This is really stupid, but it seems as if Outlook (and possibly other mail clients) either pre-fetch web pages linked from an email, or send the links to Bing for indexing. See this Stackoverflow question:
What this means is someone using Outlook might receive our confirmation mail, but when they click the link, Bing has already hit the link which makes them receive a 404, since the confirmation link is no longer valid. We can purposefully deny Bing any access here, which seems sensible, but the long-term answer is probably to change our logic.
When someone clicks the link, they should be greeted with a web page on which they must click a button to proceed. And that button should be linked with rel="nofollow" or similar, but I'm honestly not sure if that actually prevents web crawlers from clicking that button!
This is really stupid, but it seems as if Outlook (and possibly other mail clients) either pre-fetch web pages linked from an email, or send the links to Bing for indexing. See this Stackoverflow question:
https://stackoverflow.com/questions/41699071/bingpreview-invalidates-one-time-links-in-email
What this means is someone using Outlook might receive our confirmation mail, but when they click the link, Bing has already hit the link which makes them receive a 404, since the confirmation link is no longer valid. We can purposefully deny Bing any access here, which seems sensible, but the long-term answer is probably to change our logic.
When someone clicks the link, they should be greeted with a web page on which they must click a button to proceed. And that button should be linked with rel="nofollow" or similar, but I'm honestly *not sure* if that actually prevents web crawlers from clicking that button!
As we were discussing in reuse/api#10, it would be good if each application could have an optional "name" value which could be displayed on the confirmation page. This way, even if forms is generally FSFE branded, people coming from external pages like PMPC or REUSE could see that they actually sign up for what they came for.
As we were discussing in https://git.fsfe.org/reuse/api/issues/10, it would be good if each application could have an optional "name" value which could be displayed on the confirmation page. This way, even if forms is generally FSFE branded, people coming from external pages like PMPC or REUSE could see that they actually sign up for what they came for.
max.mehl
changed title from Stop Bing (and others) from hitting the URLs, by enforcing a click to Requiring a manual button click before confirmation12 months ago
This is really stupid, but it seems as if Outlook (and possibly other mail clients) either pre-fetch web pages linked from an email, or send the links to Bing for indexing. See this Stackoverflow question:
https://stackoverflow.com/questions/41699071/bingpreview-invalidates-one-time-links-in-email
What this means is someone using Outlook might receive our confirmation mail, but when they click the link, Bing has already hit the link which makes them receive a 404, since the confirmation link is no longer valid. We can purposefully deny Bing any access here, which seems sensible, but the long-term answer is probably to change our logic.
When someone clicks the link, they should be greeted with a web page on which they must click a button to proceed. And that button should be linked with rel="nofollow" or similar, but I'm honestly not sure if that actually prevents web crawlers from clicking that button!
As we were discussing in reuse/api#10, it would be good if each application could have an optional "name" value which could be displayed on the confirmation page. This way, even if forms is generally FSFE branded, people coming from external pages like PMPC or REUSE could see that they actually sign up for what they came for.
Stop Bing (and others) from hitting the URLs, by enforcing a clickto Requiring a manual button click before confirmation 12 months agoWe've received word that Microsoft 365 Outlook servers do the same. So this seems to be a more wide-spread issue. Increasing the prio therefore.
Fixed by #71