use abbreviated commit hash for revision IDs

[FND]

40 characters seems long and cumbersome. Git and GitHub often shorten these hashes to 7-10 characters in the UI - perhaps we should do the same, hoping that 10 characters are plenty to avoid collisions?

Thoughts, @cdent?

[cdent]

There is, presumably, somewhere a faq or similar on "why it is okay to refer to a commit by prefix of length X" which, when found, will then allow you to reassure yourself that you are getting the uniqueness you need.

However, given that in the tiddlyweb context you never refer a revision id without a tiddler id, your opportunities for collision have dramatically collapsed, so I suspect there's no real issue.

[FND]

There is, presumably, somewhere a faq or similar on "why it is okay to refer to a commit by prefix of length X"

Well, it's quite simple: When referencing a commit, you can use any abbreviation of the full hash ID - Git will expand that if it's unique. However, an abbreviation that was unique at one time is not necessarily unique in the future.

given that in the tiddlyweb context you never refer a revision id without a tiddler id, your opportunities for collision have dramatically collapsed

That's what I figured as well - so I'll go ahead and use 10-char IDs.

[pmario]

Github uses 7chars atm. I think at the very beginning it were less. So with 10 we are probably on the save side. :)

[FND]

done in f986fa28b645d8b01b4acc315a8c73d1235cff0e