Bug 220614 – [WPE] imported/w3c/web-platform-tests/svg/linking/reftests/url-reference-local-textpath.svg is flaky failing

• Status: RESOLVED
• Resolution: FIXED
• Component: SVG

[Lauro Moura]

imported/w3c/web-platform-tests/svg/linking/reftests/url-reference-local-textpath.svg

Rarely flaky since added. Became much more frequent since first failure in r269568. GTK also with the rare flakiness with a similar diff.

Diff:

--- /home/buildbot/worker/wpe-linux-64-release-tests/build/layout-test-results/imported/w3c/web-platform-tests/svg/linking/reftests/url-reference-local-textpath-expected.txt
+++ /home/buildbot/worker/wpe-linux-64-release-tests/build/layout-test-results/imported/w3c/web-platform-tests/svg/linking/reftests/url-reference-local-textpath-actual.txt
@@ -1,9 +1,9 @@
 layer at (0,0) size 800x600
   RenderView at (0,0) size 800x600
 layer at (0,0) size 800x600
-  RenderSVGRoot {svg} at (0,0) size 72x102
+  RenderSVGRoot {svg} at (0,0) size 100x100
     RenderSVGPath {path} at (0,0) size 0x0 [fill={[type=SOLID] [color=#000000]}] [data="M 0 80 L 100 80"]
-    RenderSVGText {text} at (0,-10) size 72x112 contains 1 chunk(s)
-      RenderSVGTextPath {textPath} at (0,0) size 72x111
-        RenderSVGInlineText {#text} at (0,0) size 72x111
-          chunk 1 text run 1 at (36.00,80.00) startOffset 0 endOffset 1 width 72.00: "X"
+    RenderSVGText {text} at (0,0) size 100x100 contains 1 chunk(s)
+      RenderSVGTextPath {textPath} at (0,0) size 100x100
+        RenderSVGInlineText {#text} at (0,0) size 100x100
+          chunk 1 text run 1 at (50.00,80.00) startOffset 0 endOffset 1 width 100.00: "X"

Could not reproduce it locally.

Checking with the browser, the reference 100x100 green square is correctly shown, both GTK and WPE. The mac expectation also has RenderSVGRoot dump as 100x100.

Shouldn't WTR show the RenderSVGRoot with 100x100 instead of 72x102? This looks like some discrepancy between GTK/WPE WTR and the actual browsers (including MiniBrowser).

[Diego Pino]

The test(s) filed under this bug ticket have consistently passed for the last 4000 revisions. Currently, there are no more tests in TestExpectations that refer to this bug. Therefore, I'm closing the bug.