hi, I'm very happy to see this, I think it will be especially useful to windows users who cant just open a a new terminal and run "bzr diff |less" Even as a linux user, I'd probably alias it into my commit. Let me know if you need any help. One other use case I'd like to see addressed in the future is: if you are halfway through editing your commit message (while reading the diff), you realize you made a trivial mistake somewhere like a typo or excessive white space.. In this case I would like to be able to quit and resume it later. In the olden days you could just keep the editor open make the changes, and the commit picked it up fine, but lately you have to copy your current message, CTRL+C the commit script, run commit again and paste it in the new editor.. Maybe commit can re-use the previous uncommitted message backup (bzr_log._uBiTA~) or maybe just if we end our commit message with DON'T COMMIT YET, it won't commit it, and automatically reuse it next time. I'll think about this a little more. thanks again marius <div>On 7/23/07, Goffredo Baroncelli < <a href="mailto:kreijack@tiscalinet.it">kreijack@tiscalinet.it</a>> wrote:<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Hi All, my last email didn't receive any complaint, so I think that my code goes in the right direction. I addressed most part of the raised concerns: - the switch --diff was renamed in --show-diff. (John) - The code handles the (wrong) case where both --show-diff and -m or --file were passed. (Jhon) - The patch include the addendum to the tutorial. (James) - The encoding of the diff (the most controversial feature) *IS REMOVED*. Now the diff content is wrote as 8-bit data on the file passed to the editor. In any case the diff is not stored in the commit message. So the worst case is that the diff in the editor is not clear. The down side is the code complication. But I think that is an acceptable compromise. (Aaron) - Added a test case (Aaron) </blockquote></div>...