view graster/hacklab-engraver/.git/hooks/pre-rebase.sample @ 11:f952052e37b7

trying a fix.
author Robert McIntyre <rlm@mit.edu>
date Tue, 24 Aug 2010 19:06:45 -0400
parents
children
line wrap: on
line source
1 #!/bin/sh
2 #
3 # Copyright (c) 2006, 2008 Junio C Hamano
4 #
5 # The "pre-rebase" hook is run just before "git-rebase" starts doing
6 # its job, and can prevent the command from running by exiting with
7 # non-zero status.
8 #
9 # The hook is called with the following parameters:
10 #
11 # $1 -- the upstream the series was forked from.
12 # $2 -- the branch being rebased (or empty when rebasing the current branch).
13 #
14 # This sample shows how to prevent topic branches that are already
15 # merged to 'next' branch from getting rebased, because allowing it
16 # would result in rebasing already published history.
18 publish=next
19 basebranch="$1"
20 if test "$#" = 2
21 then
22 topic="refs/heads/$2"
23 else
24 topic=`git symbolic-ref HEAD` ||
25 exit 0 ;# we do not interrupt rebasing detached HEAD
26 fi
28 case "$topic" in
29 refs/heads/??/*)
30 ;;
31 *)
32 exit 0 ;# we do not interrupt others.
33 ;;
34 esac
36 # Now we are dealing with a topic branch being rebased
37 # on top of master. Is it OK to rebase it?
39 # Does the topic really exist?
40 git show-ref -q "$topic" || {
41 echo >&2 "No such branch $topic"
42 exit 1
43 }
45 # Is topic fully merged to master?
46 not_in_master=`git-rev-list --pretty=oneline ^master "$topic"`
47 if test -z "$not_in_master"
48 then
49 echo >&2 "$topic is fully merged to master; better remove it."
50 exit 1 ;# we could allow it, but there is no point.
51 fi
53 # Is topic ever merged to next? If so you should not be rebasing it.
54 only_next_1=`git-rev-list ^master "^$topic" ${publish} | sort`
55 only_next_2=`git-rev-list ^master ${publish} | sort`
56 if test "$only_next_1" = "$only_next_2"
57 then
58 not_in_topic=`git-rev-list "^$topic" master`
59 if test -z "$not_in_topic"
60 then
61 echo >&2 "$topic is already up-to-date with master"
62 exit 1 ;# we could allow it, but there is no point.
63 else
64 exit 0
65 fi
66 else
67 not_in_next=`git-rev-list --pretty=oneline ^${publish} "$topic"`
68 perl -e '
69 my $topic = $ARGV[0];
70 my $msg = "* $topic has commits already merged to public branch:\n";
71 my (%not_in_next) = map {
72 /^([0-9a-f]+) /;
73 ($1 => 1);
74 } split(/\n/, $ARGV[1]);
75 for my $elem (map {
76 /^([0-9a-f]+) (.*)$/;
77 [$1 => $2];
78 } split(/\n/, $ARGV[2])) {
79 if (!exists $not_in_next{$elem->[0]}) {
80 if ($msg) {
81 print STDERR $msg;
82 undef $msg;
83 }
84 print STDERR " $elem->[1]\n";
85 }
86 }
87 ' "$topic" "$not_in_next" "$not_in_master"
88 exit 1
89 fi
91 exit 0
93 ################################################################
95 This sample hook safeguards topic branches that have been
96 published from being rewound.
98 The workflow assumed here is:
100 * Once a topic branch forks from "master", "master" is never
101 merged into it again (either directly or indirectly).
103 * Once a topic branch is fully cooked and merged into "master",
104 it is deleted. If you need to build on top of it to correct
105 earlier mistakes, a new topic branch is created by forking at
106 the tip of the "master". This is not strictly necessary, but
107 it makes it easier to keep your history simple.
109 * Whenever you need to test or publish your changes to topic
110 branches, merge them into "next" branch.
112 The script, being an example, hardcodes the publish branch name
113 to be "next", but it is trivial to make it configurable via
114 $GIT_DIR/config mechanism.
116 With this workflow, you would want to know:
118 (1) ... if a topic branch has ever been merged to "next". Young
119 topic branches can have stupid mistakes you would rather
120 clean up before publishing, and things that have not been
121 merged into other branches can be easily rebased without
122 affecting other people. But once it is published, you would
123 not want to rewind it.
125 (2) ... if a topic branch has been fully merged to "master".
126 Then you can delete it. More importantly, you should not
127 build on top of it -- other people may already want to
128 change things related to the topic as patches against your
129 "master", so if you need further changes, it is better to
130 fork the topic (perhaps with the same name) afresh from the
131 tip of "master".
133 Let's look at this example:
135 o---o---o---o---o---o---o---o---o---o "next"
136 / / / /
137 / a---a---b A / /
138 / / / /
139 / / c---c---c---c B /
140 / / / \ /
141 / / / b---b C \ /
142 / / / / \ /
143 ---o---o---o---o---o---o---o---o---o---o---o "master"
146 A, B and C are topic branches.
148 * A has one fix since it was merged up to "next".
150 * B has finished. It has been fully merged up to "master" and "next",
151 and is ready to be deleted.
153 * C has not merged to "next" at all.
155 We would want to allow C to be rebased, refuse A, and encourage
156 B to be deleted.
158 To compute (1):
160 git-rev-list ^master ^topic next
161 git-rev-list ^master next
163 if these match, topic has not merged in next at all.
165 To compute (2):
167 git-rev-list master..topic
169 if this is empty, it is fully merged to "master".