Automatically generated summary

[opk12]

Vespucci is sometimes uploading the literal changeset comment Automatically generated summary, like here and here. It is of no practical use and it would be better to either auto-generate a meaningful summary, or force the user to type something.

[opk12]

In particular, in the recent changes list (OsmCha), it is frustrating to see that comment repeated again and again - you must open each single change, just to understand what it does, and this is literally the job of the changeset comment. But a short list of changes in the comment would be extremely helpful in that context.

[simonpoole]

The problem is in most cases there is no way to generate a readable summary in the single available comment tag. That is why there is now always an automatically generated list of edits included in the changeset.

The 'standard comment' is only added when the user left the comment empty as a pointer to the automatically generated list.

Further the user has to explicitly turn off nagging about an empty comment.

[opk12]

Uhm, if an auto-generated comment does not fit the comment field, it means the cs is complex and a human-made comment is essential, no? I've never been on such a hurry that I could reshape areas, but not type a couple words.

Also, in this cs, the auto-generated comment would fit.

What is a use case for an empty comment? The wiki says good practice is small changesets and good comments. Short comments like school's shape and paths for this cs make the difference when reviewing the list of recent changes. Maybe there should not be an option about empty comments.

[simonpoole]

Nobody said that empty comments are a good idea, quite the opposite. And anybody turning the nagging off gets a notice pointing out that not providing a comment is against best practices.

You are barking up the wrong tree.

See also #2392