X-Git-Url: http://git.vanrenterghem.biz/git.ikiwiki.info.git/blobdiff_plain/49043e48fad367b4cb0fd23e6250a915f095bf1d..a605ee2c562323dc29852787c48a28bd03706718:/doc/todo/git_attribution/discussion.mdwn

diff --git a/doc/todo/git_attribution/discussion.mdwn b/doc/todo/git_attribution/discussion.mdwn
index 09b0d3c51..dfb490bc2 100644
--- a/doc/todo/git_attribution/discussion.mdwn
+++ b/doc/todo/git_attribution/discussion.mdwn
@@ -10,10 +10,13 @@ I've been thinking a bit about the GIT attribution in ikiwiki...
 
 If no email set, I think "$USERNAME" is reasonable... no point in the
 '<>' causing clutter.
+>> **adjustement wrt comments**: leave the '<>' in due to requirements in git
 
 If no username set... then something like '@[IPADDR]' makes sense...
 (not in email brackets).
 
+> Why not put it in email brackets? --[[Joey]]
+
 In the case of OpenID login.. I think that's a special case...  I don't
 think attempting to munge something meaningful out of the OpenID makes
 sense... but I think some massaging might need to be done.
@@ -42,7 +45,54 @@ Perhaps an OpenID: $OPENID_URL would make sense.  This could help w/
 making sure that no one irrefutably spoofs a post by someone (since w/
 the setup where email and effective username are configurable, there's
 no determination of uniqueness)
+>> **adj re git req**: "$OPENID_URL <>"
 
 [[harningt]]
 
-[[madduck]]: git requires "Name <Email@address>" format, as far as I know.
+[[madduck]]: git requires `Name <Email@address>` format, as far as I know.
+
+> Yes, it does:
+> 
+>     joey@kodama:~/tmp/foo/bar>git commit --author "foo"
+>     fatal: malformed --author parameter
+> 
+> It seems to be happy with anything of the form "foo <foo>" -- doesn't seem to
+> do any kind of strict checking. Even "http://joey.kitenet.net <>" will be
+> accepted. --[[Joey]]
+>>
+>>Sounds good to me, 
+>>
+>> --[[harningt]]
+
+> I think the thing to do is, as Josh suggested originally, use
+> GIT_AUTHOR_NAME and GIT_AUTHOR_EMAIL. Note that setting these
+> individually is best, so git can independently validate/sanitize both
+> (which it does do somewhat). Always put the username/openid/IP in
+> GIT_AUTHOR_NAME; if the user has configured an email address,
+> GIT_AUTHOR_EMAIL can also be set.
+> 
+> There is one thing yet to be solved, and that is how to tell the
+> difference between a web commit by 'Joey Hess <joey@kitenet.net>',
+> and a git commit by the same. I think we do want to differentiate these,
+> and the best way to do it seems to be to add a line to the end of the
+> commit message. Something like: "\n\nWeb-commit: true"
+> 
+> For backwards compatability, the code that parses the current stuff needs
+> to be left in. But it will need to take care to only parse that if the
+> commit isn't flagged as a web commit! Else web committers could forge
+> commits from others. --[[Joey]]
+> 
+> BTW, I decided not to use the user's email address in the commit, because
+> then the email becomes part of project history, and you don't really
+> expect that to happen when you give your email address on signup to a web
+> site.
+> 
+> The problem with leaving the email empty is that it confuses some things
+> that try to parse it, including: 
+> * cia (wants a username in there):
+> * git pull --rebase (?)
+> * github pushes to twitter ;-)
+> 
+> So while I tried that way at first, I'm now leaning toward encoding the
+> username in the email address. Like "user <user@web>", or
+> "joey <http://joey.kitenet.net/@web>".