Add discussion section to git-tag documentation.
Signed-off-by: Junio C Hamano <junkio@cox.net>
Junio C Hamano committed
Feb 11, 2007 at 22:29 UTC
4853534e180c237331cb8395fe87ff6e09bbcf21
1 file changed
+131
Documentation/git-tag.txt
+131
@@ -80,6 +80,137 @@ it in the repository configuration as follows:
80
[user]
81
signingkey = <gpg-key-id>
82
83
+
84
+DISCUSSION
85
+----------
86
+
87
+On Re-tagging
88
+~~~~~~~~~~~~~
89
+
90
+What should you do when you tag a wrong commit and you would
91
+want to re-tag?
92
+
93
+If you never pushed anything out, just re-tag it. Use "-f" to
94
+replace the old one. And you're done.
95
+
96
+But if you have pushed things out (or others could just read
97
+your repository directly), then others will have already seen
98
+the old tag. In that case you can do one of two things:
99
+
100
+. The sane thing.
101
+Just admit you screwed up, and use a different name. Others have
102
+already seen one tag-name, and if you keep the same name, you
103
+may be in the situation that two people both have "version X",
104
+but they actually have 'different' "X"'s. So just call it "X.1"
105
+and be done with it.
106
+
107
+. The insane thing.
108
+You really want to call the new version "X" too, 'even though'
109
+others have already seen the old one. So just use "git tag -f"
110
+again, as if you hadn't already published the old one.
111
+
112
+However, Git does *not* (and it should not)change tags behind
113
+users back. So if somebody already got the old tag, doing a "git
114
+pull" on your tree shouldn't just make them overwrite the old
115
+one.
116
+
117
+If somebody got a release tag from you, you cannot just change
118
+the tag for them by updating your own one. This is a big
119
+security issue, in that people MUST be able to trust their
120
+tag-names. If you really want to do the insane thing, you need
121
+to just fess up to it, and tell people that you messed up. You
122
+can do that by making a very public announcement saying:
123
+
124
+------------
125
+Ok, I messed up, and I pushed out an earlier version tagged as X. I
126
+then fixed something, and retagged the *fixed* tree as X again.
127
+
128
+If you got the wrong tag, and want the new one, please delete
129
+the old one and fetch the new one by doing:
130
+
131
+ git tag -d X
132
+ git fetch origin tag X
133
+
134
+to get my updated tag.
135
+
136
+You can test which tag you have by doing
137
+
138
+ git rev-parse X
139
+
140
+which should return 0123456789abcdef.. if you have the new version.
141
+
142
+Sorry for inconvenience.
143
+------------
144
+
145
+Does this seem a bit complicated? It *should* be. There is no
146
+way that it would be correct to just "fix" it behind peoples
147
+backs. People need to know that their tags might have been
148
+changed.
149
+
150
+
151
+On Automatic following
152
+~~~~~~~~~~~~~~~~~~~~~~
153
+
154
+If you are following somebody else's tree, you are most likely
155
+using tracking branches (`refs/heads/origin` in traditional
156
+layout, or `refs/remotes/origin/master` in the separate-remote
157
+layout). You usually want the tags from the other end.
158
+
159
+On the other hand, if you are fetching because you would want a
160
+one-shot merge from somebody else, you typically do not want to
161
+get tags from there. This happens more often for people near
162
+the toplevel but not limited to them. Mere mortals when pulling
163
+from each other do not necessarily want to automatically get
164
+private anchor point tags from the other person.
165
+
166
+You would notice "please pull" messages on the mailing list says
167
+repo URL and branch name alone. This is designed to be easily
168
+cut&pasted to "git fetch" command line:
169
+
170
+------------
171
+Linus, please pull from
172
+
173
+ git://git..../proj.git master
174
+
175
+to get the following updates...
176
+------------
177
+
178
+becomes:
179
+
180
+------------
181
+$ git pull git://git..../proj.git master
182
+------------
183
+
184
+In such a case, you do not want to automatically follow other's
185
+tags.
186
+
187
+One important aspect of git is it is distributed, and being
188
+distributed largely means there is no inherent "upstream" or
189
+"downstream" in the system. On the face of it, the above
190
+example might seem to indicate that the tag namespace is owned
191
+by upper echelon of people and tags only flow downwards, but
192
+that is not the case. It only shows that the usage pattern
193
+determines who are interested in whose tags.
194
+
195
+A one-shot pull is a sign that a commit history is now crossing
196
+the boundary between one circle of people (e.g. "people who are
197
+primarily interested in networking part of the kernel") who may
198
+have their own set of tags (e.g. "this is the third release
199
+candidate from the networking group to be proposed for general
200
+consumption with 2.6.21 release") to another circle of people
201
+(e.g. "people who integrate various subsystem improvements").
202
+The latter are usually not interested in the detailed tags used
203
+internally in the former group (that is what "internal" means).
204
+That is why it is desirable not to follow tags automatically in
205
+this case.
206
+
207
+It may well be that among networking people, they may want to
208
+exchange the tags internal to their group, but in that workflow
209
+they are most likely tracking with each other's progress by
210
+having tracking branches. Again, the heuristic to automatically
211
+follow such tags is a good thing.
212
+
213
+
214
Author
215
------
216
Written by Linus Torvalds <torvalds@osdl.org>,