | | | | | | referenced symbol We meant to look at the referenced <symbol>'s computed values, not the <use>'s computed values. Fixes https://gitlab.gnome.org/GNOME/librsvg/-/issues/92 Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/737>
* Try changing the invalid URL in 524-invalid-stylesheet-href.svgFederico Mena Quintero2022-08-091-1/+1
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/728>
* tests/api.c: Test the case where get_intrinsic_dimensions() reports no viewBoxFederico Mena Quintero2022-08-091-0/+1
| | | | | | | | After SVG2, where width/height always compute to 100% because their initial value is "auto", has_width and has_height will always be TRUE. It is only has_viewbox that needs testing. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/727>
* Add missing test for rsvg_handle_set_stylesheet()Federico Mena Quintero2022-08-092-0/+9
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/727>
* (#880): Accept patterns with userSpaceOnUse units even if there is no object ↵Federico Mena Quintero2022-07-112-0/+17
| | | | | | | | bounding box Fixes https://gitlab.gnome.org/GNOME/librsvg/-/issues/880 Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/713>
* Add test for a rect in an isolation contextFederico Mena Quintero2022-05-092-0/+20
| | | | | | This still passes; the next one will fail. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/697>
* Use JSON fixtures for the primitive_geometries testsFederico Mena Quintero2022-05-096-0/+60
| | | | | | Finally, something I've been wanting to do for a while. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/697>
* Add test files from the Horizon EDA projectFederico Mena Quintero2022-05-069-0/+3229
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/697>
* (#823): Regression in geometry computation when the SVG viewBox does not ↵Federico Mena Quintero2022-04-291-0/+8
| | | | | | | | | | | | | | | | | match the default viewport Curiously, this manifests in the same way as #521, and is fixed by removing the code that I added to fix that one. I'm thinking that all the cleanups around viewport and viewBox handling caused the old code to not be needed anymore. (Found by adding the same debug output to librsvg-2.48 and the main branch, and comparing executions.) Fixes https://gitlab.gnome.org/GNOME/librsvg/-/issues/823 Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/695>
* (#721): Catch circular references when drawing patternsFederico Mena Quintero2022-03-142-0/+17
| | | | | | | | | It seems I broke this when splitting patterns into the resolve stage and the drawing stage. Fixes https://gitlab.gnome.org/GNOME/librsvg/-/issues/721 Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/679>
* Enable the SVG1.1 tests with text scaled non-proportionallyFederico Mena Quintero2022-03-094-0/+0
| | | | | | | | | | | These tests: pservers-grad-07-b-ref.png pservers-grad-08-b-ref.png They both have text that has different scaling in the x/y dimensions; they are rendered correctly now. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/673>
* (#599) - Draw all text as paths, to fix text scaled non-proportionally.Federico Mena Quintero2022-03-09176-0/+0
| | | | | | | | | | | | | | | | | | It looks like this won't be fixed in Pango soon, so let's draw all text as paths instead. We create the text-as-path with an identity transform, and then replay the path to the Cairo context with the current transform. I disabled tests/fixtures/text/804-tspan-direction-change.svg in tests/src/text.rs because its spacing is different - need to figure that one out. Also regenerated the test reference files that have text in them; with text-as-path they render slightly differently. Fixes https://gitlab.gnome.org/GNOME/librsvg/-/issues/599 Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/673>
* New test file for text elementsFederico Mena Quintero2022-03-081-0/+11
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/673>
* Revert "(#599) - Draw all text as paths, to fix text scaled non-proportionally."Federico Mena Quintero2022-03-05176-0/+0
| | | | | | This reverts commit fc5579d408001b1063ac342983d0eadf3265c156. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/672>
* Revert "Enable the SVG1.1 tests with text scaled non-proportionally"Federico Mena Quintero2022-03-054-0/+0
| | | | | | This reverts commit 4887a02b7f1b22286c4600349d1abc005657c4e0. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/672>
* Add tests for when ellipse's rx/ry properties default to "auto"Federico Mena Quintero2022-03-044-0/+22
| | | | | | Yay, code coverage! Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/670>
* Enable the SVG1.1 tests with text scaled non-proportionallyFederico Mena Quintero2022-02-044-0/+0
| | | | | | | | | | | These tests: pservers-grad-07-b-ref.png pservers-grad-08-b-ref.png They both have text that has different scaling in the x/y dimensions; they are rendered correctly now. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/660>
* (#599) - Draw all text as paths, to fix text scaled non-proportionally.Federico Mena Quintero2022-02-04176-0/+0
| | | | | | | | | | | | | | | | | | It looks like this won't be fixed in Pango soon, so let's draw all text as paths instead. We create the text-as-path with an identity transform, and then replay the path to the Cairo context with the current transform. I disabled tests/fixtures/text/804-tspan-direction-change.svg in tests/src/text.rs because its spacing is different - need to figure that one out. Also regenerated the test reference files that have text in them; with text-as-path they render slightly differently. Fixes https://gitlab.gnome.org/GNOME/librsvg/-/issues/599 Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/660>
* Remove obsolete tests for 777834-empty-text-children.svgFederico Mena Quintero2022-02-042-78/+0
| | | | | | | | | | | | | | | | 1. This was one of the tests that we started ignoring recently, when Pango reverted its fix for for fonts that are not scaled the same in the x/y dimensions. 2. That test was for the old https://bugzilla.gnome.org/show_bug.cgi?id=777834, where the Rust code for patterns would panic when the C code for text generated an empty bounding box. Since the code for patterns in (2) is fixed since a long time ago, we no longer need this test. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/660>
* New test for the computed bounds of text elements at the API levelFederico Mena Quintero2022-02-041-4/+4
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/660>
* Add a test for the bounds of a text elementFederico Mena Quintero2022-02-042-0/+68
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/660>
* Extract function to compute bounding rectangles from compute_stroke_and_fill_boxFederico Mena Quintero2022-02-041-0/+0
| | | | | | | | | That function now puts the rectangles in a BoundingBox, with their coordinate system. The test filters-composite-02-b.svg seems to have lost a pixel column... Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/660>
* Refresh text-text-03-b-ref.pngFederico Mena Quintero2022-01-211-0/+0
| | | | Did Pango change its strike/underline metrics...?
* Ignore the three tests that have text scaled non-uniformlyFederico Mena Quintero2022-01-213-0/+0
| | | | | | | | | | | | The three files in question are the only ones in the test suite that have text elements scaled non-uniformly in the X/Y directions. Pango had to revert its fix for https://gitlab.gnome.org/GNOME/pango/-/issues/620 (now marked as duplicate of https://gitlab.gnome.org/GNOME/pango/-/issues/520), which came from https://gitlab.gnome.org/GNOME/librsvg/-/issues/599 Summary: this is a Pango bug; unfortunately librsvg can't do anything about it right now.
* Update test reference files for PangoFederico Mena Quintero2022-01-214-0/+0
| | | | | | The fix for https://gitlab.gnome.org/GNOME/pango/-/issues/620 (https://gitlab.gnome.org/GNOME/librsvg/-/issues/599) causes our tests that have scaled fonts to change. Fortunately this really fixes the text elements that are not scaled uniformly.
* Add test cases for display:none and visibility:hiddenMichael Howell2022-01-075-0/+23
| | | | Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/647>
* (#818): Allow having a mask and opacity in the same elementFederico Mena Quintero2021-11-112-0/+10
| | | | | | | | | | The code structure made it so that an element with a `mask` attribute, together with `opacity` and/or `mix-blend-mode`, would only apply the mask. Now they can all be used together. Fixes https://gitlab.gnome.org/GNOME/librsvg/-/issues/818 Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/631>
* pattern: Don't panic when trying to resolve with an empty bounding boxFederico Mena Quintero2021-11-112-0/+30
| | | | | | | This got triggered from DrawingCtx::with_discrete_layer(), where it tries to resolve the paint servers to pass on to the filters code. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/631>
* Handle the isolation property in the drawing machineryFederico Mena Quintero2021-11-112-0/+11
| | | | | | | | This is an easy one; Isolation::Isolate basically means to unconditionally create a temporary surface for compositing, which we already know how to do. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/630>
* Convert the test for #804 into a reference testFederico Mena Quintero2021-11-053-2/+9
| | | | | | | The PNG reference wasn't the same as the result when using Harfbuzz 3.0.0 (the test suite still assumes Harfbuzz 2.9.x). Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/626>
* (#249): Basic implementation of the unicode-bidi propertyFederico Mena Quintero2021-10-252-0/+12
| | | | | | | | | | | | | This commit causes each span's text to be wrapped with the directional control characters specified in https://www.w3.org/TR/css-writing-modes-3/#unicode-bidi For a more complete implementation of unicode-bidi, we'll have to change the code to actually bidi/shape the whole character content of a <text> element and its children, not just each individual span. That will come soon. Part-of: <https://gitlab.gnome.org/GNOME/librsvg/-/merge_requests/621>
View Lorry Controller Status