This basically means that GIFs made through (or uploaded to) Gfycat or Imgur aren’t actual GIFs, they’re MP4 or WebM videos. In an attempt to improve the GIF format, websites like Gfycat and Imgur rely on an HTML5 video element extension called GIFV. GIFs Are Easy to Make: Why move to a new format when GIFs are so easy to make? GIF-making websites have been around for eons, and most photo editing software can be used to make GIFs.ĭon’t worry things are getting better.Many websites used Adobe’s Flash plug-in for videos, but Flash didn’t work on mobile devices like iPhones. This meant that it was easier to share GIFs than actual videos, so GIFs stuck around. HTML Didn’t Support Video: Before HTML5 launched in 2014, the HTML standard didn’t support video.(In other words, if this animation doesn’t work, then you’re using an out-of-date version of the Edge browser.) All browsers have supported animated GIFs for a long time now. Need a specific example? Mozilla’s APNG format came out in 2008, but the Microsoft Edge browser only began to support the format this year. All Browsers Are Different: Browsers have their quirks, and sometimes one stinkin’ browser can prevent the web from moving forward.Verified code coverage for the changes made. Added relevant unit test for your changes. Addresses an existing issue: Partially addresses #571 `rule-exclusions.ts` during the 4.6.3 update). The `empty-table-header` rule, which was accidentally omitted from The service reported a slightly distinct set of errors (one violation of Microsoft/accessibility-insights-web#6449īesides the PR checklist, I validated that a local build of the scanĬonsistent with results (no violations), where previously Microsoft/accessibility-insights-web#6448 and To that end state without the NPM package layer. This is still an incremental improvement that moves us closer Tie up the fix for the rule inconsistency in service vs web with that * I considered doing a fuller implementation that would involve creatingĪ publishing pipeline for the axe-config package, but I didn't want to * Partially implements (but doesn't complete) #571 Microsoft/accessibility-insights-web#6395 * Disables the unintentionally-enabled `empty-table-header` * Make it more reliable to keep web and service rules in sync. Pages with multiple levels of iframe nesting. The page level timeouts used in `page-timeout-config.ts`, even assuming The new value from the web config for consistency with web and becauseīoth values are still very comfortably within the margin of error set by (`"pingWaitTime": 1000` instead of the default 500ms). This update bundles a copy of theĬurrent output of that package into `axe-configuration.ts` to form theīasis of service's axe configuration, and removes `rule-exclusion.ts`īesides keeping the rules in sync, this also means adds a new axe-coreĬonfig option that was previously overridden by web but not service Package builds an axe config blob based on web's `rule-exclusion.ts` with a new mechanism based on the package in the web repo. To ensure consistency between web and service, this PR replaces Update where the rulesets became inconsistent between web and service. Overrides]( difference in mechanisms led to a mistake during the axe-core 4.6.3 Ruleset (via a combination of [axe-core's rule tags and explicit This is inconsistent with how accessibility-insights-web defines its Mechanism that starts from axe-core's default ruleset an disables rules Today, Service decides which axe-core rules to use via a `RuleExclusion` (UI changes only) Verified usability with NVDA/JAWS.(UI changes only) Added screenshots/GIFs to description above.PR title AND final merge commit title both start with a semantic tag ( fix:, chore:, feat(feature-name):, refactor:).Check coverage report at: /test-results/unit/coverage Verified code coverage for the changes made.Added/updated relevant unit test(s) (and ran yarn test).marginal perf improvement by not running unnecessary rule.Preparation for using the axe-config output in service.This change means that a user relying on axe-config won't include frame-tested, which will avoid the appearance of issues reported by axe-config but not by automated checks. We include this rule in AI4Web's scans, but then post-process the results such that we don't present it to users as an automated checks failure (we instead present it as a special warning bar). Updates the new package to not include an entry for frame-tested.Service already excluded this rule explicitly. Reverted this change, see comment below: Adds an exclusion for the bypass rule, which axe-core changed in 4.2.0 to be a "reviewOnFail" rule (ie, it only ever returns "pass" or "needs review"), but which we were still running anyway and just never reporting any results for (since we don't have needs review content for bypass).
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |