From: https://www.google.com/accounts/o8/id?id=AItOawmUWmB1M35_jviFvGPYDIH-a-_Al-7OrXM Date: Tue, 19 Jul 2011 16:18:51 +0000 (-0400) Subject: (no commit message) X-Git-Url: http://git.tremily.us/?a=commitdiff_plain;h=758a4c23fd0fc30b89739f0d2be7bf0fe2f81be8;p=ikiwiki.git --- diff --git a/doc/todo/Add_instructive_commit_messages_for_add__47__edit_pages.mdwn b/doc/todo/Add_instructive_commit_messages_for_add__47__edit_pages.mdwn index c7438feb6..571e1a6eb 100644 --- a/doc/todo/Add_instructive_commit_messages_for_add__47__edit_pages.mdwn +++ b/doc/todo/Add_instructive_commit_messages_for_add__47__edit_pages.mdwn @@ -9,10 +9,14 @@ Diff follows. --[[Daniel Andersson]] > itself and it would just be clutter to mention what file was changed, > since any reasonable interface will show the filename, or a link, > or some summary of what files were affected when showing a change. -> + +>> I use the Mercurial backend, and Mercurial doesn't allow empty commit messages, so if there were no message, it would default to "no message given" (hardcoded in `mercurial.pm`), which is also clutter, and non-descriptive at that. But I'm on board with your reasoning. It's a matter of taste (and somewhat backend), I guess. I might continue to locally use this patch (with the caveat below fixed when commit message is given), but I won't push for it to be included upstream. --[[Daniel Andersson]] + > Also your patch stomps over any commit message that the user *does* > provide, so certianly cannot be applied as-is. --[[Joey]] +>> Yes, "naive" was the word :-) . --[[Daniel Andersson]] + [[!tag patch]] ---