Provide some safety fallback UX when "Updating failed"

[danielbachhuber]

As a Gutenberg user, seeing this error message is quite terrifying:

Specifically, my emotional state is:

"Updating failed"?! Does this mean I'm going to lose all of the content I published?! How do I get back to safely updating?!

It would be ideal to have some secondary UX in this error state ("save your post to your browser" or similar).

The specific issue I ran into is CloudFlare deciding to block my PUT requests.

[aduth]

The specific issue I ran into is CloudFlare deciding to block my PUT requests.

Unrelated to the UX concern you mention, but noting that this specific issue should be resolved as of #4396.

[jorgefilipecosta]

Maybe we can use the new copy all content mechanism and offer the user a link that allows copying all content to the clipboard so they feel safer.

[karmatosed]

One thing discussed before is adding context to errors. If we know what the error is, let's say it here. Design wise we don't have to change the color or nature as it is an error message. What we can do is be a lot more human in our content.

A further addition could be yes adding a 'would you like to copy the content' or just to maybe offer to create a new post. Some way they get to the end result of publishing, without having to go through the entire process again or at least less of the process.

If the message is temporary and the update may succeed again, a simple "would you like to try again", would be great.

To summarise:

• If we know the context lets add the reason why it failed to the error message - let's extend that to any messages.
• If we can offer a fast path to publishing success, let's do that and allow steps to be skipped.

[ghost]

Showing more specific error messages would also be useful for server side validation : #4063

[mtias]

Related #7367.

[paaljoachim]

What is the status of this issue? I can not remember seeing the error in a while.

Greg @gziolo

[gziolo]

@paaljoachim, I don't know this issue.

[paaljoachim]

Thanks for the response Greg!

[paaljoachim]

Thoughts on this issue?
@talldan @tellthemachines @mcsf
(Making my best guess whom I can ping that might know something about the update/save process.)

[talldan]

Not sure that it's much of an issue any longer. The underlying reason for the error seems to have been fixed.

There's now also a process that backs up posts to local browser storage, which offers another recovery option.

To add to the previous comment (#4590 (comment)), offering a Copy All Content option seems like a reasonable thing to do for errors relating to save.

Not sure what else can be done, usually when saving fails there's a good reason that can't be easily solved.

[mcsf]

I'm up for closing this.

[jorgefilipecosta]

Closing this issue as there are no additional action items.