Change Details

= Background We would like to show a different type of preview for pages where it is not possible to get extracts. A generic special preview "We cannot show the preview" should be shown when we hit these pages. {F8503553} At the moment, Page Previews processes an extract that it gets from the service (MediaWiki API and RESTBase endpoint). It then tests whether the extract looks empty-ish in client side JavaScript. When the service can indicate that a preview is empty, then these tests should be removed from the codebase/PP should only rely on the service. Related: T168392 = Implementation Possible options: * We should 404 if the extract is empty * We should omit the extracts field * We should provide an empty string = Definition of empty * If an extract field is an empty string and extract_html is also empty * extract_html is empty when rendered has no [[ https://developer.mozilla.org/en-US/docs/Web/API/Node/textContent | textContent ]] = Open questions * Which implementation? * Which service? * ~~ ```Should we move extract processing to the server before tackling this? Or is it a part of this?~~ JR: I think this is out of scope. This is about what to do when the extract is empty.

= Background We would like to show a different type of preview for pages where it is not possible to get extracts. A generic special preview "We cannot show the preview" should be shown when we hit these pages. {F8503553} At the moment, Page Previews processes an extract that it gets from the service (MediaWiki API and RESTBase endpoint). It then tests whether the extract looks empty-ish in client side JavaScript. When the service can indicate that a preview is empty, then these tests should be removed from the codebase/PP should only rely on the service. Related: T168392 = Implementation Possible options: * ~~We should 404 if the extract is empty~~ * ~~We should omit the extracts field~~ * We should provide an empty string = Definition of empty * If an extract field is an empty string and extract_html is also empty * extract_html is empty when rendered has no [[ https://developer.mozilla.org/en-US/docs/Web/API/Node/textContent | textContent ]] = Closed Questions [x] Which implementation? @phuedx: The summary-html route in MCS (T168848). [x] Which service? @phuedx: The Page Summary API (see above and [the spec](https://www.mediawiki.org/wiki/User:Phuedx_(WMF)/Reading/Web/Page_Preview_API). [x] Should we move extract processing to the server before tackling this? Or is it a part of this? @Jdlrobson: I think this is out of scope. This is about what to do when the extract is empty.

= Background We would like to show a different type of preview for pages where it is not possible to get extracts. A generic special preview "We cannot show the preview" should be shown when we hit these pages. {F8503553} At the moment, Page Previews processes an extract that it gets from the service (MediaWiki API and RESTBase endpoint). It then tests whether the extract looks empty-ish in client side JavaScript. When the service can indicate that a preview is empty, then these tests should be removed from the codebase/PP should only rely on the service. Related: T168392 = Implementation Possible options: * ~~We should 404 if the extract is empty~~ * ~~We should omit the extracts field~~ * We should provide an empty string = Definition of empty * If an extract field is an empty string and extract_html is also empty * extract_html is empty when rendered has no [[ https://developer.mozilla.org/en-US/docs/Web/API/Node/textContent | textContent ]] = Open q= Closed Questions *[x] Which implementation? * Which service? * ~~@phuedx: The summary-html route in MCS (T168848). [x] Which service? @phuedx: The Page Summary API (see above and [the spec](https://www.mediawiki.org/wiki/User:Phuedx_(WMF)/Reading/Web/Page_Preview_API). [x] Should we move extract processing to the server before tackling this? Or is it a part of this? ```Should we move extract processing to the server before tackling this? Or is it a part of this?~~ JR:@Jdlrobson: I think this is out of scope. This is about what to do when the extract is empty.