The extractor was incorrectly setting the value of data-srcset into the
src attribute, however srcset does not use the same format leading to
loading error with incorrect paths. Now data-srcset sets the value of
srcset which will let the browser to use paths of either src or srcset
attributes.
This PR may require new tests to ensure non-regression.
Here is an example with https://www.numerama.com/sciences/661291-covid-19-le-modele-emmental-montre-aussi-pourquoi-la-desinformation-est-si-grave.html:
Coverage increased (+0.01%) to 96.469% when pulling 20d62507b8993de7157ab537ee95b7a6a51ee3b4 on Kdecherf:fix/data-srcset into 1c581bb80076d933d184dbaf540ee63dec74501c on j0k3r:master.
The extractor was incorrectly setting the value of data-srcset into the src attribute, however srcset does not use the same format leading to loading error with incorrect paths. Now data-srcset sets the value of srcset which will let the browser to use paths of either src or srcset attributes.
This PR may require new tests to ensure non-regression.
Here is an example with
https://www.numerama.com/sciences/661291-covid-19-le-modele-emmental-montre-aussi-pourquoi-la-desinformation-est-si-grave.html
:Fixes https://github.com/wallabag/wallabag/issues/4914