René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"No longer possible to run tests for components which import plotly without a browser","articleBody":"Recently a change was made to _svg_text_utils.js_ which instantiates a DOMParser at the top level of that module: https://github.com/plotly/plotly.js/blob/ce970f112a8fff0d09909529be8a713f5d4f36cc/src/lib/svg_text_utils.js#L22.\r\n\r\nThis has broken a number of tests I have which do not use a browser (or jsdom etc..) because the DOMParser will be instantiated on import, giving the following:\r\n```\r\nvar parser = new DOMParser();\r\n                 ^\r\n\r\nReferenceError: DOMParser is not defined\r\n    at Object.\u003canonymous\u003e (/home/user/repo/node_modules/plotly.js/src/lib/svg_text_utils.js:22:18)\r\n...\r\n```\r\nFor now I've applied a very quick fix:\r\n```\r\nvar parser;\r\nfunction cachingParser() {\r\n  parser = parser === undefined ? new DOMParser(): parser;\r\n  return parser\r\n};\r\n...\r\nvar dom = cachingParser().parseFromString(skeleton, 'application/xml'),\r\n        childNode = dom.documentElement.firstChild;\r\n...\r\n```\r\nCan this be updated to something like the above? If not, then does the DOMParser really need to be at the top level?","author":{"url":"https://github.com/wayofthepie","@type":"Person","name":"wayofthepie"},"datePublished":"2017-05-11T21:45:01.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":3},"url":"https://github.com/1675/plotly.js/issues/1675"}