]>
Commit | Line | Data |
---|---|---|
8c7fa247 LT |
1 | A short git tutorial |
2 | ==================== | |
8c7fa247 LT |
3 | |
4 | Introduction | |
5 | ------------ | |
6 | ||
7 | This is trying to be a short tutorial on setting up and using a git | |
2a29da7c | 8 | repository, mainly because being hands-on and using explicit examples is |
8c7fa247 LT |
9 | often the best way of explaining what is going on. |
10 | ||
11 | In normal life, most people wouldn't use the "core" git programs | |
12 | directly, but rather script around them to make them more palatable. | |
13 | Understanding the core git stuff may help some people get those scripts | |
14 | done, though, and it may also be instructive in helping people | |
15 | understand what it is that the higher-level helper scripts are actually | |
16 | doing. | |
17 | ||
18 | The core git is often called "plumbing", with the prettier user | |
2a29da7c | 19 | interfaces on top of it called "porcelain". You may not want to use the |
f35ca9ed LT |
20 | plumbing directly very often, but it can be good to know what the |
21 | plumbing does for when the porcelain isn't flushing... | |
8c7fa247 LT |
22 | |
23 | ||
2a29da7c JH |
24 | Creating a git repository |
25 | ------------------------- | |
8c7fa247 | 26 | |
2a29da7c | 27 | Creating a new git repository couldn't be easier: all git repositories start |
8c7fa247 LT |
28 | out empty, and the only thing you need to do is find yourself a |
29 | subdirectory that you want to use as a working tree - either an empty | |
30 | one for a totally new project, or an existing working tree that you want | |
31 | to import into git. | |
32 | ||
2a29da7c | 33 | For our first example, we're going to start a totally new repository from |
8db9307c | 34 | scratch, with no pre-existing files, and we'll call it `git-tutorial`. |
8c7fa247 | 35 | To start up, create a subdirectory for it, change into that |
8db9307c | 36 | subdirectory, and initialize the git infrastructure with `git-init-db`: |
8c7fa247 | 37 | |
8db9307c | 38 | ------------------------------------------------ |
f2416c27 JL |
39 | $ mkdir git-tutorial |
40 | $ cd git-tutorial | |
41 | $ git-init-db | |
8db9307c | 42 | ------------------------------------------------ |
8c7fa247 LT |
43 | |
44 | to which git will reply | |
45 | ||
f2416c27 JL |
46 | ---------------- |
47 | defaulting to local storage area | |
48 | ---------------- | |
8c7fa247 | 49 | |
837eedf4 | 50 | which is just git's way of saying that you haven't been doing anything |
914328ac | 51 | strange, and that it will have created a local `.git` directory setup for |
8db9307c JH |
52 | your new project. You will now have a `.git` directory, and you can |
53 | inspect that with `ls`. For your new empty project, it should show you | |
2a29da7c | 54 | three entries, among other things: |
8c7fa247 | 55 | |
2ae6c706 JS |
56 | - a symlink called `HEAD`, pointing to `refs/heads/master` (if your |
57 | platform does not have native symlinks, it is a file containing the | |
58 | line "ref: refs/heads/master") | |
8db9307c JH |
59 | + |
60 | Don't worry about the fact that the file that the `HEAD` link points to | |
914328ac | 61 | doesn't even exist yet -- you haven't created the commit that will |
8db9307c | 62 | start your `HEAD` development branch yet. |
8c7fa247 | 63 | |
8db9307c | 64 | - a subdirectory called `objects`, which will contain all the |
8c7fa247 LT |
65 | objects of your project. You should never have any real reason to |
66 | look at the objects directly, but you might want to know that these | |
8db9307c JH |
67 | objects are what contains all the real 'data' in your repository. |
68 | ||
69 | - a subdirectory called `refs`, which contains references to objects. | |
70 | ||
71 | In particular, the `refs` subdirectory will contain two other | |
72 | subdirectories, named `heads` and `tags` respectively. They do | |
73 | exactly what their names imply: they contain references to any number | |
74 | of different 'heads' of development (aka 'branches'), and to any | |
75 | 'tags' that you have created to name specific versions in your | |
76 | repository. | |
77 | ||
78 | One note: the special `master` head is the default branch, which is | |
79 | why the `.git/HEAD` file was created as a symlink to it even if it | |
80 | doesn't yet exist. Basically, the `HEAD` link is supposed to always | |
81 | point to the branch you are working on right now, and you always | |
82 | start out expecting to work on the `master` branch. | |
83 | ||
84 | However, this is only a convention, and you can name your branches | |
85 | anything you want, and don't have to ever even 'have' a `master` | |
86 | branch. A number of the git tools will assume that `.git/HEAD` is | |
87 | valid, though. | |
88 | ||
89 | [NOTE] | |
914328ac | 90 | An 'object' is identified by its 160-bit SHA1 hash, aka 'object name', |
8db9307c | 91 | and a reference to an object is always the 40-byte hex |
914328ac | 92 | representation of that SHA1 name. The files in the `refs` |
8db9307c | 93 | subdirectory are expected to contain these hex references |
914328ac | 94 | (usually with a final `\'\n\'` at the end), and you should thus |
8db9307c | 95 | expect to see a number of 41-byte files containing these |
914328ac | 96 | references in these `refs` subdirectories when you actually start |
8db9307c | 97 | populating your tree. |
8c7fa247 | 98 | |
a1d4aa74 JH |
99 | [NOTE] |
100 | An advanced user may want to take a look at the | |
101 | link:repository-layout.html[repository layout] document | |
102 | after finishing this tutorial. | |
103 | ||
2a29da7c | 104 | You have now created your first git repository. Of course, since it's |
8c7fa247 LT |
105 | empty, that's not very useful, so let's start populating it with data. |
106 | ||
107 | ||
2a29da7c JH |
108 | Populating a git repository |
109 | --------------------------- | |
8c7fa247 LT |
110 | |
111 | We'll keep this simple and stupid, so we'll start off with populating a | |
112 | few trivial files just to get a feel for it. | |
113 | ||
114 | Start off with just creating any random files that you want to maintain | |
2a29da7c | 115 | in your git repository. We'll start off with a few bad examples, just to |
8c7fa247 LT |
116 | get a feel for how this works: |
117 | ||
8db9307c | 118 | ------------------------------------------------ |
f2416c27 JL |
119 | $ echo "Hello World" >hello |
120 | $ echo "Silly example" >example | |
8db9307c | 121 | ------------------------------------------------ |
8c7fa247 | 122 | |
914328ac | 123 | you have now created two files in your working tree (aka 'working directory'), but to |
8c7fa247 LT |
124 | actually check in your hard work, you will have to go through two steps: |
125 | ||
914328ac | 126 | - fill in the 'index' file (aka 'cache') with the information about your |
2a29da7c | 127 | working tree state. |
8c7fa247 LT |
128 | |
129 | - commit that index file as an object. | |
130 | ||
131 | The first step is trivial: when you want to tell git about any changes | |
215a7ad1 | 132 | to your working tree, you use the `git-update-index` program. That |
8c7fa247 | 133 | program normally just takes a list of filenames you want to update, but |
5f3aa197 | 134 | to avoid trivial mistakes, it refuses to add new entries to the index |
8c7fa247 | 135 | (or remove existing ones) unless you explicitly tell it that you're |
914328ac JH |
136 | adding a new entry with the `\--add` flag (or removing an entry with the |
137 | `\--remove`) flag. | |
8c7fa247 LT |
138 | |
139 | So to populate the index with the two files you just created, you can do | |
140 | ||
8db9307c | 141 | ------------------------------------------------ |
f2416c27 | 142 | $ git-update-index --add hello example |
8db9307c | 143 | ------------------------------------------------ |
8c7fa247 LT |
144 | |
145 | and you have now told git to track those two files. | |
146 | ||
147 | In fact, as you did that, if you now look into your object directory, | |
837eedf4 | 148 | you'll notice that git will have added two new objects to the object |
2a29da7c | 149 | database. If you did exactly the steps above, you should now be able to do |
8c7fa247 | 150 | |
f2416c27 JL |
151 | |
152 | ---------------- | |
153 | $ ls .git/objects/??/* | |
154 | ---------------- | |
8c7fa247 LT |
155 | |
156 | and see two files: | |
157 | ||
f2416c27 JL |
158 | ---------------- |
159 | .git/objects/55/7db03de997c86a4a028e1ebd3a1ceb225be238 | |
160 | .git/objects/f2/4c74a2e500f5ee1332c86b94199f52b1d1d962 | |
161 | ---------------- | |
8c7fa247 | 162 | |
2a29da7c | 163 | which correspond with the objects with names of 557db... and f24c7.. |
8c7fa247 LT |
164 | respectively. |
165 | ||
914328ac | 166 | If you want to, you can use `git-cat-file` to look at those objects, but |
8c7fa247 LT |
167 | you'll have to use the object name, not the filename of the object: |
168 | ||
f2416c27 JL |
169 | ---------------- |
170 | $ git-cat-file -t 557db03de997c86a4a028e1ebd3a1ceb225be238 | |
171 | ---------------- | |
8c7fa247 | 172 | |
914328ac | 173 | where the `-t` tells `git-cat-file` to tell you what the "type" of the |
72e9340c | 174 | object is. git will tell you that you have a "blob" object (ie just a |
8c7fa247 LT |
175 | regular file), and you can see the contents with |
176 | ||
f2416c27 JL |
177 | ---------------- |
178 | $ git-cat-file "blob" 557db03 | |
179 | ---------------- | |
8c7fa247 | 180 | |
2a29da7c | 181 | which will print out "Hello World". The object 557db03 is nothing |
914328ac | 182 | more than the contents of your file `hello`. |
8c7fa247 | 183 | |
8db9307c | 184 | [NOTE] |
914328ac JH |
185 | Don't confuse that object with the file `hello` itself. The |
186 | object is literally just those specific *contents* of the file, and | |
187 | however much you later change the contents in file `hello`, the object | |
188 | we just looked at will never change. Objects are immutable. | |
2a29da7c | 189 | |
8db9307c JH |
190 | [NOTE] |
191 | The second example demonstrates that you can | |
192 | abbreviate the object name to only the first several | |
193 | hexadecimal digits in most places. | |
8c7fa247 LT |
194 | |
195 | Anyway, as we mentioned previously, you normally never actually take a | |
2a29da7c JH |
196 | look at the objects themselves, and typing long 40-character hex |
197 | names is not something you'd normally want to do. The above digression | |
215a7ad1 | 198 | was just to show that `git-update-index` did something magical, and |
2a29da7c JH |
199 | actually saved away the contents of your files into the git object |
200 | database. | |
8c7fa247 | 201 | |
5f3aa197 | 202 | Updating the index did something else too: it created a `.git/index` |
2a29da7c JH |
203 | file. This is the index that describes your current working tree, and |
204 | something you should be very aware of. Again, you normally never worry | |
8c7fa247 LT |
205 | about the index file itself, but you should be aware of the fact that |
206 | you have not actually really "checked in" your files into git so far, | |
914328ac | 207 | you've only *told* git about them. |
8c7fa247 | 208 | |
f35ca9ed | 209 | However, since git knows about them, you can now start using some of the |
8c7fa247 LT |
210 | most basic git commands to manipulate the files or look at their status. |
211 | ||
212 | In particular, let's not even check in the two files into git yet, we'll | |
914328ac | 213 | start off by adding another line to `hello` first: |
8c7fa247 | 214 | |
8db9307c | 215 | ------------------------------------------------ |
f2416c27 | 216 | $ echo "It's a new day for git" >>hello |
8db9307c | 217 | ------------------------------------------------ |
8c7fa247 | 218 | |
914328ac | 219 | and you can now, since you told git about the previous state of `hello`, ask |
8c7fa247 | 220 | git what has changed in the tree compared to your old index, using the |
914328ac | 221 | `git-diff-files` command: |
8c7fa247 | 222 | |
914328ac | 223 | ------------ |
f2416c27 | 224 | $ git-diff-files |
914328ac | 225 | ------------ |
8c7fa247 | 226 | |
2a29da7c | 227 | Oops. That wasn't very readable. It just spit out its own internal |
914328ac | 228 | version of a `diff`, but that internal version really just tells you |
a05ed8cb | 229 | that it has noticed that "hello" has been modified, and that the old object |
8c7fa247 LT |
230 | contents it had have been replaced with something else. |
231 | ||
232 | To make it readable, we can tell git-diff-files to output the | |
914328ac | 233 | differences as a patch, using the `-p` flag: |
8c7fa247 | 234 | |
914328ac | 235 | ------------ |
f2416c27 | 236 | $ git-diff-files -p |
914328ac | 237 | diff --git a/hello b/hello |
2ae6c706 | 238 | index 557db03..263414f 100644 |
914328ac JH |
239 | --- a/hello |
240 | +++ b/hello | |
241 | @@ -1 +1,2 @@ | |
242 | Hello World | |
243 | +It's a new day for git | |
244 | ---- | |
8c7fa247 | 245 | |
914328ac | 246 | i.e. the diff of the change we caused by adding another line to `hello`. |
8c7fa247 | 247 | |
914328ac | 248 | In other words, `git-diff-files` always shows us the difference between |
8c7fa247 LT |
249 | what is recorded in the index, and what is currently in the working |
250 | tree. That's very useful. | |
251 | ||
914328ac JH |
252 | A common shorthand for `git-diff-files -p` is to just write `git |
253 | diff`, which will do the same thing. | |
ed616049 | 254 | |
8c7fa247 | 255 | |
2a29da7c JH |
256 | Committing git state |
257 | -------------------- | |
8c7fa247 LT |
258 | |
259 | Now, we want to go to the next stage in git, which is to take the files | |
260 | that git knows about in the index, and commit them as a real tree. We do | |
914328ac JH |
261 | that in two phases: creating a 'tree' object, and committing that 'tree' |
262 | object as a 'commit' object together with an explanation of what the | |
8c7fa247 LT |
263 | tree was all about, along with information of how we came to that state. |
264 | ||
914328ac | 265 | Creating a tree object is trivial, and is done with `git-write-tree`. |
8c7fa247 LT |
266 | There are no options or other input: git-write-tree will take the |
267 | current index state, and write an object that describes that whole | |
2a29da7c | 268 | index. In other words, we're now tying together all the different |
8c7fa247 LT |
269 | filenames with their contents (and their permissions), and we're |
270 | creating the equivalent of a git "directory" object: | |
271 | ||
8db9307c | 272 | ------------------------------------------------ |
f2416c27 | 273 | $ git-write-tree |
8db9307c | 274 | ------------------------------------------------ |
8c7fa247 LT |
275 | |
276 | and this will just output the name of the resulting tree, in this case | |
ade75a59 | 277 | (if you have done exactly as I've described) it should be |
8c7fa247 | 278 | |
f2416c27 JL |
279 | ---------------- |
280 | 8988da15d077d4829fc51d8544c097def6644dbb | |
281 | ---------------- | |
8c7fa247 LT |
282 | |
283 | which is another incomprehensible object name. Again, if you want to, | |
914328ac | 284 | you can use `git-cat-file -t 8988d\...` to see that this time the object |
8c7fa247 | 285 | is not a "blob" object, but a "tree" object (you can also use |
914328ac | 286 | `git-cat-file` to actually output the raw object contents, but you'll see |
8c7fa247 LT |
287 | mainly a binary mess, so that's less interesting). |
288 | ||
914328ac | 289 | However -- normally you'd never use `git-write-tree` on its own, because |
8c7fa247 | 290 | normally you always commit a tree into a commit object using the |
914328ac JH |
291 | `git-commit-tree` command. In fact, it's easier to not actually use |
292 | `git-write-tree` on its own at all, but to just pass its result in as an | |
293 | argument to `git-commit-tree`. | |
8c7fa247 | 294 | |
914328ac JH |
295 | `git-commit-tree` normally takes several arguments -- it wants to know |
296 | what the 'parent' of a commit was, but since this is the first commit | |
2a29da7c | 297 | ever in this new repository, and it has no parents, we only need to pass in |
914328ac JH |
298 | the object name of the tree. However, `git-commit-tree` |
299 | also wants to get a commit message | |
2a29da7c | 300 | on its standard input, and it will write out the resulting object name for the |
8c7fa247 LT |
301 | commit to its standard output. |
302 | ||
5990efb0 JH |
303 | And this is where we create the `.git/refs/heads/master` file |
304 | which is pointed at by `HEAD`. This file is supposed to contain | |
305 | the reference to the top-of-tree of the master branch, and since | |
306 | that's exactly what `git-commit-tree` spits out, we can do this | |
307 | all with a sequence of simple shell commands: | |
8c7fa247 | 308 | |
8db9307c | 309 | ------------------------------------------------ |
f2416c27 JL |
310 | $ tree=$(git-write-tree) |
311 | $ commit=$(echo 'Initial commit' | git-commit-tree $tree) | |
312 | $ git-update-ref HEAD $commit | |
8db9307c | 313 | ------------------------------------------------ |
8c7fa247 LT |
314 | |
315 | which will say: | |
316 | ||
f2416c27 JL |
317 | ---------------- |
318 | Committing initial tree 8988da15d077d4829fc51d8544c097def6644dbb | |
319 | ---------------- | |
8c7fa247 LT |
320 | |
321 | just to warn you about the fact that it created a totally new commit | |
914328ac | 322 | that is not related to anything else. Normally you do this only *once* |
8c7fa247 LT |
323 | for a project ever, and all later commits will be parented on top of an |
324 | earlier commit, and you'll never see this "Committing initial tree" | |
325 | message ever again. | |
326 | ||
2a29da7c | 327 | Again, normally you'd never actually do this by hand. There is a |
914328ac JH |
328 | helpful script called `git commit` that will do all of this for you. So |
329 | you could have just written `git commit` | |
ed616049 LT |
330 | instead, and it would have done the above magic scripting for you. |
331 | ||
8c7fa247 | 332 | |
2a29da7c JH |
333 | Making a change |
334 | --------------- | |
8c7fa247 | 335 | |
215a7ad1 | 336 | Remember how we did the `git-update-index` on file `hello` and then we |
914328ac | 337 | changed `hello` afterward, and could compare the new state of `hello` with the |
8c7fa247 LT |
338 | state we saved in the index file? |
339 | ||
914328ac JH |
340 | Further, remember how I said that `git-write-tree` writes the contents |
341 | of the *index* file to the tree, and thus what we just committed was in | |
342 | fact the *original* contents of the file `hello`, not the new ones. We did | |
8c7fa247 | 343 | that on purpose, to show the difference between the index state, and the |
2a29da7c | 344 | state in the working tree, and how they don't have to match, even |
8c7fa247 LT |
345 | when we commit things. |
346 | ||
914328ac | 347 | As before, if we do `git-diff-files -p` in our git-tutorial project, |
8c7fa247 | 348 | we'll still see the same difference we saw last time: the index file |
2a29da7c | 349 | hasn't changed by the act of committing anything. However, now that we |
8c7fa247 | 350 | have committed something, we can also learn to use a new command: |
215a7ad1 | 351 | `git-diff-index`. |
8c7fa247 | 352 | |
914328ac | 353 | Unlike `git-diff-files`, which showed the difference between the index |
215a7ad1 | 354 | file and the working tree, `git-diff-index` shows the differences |
914328ac | 355 | between a committed *tree* and either the index file or the working |
215a7ad1 | 356 | tree. In other words, `git-diff-index` wants a tree to be diffed |
a7b20909 LT |
357 | against, and before we did the commit, we couldn't do that, because we |
358 | didn't have anything to diff against. | |
8c7fa247 | 359 | |
914328ac | 360 | But now we can do |
8c7fa247 | 361 | |
f2416c27 JL |
362 | ---------------- |
363 | $ git-diff-index -p HEAD | |
364 | ---------------- | |
8c7fa247 | 365 | |
914328ac | 366 | (where `-p` has the same meaning as it did in `git-diff-files`), and it |
8c7fa247 | 367 | will show us the same difference, but for a totally different reason. |
2a29da7c JH |
368 | Now we're comparing the working tree not against the index file, |
369 | but against the tree we just wrote. It just so happens that those two | |
a7b20909 LT |
370 | are obviously the same, so we get the same result. |
371 | ||
ed616049 LT |
372 | Again, because this is a common operation, you can also just shorthand |
373 | it with | |
374 | ||
f2416c27 JL |
375 | ---------------- |
376 | $ git diff HEAD | |
377 | ---------------- | |
ed616049 LT |
378 | |
379 | which ends up doing the above for you. | |
380 | ||
215a7ad1 | 381 | In other words, `git-diff-index` normally compares a tree against the |
914328ac | 382 | working tree, but when given the `\--cached` flag, it is told to |
a7b20909 | 383 | instead compare against just the index cache contents, and ignore the |
2a29da7c | 384 | current working tree state entirely. Since we just wrote the index |
215a7ad1 | 385 | file to HEAD, doing `git-diff-index \--cached -p HEAD` should thus return |
a7b20909 LT |
386 | an empty set of differences, and that's exactly what it does. |
387 | ||
8db9307c | 388 | [NOTE] |
914328ac | 389 | ================ |
215a7ad1 | 390 | `git-diff-index` really always uses the index for its |
8db9307c JH |
391 | comparisons, and saying that it compares a tree against the working |
392 | tree is thus not strictly accurate. In particular, the list of | |
914328ac JH |
393 | files to compare (the "meta-data") *always* comes from the index file, |
394 | regardless of whether the `\--cached` flag is used or not. The `\--cached` | |
395 | flag really only determines whether the file *contents* to be compared | |
8db9307c | 396 | come from the working tree or not. |
914328ac | 397 | |
8db9307c JH |
398 | This is not hard to understand, as soon as you realize that git simply |
399 | never knows (or cares) about files that it is not told about | |
72e9340c | 400 | explicitly. git will never go *looking* for files to compare, it |
8db9307c JH |
401 | expects you to tell it what the files are, and that's what the index |
402 | is there for. | |
914328ac | 403 | ================ |
8c7fa247 | 404 | |
914328ac | 405 | However, our next step is to commit the *change* we did, and again, to |
837eedf4 | 406 | understand what's going on, keep in mind the difference between "working |
2a29da7c JH |
407 | tree contents", "index file" and "committed tree". We have changes |
408 | in the working tree that we want to commit, and we always have to | |
8c7fa247 LT |
409 | work through the index file, so the first thing we need to do is to |
410 | update the index cache: | |
411 | ||
8db9307c | 412 | ------------------------------------------------ |
f2416c27 | 413 | $ git-update-index hello |
8db9307c | 414 | ------------------------------------------------ |
8c7fa247 | 415 | |
914328ac | 416 | (note how we didn't need the `\--add` flag this time, since git knew |
8c7fa247 LT |
417 | about the file already). |
418 | ||
914328ac JH |
419 | Note what happens to the different `git-diff-\*` versions here. After |
420 | we've updated `hello` in the index, `git-diff-files -p` now shows no | |
215a7ad1 | 421 | differences, but `git-diff-index -p HEAD` still *does* show that the |
2a29da7c | 422 | current state is different from the state we committed. In fact, now |
215a7ad1 | 423 | `git-diff-index` shows the same difference whether we use the `--cached` |
2a29da7c | 424 | flag or not, since now the index is coherent with the working tree. |
8c7fa247 | 425 | |
914328ac | 426 | Now, since we've updated `hello` in the index, we can commit the new |
2a29da7c | 427 | version. We could do it by writing the tree by hand again, and |
914328ac JH |
428 | committing the tree (this time we'd have to use the `-p HEAD` flag to |
429 | tell commit that the HEAD was the *parent* of the new commit, and that | |
ed616049 LT |
430 | this wasn't an initial commit any more), but you've done that once |
431 | already, so let's just use the helpful script this time: | |
8c7fa247 | 432 | |
8db9307c | 433 | ------------------------------------------------ |
f2416c27 | 434 | $ git commit |
8db9307c | 435 | ------------------------------------------------ |
8c7fa247 | 436 | |
ed616049 | 437 | which starts an editor for you to write the commit message and tells you |
2a29da7c | 438 | a bit about what you have done. |
ed616049 | 439 | |
8c7fa247 LT |
440 | Write whatever message you want, and all the lines that start with '#' |
441 | will be pruned out, and the rest will be used as the commit message for | |
442 | the change. If you decide you don't want to commit anything after all at | |
5f3aa197 | 443 | this point (you can continue to edit things and update the index), you |
914328ac | 444 | can just leave an empty message. Otherwise `git commit` will commit |
8c7fa247 LT |
445 | the change for you. |
446 | ||
8c7fa247 | 447 | You've now made your first real git commit. And if you're interested in |
914328ac | 448 | looking at what `git commit` really does, feel free to investigate: |
8c7fa247 | 449 | it's a few very simple shell scripts to generate the helpful (?) commit |
914328ac | 450 | message headers, and a few one-liners that actually do the |
215a7ad1 | 451 | commit itself (`git-commit`). |
8c7fa247 LT |
452 | |
453 | ||
d154dcc8 JL |
454 | Inspecting Changes |
455 | ------------------ | |
8c7fa247 LT |
456 | |
457 | While creating changes is useful, it's even more useful if you can tell | |
2a29da7c | 458 | later what changed. The most useful command for this is another of the |
914328ac | 459 | `diff` family, namely `git-diff-tree`. |
8c7fa247 | 460 | |
914328ac | 461 | `git-diff-tree` can be given two arbitrary trees, and it will tell you the |
8c7fa247 LT |
462 | differences between them. Perhaps even more commonly, though, you can |
463 | give it just a single commit object, and it will figure out the parent | |
464 | of that commit itself, and show the difference directly. Thus, to get | |
465 | the same diff that we've already seen several times, we can now do | |
466 | ||
f2416c27 JL |
467 | ---------------- |
468 | $ git-diff-tree -p HEAD | |
469 | ---------------- | |
8c7fa247 | 470 | |
914328ac JH |
471 | (again, `-p` means to show the difference as a human-readable patch), |
472 | and it will show what the last commit (in `HEAD`) actually changed. | |
8c7fa247 | 473 | |
36f05ef4 JH |
474 | [NOTE] |
475 | ============ | |
476 | Here is an ASCII art by Jon Loeliger that illustrates how | |
477 | various diff-\* commands compare things. | |
478 | ||
479 | diff-tree | |
480 | +----+ | |
481 | | | | |
482 | | | | |
483 | V V | |
484 | +-----------+ | |
485 | | Object DB | | |
486 | | Backing | | |
487 | | Store | | |
488 | +-----------+ | |
489 | ^ ^ | |
490 | | | | |
491 | | | diff-index --cached | |
492 | | | | |
493 | diff-index | V | |
494 | | +-----------+ | |
495 | | | Index | | |
496 | | | "cache" | | |
497 | | +-----------+ | |
498 | | ^ | |
499 | | | | |
500 | | | diff-files | |
501 | | | | |
502 | V V | |
503 | +-----------+ | |
504 | | Working | | |
505 | | Directory | | |
506 | +-----------+ | |
507 | ============ | |
508 | ||
914328ac | 509 | More interestingly, you can also give `git-diff-tree` the `-v` flag, which |
8c7fa247 LT |
510 | tells it to also show the commit message and author and date of the |
511 | commit, and you can tell it to show a whole series of diffs. | |
512 | Alternatively, you can tell it to be "silent", and not show the diffs at | |
513 | all, but just show the actual commit message. | |
514 | ||
914328ac JH |
515 | In fact, together with the `git-rev-list` program (which generates a |
516 | list of revisions), `git-diff-tree` ends up being a veritable fount of | |
517 | changes. A trivial (but very useful) script called `git-whatchanged` is | |
8c7fa247 | 518 | included with git which does exactly this, and shows a log of recent |
2a29da7c | 519 | activities. |
8c7fa247 | 520 | |
81bb573e | 521 | To see the whole history of our pitiful little git-tutorial project, you |
8c7fa247 LT |
522 | can do |
523 | ||
f2416c27 JL |
524 | ---------------- |
525 | $ git log | |
526 | ---------------- | |
81bb573e LT |
527 | |
528 | which shows just the log messages, or if we want to see the log together | |
cc29f732 | 529 | with the associated patches use the more complex (and much more |
81bb573e LT |
530 | powerful) |
531 | ||
f2416c27 JL |
532 | ---------------- |
533 | $ git-whatchanged -p --root | |
534 | ---------------- | |
8c7fa247 | 535 | |
81bb573e LT |
536 | and you will see exactly what has changed in the repository over its |
537 | short history. | |
538 | ||
8db9307c | 539 | [NOTE] |
914328ac JH |
540 | The `\--root` flag is a flag to `git-diff-tree` to tell it to |
541 | show the initial aka 'root' commit too. Normally you'd probably not | |
8db9307c JH |
542 | want to see the initial import diff, but since the tutorial project |
543 | was started from scratch and is so small, we use it to make the result | |
544 | a bit more interesting. | |
8c7fa247 | 545 | |
837eedf4 | 546 | With that, you should now be having some inkling of what git does, and |
8c7fa247 LT |
547 | can explore on your own. |
548 | ||
8db9307c JH |
549 | [NOTE] |
550 | Most likely, you are not directly using the core | |
551 | git Plumbing commands, but using Porcelain like Cogito on top | |
552 | of it. Cogito works a bit differently and you usually do not | |
215a7ad1 | 553 | have to run `git-update-index` yourself for changed files (you |
8db9307c | 554 | do tell underlying git about additions and removals via |
914328ac JH |
555 | `cg-add` and `cg-rm` commands). Just before you make a commit |
556 | with `cg-commit`, Cogito figures out which files you modified, | |
215a7ad1 | 557 | and runs `git-update-index` on them for you. |
3eb5128a JH |
558 | |
559 | ||
2a29da7c JH |
560 | Tagging a version |
561 | ----------------- | |
3eb5128a | 562 | |
2a29da7c | 563 | In git, there are two kinds of tags, a "light" one, and an "annotated tag". |
3eb5128a JH |
564 | |
565 | A "light" tag is technically nothing more than a branch, except we put | |
914328ac | 566 | it in the `.git/refs/tags/` subdirectory instead of calling it a `head`. |
3eb5128a JH |
567 | So the simplest form of tag involves nothing more than |
568 | ||
8db9307c | 569 | ------------------------------------------------ |
f2416c27 | 570 | $ git tag my-first-tag |
8db9307c | 571 | ------------------------------------------------ |
3eb5128a | 572 | |
914328ac | 573 | which just writes the current `HEAD` into the `.git/refs/tags/my-first-tag` |
a7333f9e | 574 | file, after which point you can then use this symbolic name for that |
2a29da7c | 575 | particular state. You can, for example, do |
3eb5128a | 576 | |
f2416c27 JL |
577 | ---------------- |
578 | $ git diff my-first-tag | |
579 | ---------------- | |
3eb5128a JH |
580 | |
581 | to diff your current state against that tag (which at this point will | |
582 | obviously be an empty diff, but if you continue to develop and commit | |
ade75a59 | 583 | stuff, you can use your tag as an "anchor-point" to see what has changed |
3eb5128a JH |
584 | since you tagged it. |
585 | ||
2a29da7c | 586 | An "annotated tag" is actually a real git object, and contains not only a |
3eb5128a | 587 | pointer to the state you want to tag, but also a small tag name and |
914328ac JH |
588 | message, along with optionally a PGP signature that says that yes, |
589 | you really did | |
590 | that tag. You create these annotated tags with either the `-a` or | |
591 | `-s` flag to `git tag`: | |
3eb5128a | 592 | |
f2416c27 JL |
593 | ---------------- |
594 | $ git tag -s <tagname> | |
595 | ---------------- | |
3eb5128a | 596 | |
914328ac | 597 | which will sign the current `HEAD` (but you can also give it another |
3eb5128a | 598 | argument that specifies the thing to tag, ie you could have tagged the |
914328ac | 599 | current `mybranch` point by using `git tag <tagname> mybranch`). |
3eb5128a JH |
600 | |
601 | You normally only do signed tags for major releases or things | |
602 | like that, while the light-weight tags are useful for any marking you | |
914328ac | 603 | want to do -- any time you decide that you want to remember a certain |
3eb5128a JH |
604 | point, just create a private tag for it, and you have a nice symbolic |
605 | name for the state at that point. | |
606 | ||
607 | ||
2a29da7c JH |
608 | Copying repositories |
609 | -------------------- | |
f35ca9ed | 610 | |
f2416c27 JL |
611 | git repositories are normally totally self-sufficient and relocatable |
612 | Unlike CVS, for example, there is no separate notion of | |
914328ac JH |
613 | "repository" and "working tree". A git repository normally *is* the |
614 | working tree, with the local git information hidden in the `.git` | |
2a29da7c | 615 | subdirectory. There is nothing else. What you see is what you got. |
f35ca9ed | 616 | |
8db9307c JH |
617 | [NOTE] |
618 | You can tell git to split the git internal information from | |
619 | the directory that it tracks, but we'll ignore that for now: it's not | |
620 | how normal projects work, and it's really only meant for special uses. | |
621 | So the mental model of "the git information is always tied directly to | |
622 | the working tree that it describes" may not be technically 100% | |
623 | accurate, but it's a good model for all normal use. | |
f35ca9ed LT |
624 | |
625 | This has two implications: | |
626 | ||
2a29da7c | 627 | - if you grow bored with the tutorial repository you created (or you've |
f35ca9ed | 628 | made a mistake and want to start all over), you can just do simple |
f2416c27 JL |
629 | + |
630 | ---------------- | |
631 | $ rm -rf git-tutorial | |
632 | ---------------- | |
8db9307c JH |
633 | + |
634 | and it will be gone. There's no external repository, and there's no | |
635 | history outside the project you created. | |
f35ca9ed | 636 | |
2a29da7c | 637 | - if you want to move or duplicate a git repository, you can do so. There |
914328ac | 638 | is `git clone` command, but if all you want to do is just to |
2a29da7c | 639 | create a copy of your repository (with all the full history that |
e7c1ca42 | 640 | went along with it), you can do so with a regular |
914328ac | 641 | `cp -a git-tutorial new-git-tutorial`. |
8db9307c JH |
642 | + |
643 | Note that when you've moved or copied a git repository, your git index | |
644 | file (which caches various information, notably some of the "stat" | |
645 | information for the files involved) will likely need to be refreshed. | |
914328ac | 646 | So after you do a `cp -a` to create a new copy, you'll want to do |
f2416c27 JL |
647 | + |
648 | ---------------- | |
649 | $ git-update-index --refresh | |
650 | ---------------- | |
8db9307c JH |
651 | + |
652 | in the new repository to make sure that the index file is up-to-date. | |
f35ca9ed | 653 | |
2a29da7c | 654 | Note that the second point is true even across machines. You can |
914328ac JH |
655 | duplicate a remote git repository with *any* regular copy mechanism, be it |
656 | `scp`, `rsync` or `wget`. | |
f35ca9ed LT |
657 | |
658 | When copying a remote repository, you'll want to at a minimum update the | |
2a29da7c | 659 | index cache when you do this, and especially with other peoples' |
f35ca9ed | 660 | repositories you often want to make sure that the index cache is in some |
914328ac | 661 | known state (you don't know *what* they've done and not yet checked in), |
215a7ad1 | 662 | so usually you'll precede the `git-update-index` with a |
f35ca9ed | 663 | |
f2416c27 JL |
664 | ---------------- |
665 | $ git-read-tree --reset HEAD | |
666 | $ git-update-index --refresh | |
667 | ---------------- | |
f35ca9ed | 668 | |
914328ac | 669 | which will force a total index re-build from the tree pointed to by `HEAD`. |
215a7ad1 | 670 | It resets the index contents to `HEAD`, and then the `git-update-index` |
2a29da7c JH |
671 | makes sure to match up all index entries with the checked-out files. |
672 | If the original repository had uncommitted changes in its | |
215a7ad1 | 673 | working tree, `git-update-index --refresh` notices them and |
2a29da7c | 674 | tells you they need to be updated. |
f35ca9ed | 675 | |
ce30a4b6 LT |
676 | The above can also be written as simply |
677 | ||
f2416c27 JL |
678 | ---------------- |
679 | $ git reset | |
680 | ---------------- | |
ce30a4b6 LT |
681 | |
682 | and in fact a lot of the common git command combinations can be scripted | |
c87d41eb JL |
683 | with the `git xyz` interfaces. You can learn things by just looking |
684 | at what the various git scripts do. For example, `git reset` is the | |
685 | above two lines implemented in `git-reset`, but some things like | |
686 | `git status` and `git commit` are slightly more complex scripts around | |
687 | the basic git commands. | |
ce30a4b6 | 688 | |
2a29da7c | 689 | Many (most?) public remote repositories will not contain any of |
914328ac | 690 | the checked out files or even an index file, and will *only* contain the |
2a29da7c | 691 | actual core git files. Such a repository usually doesn't even have the |
8db9307c | 692 | `.git` subdirectory, but has all the git files directly in the |
ce30a4b6 | 693 | repository. |
f35ca9ed LT |
694 | |
695 | To create your own local live copy of such a "raw" git repository, you'd | |
cc29f732 | 696 | first create your own subdirectory for the project, and then copy the |
914328ac | 697 | raw repository contents into the `.git` directory. For example, to |
f35ca9ed LT |
698 | create your own copy of the git repository, you'd do the following |
699 | ||
f2416c27 JL |
700 | ---------------- |
701 | $ mkdir my-git | |
702 | $ cd my-git | |
703 | $ rsync -rL rsync://rsync.kernel.org/pub/scm/git/git.git/ .git | |
704 | ---------------- | |
f35ca9ed LT |
705 | |
706 | followed by | |
707 | ||
f2416c27 JL |
708 | ---------------- |
709 | $ git-read-tree HEAD | |
710 | ---------------- | |
f35ca9ed LT |
711 | |
712 | to populate the index. However, now you have populated the index, and | |
713 | you have all the git internal files, but you will notice that you don't | |
2a29da7c | 714 | actually have any of the working tree files to work on. To get |
f35ca9ed LT |
715 | those, you'd check them out with |
716 | ||
f2416c27 JL |
717 | ---------------- |
718 | $ git-checkout-index -u -a | |
719 | ---------------- | |
f35ca9ed | 720 | |
914328ac | 721 | where the `-u` flag means that you want the checkout to keep the index |
cc29f732 | 722 | up-to-date (so that you don't have to refresh it afterward), and the |
914328ac JH |
723 | `-a` flag means "check out all files" (if you have a stale copy or an |
724 | older version of a checked out tree you may also need to add the `-f` | |
215a7ad1 | 725 | flag first, to tell git-checkout-index to *force* overwriting of any old |
f35ca9ed LT |
726 | files). |
727 | ||
ed616049 LT |
728 | Again, this can all be simplified with |
729 | ||
f2416c27 JL |
730 | ---------------- |
731 | $ git clone rsync://rsync.kernel.org/pub/scm/git/git.git/ my-git | |
732 | $ cd my-git | |
733 | $ git checkout | |
734 | ---------------- | |
ed616049 LT |
735 | |
736 | which will end up doing all of the above for you. | |
737 | ||
cc29f732 | 738 | You have now successfully copied somebody else's (mine) remote |
f35ca9ed LT |
739 | repository, and checked it out. |
740 | ||
ed616049 | 741 | |
2a29da7c JH |
742 | Creating a new branch |
743 | --------------------- | |
ed616049 LT |
744 | |
745 | Branches in git are really nothing more than pointers into the git | |
914328ac JH |
746 | object database from within the `.git/refs/` subdirectory, and as we |
747 | already discussed, the `HEAD` branch is nothing but a symlink to one of | |
ed616049 LT |
748 | these object pointers. |
749 | ||
750 | You can at any time create a new branch by just picking an arbitrary | |
751 | point in the project history, and just writing the SHA1 name of that | |
914328ac | 752 | object into a file under `.git/refs/heads/`. You can use any filename you |
ed616049 | 753 | want (and indeed, subdirectories), but the convention is that the |
914328ac | 754 | "normal" branch is called `master`. That's just a convention, though, |
ed616049 LT |
755 | and nothing enforces it. |
756 | ||
2a29da7c | 757 | To show that as an example, let's go back to the git-tutorial repository we |
a7333f9e LT |
758 | used earlier, and create a branch in it. You do that by simply just |
759 | saying that you want to check out a new branch: | |
ed616049 | 760 | |
8db9307c | 761 | ------------ |
f2416c27 | 762 | $ git checkout -b mybranch |
8db9307c | 763 | ------------ |
ed616049 | 764 | |
914328ac | 765 | will create a new branch based at the current `HEAD` position, and switch |
a7333f9e | 766 | to it. |
ed616049 | 767 | |
8db9307c JH |
768 | [NOTE] |
769 | ================================================ | |
770 | If you make the decision to start your new branch at some | |
914328ac JH |
771 | other point in the history than the current `HEAD`, you can do so by |
772 | just telling `git checkout` what the base of the checkout would be. | |
8db9307c | 773 | In other words, if you have an earlier tag or branch, you'd just do |
ed616049 | 774 | |
2ae6c706 | 775 | ------------ |
f2416c27 | 776 | $ git checkout -b mybranch earlier-commit |
2ae6c706 | 777 | ------------ |
ed616049 | 778 | |
914328ac | 779 | and it would create the new branch `mybranch` at the earlier commit, |
8db9307c JH |
780 | and check out the state at that time. |
781 | ================================================ | |
ed616049 | 782 | |
914328ac | 783 | You can always just jump back to your original `master` branch by doing |
ed616049 | 784 | |
2ae6c706 | 785 | ------------ |
f2416c27 | 786 | $ git checkout master |
2ae6c706 | 787 | ------------ |
ed616049 | 788 | |
a7333f9e LT |
789 | (or any other branch-name, for that matter) and if you forget which |
790 | branch you happen to be on, a simple | |
ed616049 | 791 | |
2ae6c706 | 792 | ------------ |
f2416c27 | 793 | $ ls -l .git/HEAD |
2ae6c706 | 794 | ------------ |
ed616049 | 795 | |
2ae6c706 JS |
796 | will tell you where it's pointing (Note that on platforms with bad or no |
797 | symlink support, you have to execute | |
ed616049 | 798 | |
2ae6c706 | 799 | ------------ |
f2416c27 | 800 | $ cat .git/HEAD |
2ae6c706 JS |
801 | ------------ |
802 | ||
803 | instead). To get the list of branches you have, you can say | |
804 | ||
805 | ------------ | |
f2416c27 | 806 | $ git branch |
2ae6c706 | 807 | ------------ |
2a29da7c | 808 | |
8db9307c | 809 | which is nothing more than a simple script around `ls .git/refs/heads`. |
928f74de | 810 | There will be asterisk in front of the branch you are currently on. |
2a29da7c JH |
811 | |
812 | Sometimes you may wish to create a new branch _without_ actually | |
a7333f9e LT |
813 | checking it out and switching to it. If so, just use the command |
814 | ||
2ae6c706 | 815 | ------------ |
f2416c27 | 816 | $ git branch <branchname> [startingpoint] |
2ae6c706 | 817 | ------------ |
a7333f9e LT |
818 | |
819 | which will simply _create_ the branch, but will not do anything further. | |
914328ac JH |
820 | You can then later -- once you decide that you want to actually develop |
821 | on that branch -- switch to that branch with a regular `git checkout` | |
a7333f9e LT |
822 | with the branchname as the argument. |
823 | ||
ed616049 | 824 | |
2a29da7c JH |
825 | Merging two branches |
826 | -------------------- | |
ed616049 LT |
827 | |
828 | One of the ideas of having a branch is that you do some (possibly | |
829 | experimental) work in it, and eventually merge it back to the main | |
914328ac JH |
830 | branch. So assuming you created the above `mybranch` that started out |
831 | being the same as the original `master` branch, let's make sure we're in | |
ed616049 LT |
832 | that branch, and do some work there. |
833 | ||
914328ac | 834 | ------------------------------------------------ |
f2416c27 JL |
835 | $ git checkout mybranch |
836 | $ echo "Work, work, work" >>hello | |
837 | $ git commit -m 'Some work.' hello | |
914328ac | 838 | ------------------------------------------------ |
ed616049 | 839 | |
914328ac | 840 | Here, we just added another line to `hello`, and we used a shorthand for |
96768e31 | 841 | doing both `git-update-index hello` and `git commit` by just giving the |
914328ac | 842 | filename directly to `git commit`. The `-m` flag is to give the |
2a29da7c | 843 | commit log message from the command line. |
ed616049 LT |
844 | |
845 | Now, to make it a bit more interesting, let's assume that somebody else | |
846 | does some work in the original branch, and simulate that by going back | |
847 | to the master branch, and editing the same file differently there: | |
848 | ||
8db9307c | 849 | ------------ |
f2416c27 | 850 | $ git checkout master |
8db9307c | 851 | ------------ |
ed616049 | 852 | |
914328ac JH |
853 | Here, take a moment to look at the contents of `hello`, and notice how they |
854 | don't contain the work we just did in `mybranch` -- because that work | |
855 | hasn't happened in the `master` branch at all. Then do | |
ed616049 | 856 | |
8db9307c | 857 | ------------ |
f2416c27 JL |
858 | $ echo "Play, play, play" >>hello |
859 | $ echo "Lots of fun" >>example | |
860 | $ git commit -m 'Some fun.' hello example | |
8db9307c | 861 | ------------ |
ed616049 LT |
862 | |
863 | since the master branch is obviously in a much better mood. | |
864 | ||
865 | Now, you've got two branches, and you decide that you want to merge the | |
866 | work done. Before we do that, let's introduce a cool graphical tool that | |
867 | helps you view what's going on: | |
868 | ||
f2416c27 JL |
869 | ---------------- |
870 | $ gitk --all | |
871 | ---------------- | |
ed616049 | 872 | |
914328ac JH |
873 | will show you graphically both of your branches (that's what the `\--all` |
874 | means: normally it will just show you your current `HEAD`) and their | |
2a29da7c | 875 | histories. You can also see exactly how they came to be from a common |
ed616049 LT |
876 | source. |
877 | ||
914328ac JH |
878 | Anyway, let's exit `gitk` (`^Q` or the File menu), and decide that we want |
879 | to merge the work we did on the `mybranch` branch into the `master` | |
880 | branch (which is currently our `HEAD` too). To do that, there's a nice | |
44760f1d | 881 | script called `git merge`, which wants to know which branches you want |
ed616049 LT |
882 | to resolve and what the merge is all about: |
883 | ||
8db9307c | 884 | ------------ |
f2416c27 | 885 | $ git merge "Merge work in mybranch" HEAD mybranch |
8db9307c | 886 | ------------ |
ed616049 | 887 | |
44760f1d | 888 | where the first argument is going to be used as the commit message if |
ed616049 LT |
889 | the merge can be resolved automatically. |
890 | ||
891 | Now, in this case we've intentionally created a situation where the | |
892 | merge will need to be fixed up by hand, though, so git will do as much | |
914328ac JH |
893 | of it as it can automatically (which in this case is just merge the `example` |
894 | file, which had no differences in the `mybranch` branch), and say: | |
ed616049 | 895 | |
f2416c27 | 896 | ---------------- |
44760f1d JH |
897 | Trying really trivial in-index merge... |
898 | fatal: Merge requires file-level merging | |
899 | Nope. | |
900 | ... | |
0501c240 JH |
901 | Auto-merging hello |
902 | CONFLICT (content): Merge conflict in hello | |
44760f1d | 903 | Automatic merge failed/prevented; fix up by hand |
f2416c27 | 904 | ---------------- |
ed616049 LT |
905 | |
906 | which is way too verbose, but it basically tells you that it failed the | |
907 | really trivial merge ("Simple merge") and did an "Automatic merge" | |
914328ac | 908 | instead, but that too failed due to conflicts in `hello`. |
ed616049 | 909 | |
914328ac | 910 | Not to worry. It left the (trivial) conflict in `hello` in the same form you |
ed616049 | 911 | should already be well used to if you've ever used CVS, so let's just |
914328ac JH |
912 | open `hello` in our editor (whatever that may be), and fix it up somehow. |
913 | I'd suggest just making it so that `hello` contains all four lines: | |
ed616049 | 914 | |
8db9307c JH |
915 | ------------ |
916 | Hello World | |
917 | It's a new day for git | |
918 | Play, play, play | |
919 | Work, work, work | |
920 | ------------ | |
ed616049 LT |
921 | |
922 | and once you're happy with your manual merge, just do a | |
923 | ||
8db9307c | 924 | ------------ |
f2416c27 | 925 | $ git commit hello |
8db9307c | 926 | ------------ |
ed616049 LT |
927 | |
928 | which will very loudly warn you that you're now committing a merge | |
929 | (which is correct, so never mind), and you can write a small merge | |
914328ac | 930 | message about your adventures in git-merge-land. |
ed616049 | 931 | |
e1ccf53a | 932 | After you're done, start up `gitk \--all` to see graphically what the |
914328ac | 933 | history looks like. Notice that `mybranch` still exists, and you can |
2a29da7c | 934 | switch to it, and continue to work with it if you want to. The |
914328ac JH |
935 | `mybranch` branch will not contain the merge, but next time you merge it |
936 | from the `master` branch, git will know how you merged it, so you'll not | |
ed616049 LT |
937 | have to do _that_ merge again. |
938 | ||
8db9307c | 939 | Another useful tool, especially if you do not always work in X-Window |
914328ac | 940 | environment, is `git show-branch`. |
2a29da7c JH |
941 | |
942 | ------------------------------------------------ | |
943 | $ git show-branch master mybranch | |
0501c240 | 944 | * [master] Merge work in mybranch |
2a29da7c JH |
945 | ! [mybranch] Some work. |
946 | -- | |
0501c240 | 947 | + [master] Merge work in mybranch |
2a29da7c JH |
948 | ++ [mybranch] Some work. |
949 | ------------------------------------------------ | |
950 | ||
951 | The first two lines indicate that it is showing the two branches | |
952 | and the first line of the commit log message from their | |
914328ac JH |
953 | top-of-the-tree commits, you are currently on `master` branch |
954 | (notice the asterisk `*` character), and the first column for | |
2a29da7c | 955 | the later output lines is used to show commits contained in the |
914328ac | 956 | `master` branch, and the second column for the `mybranch` |
2a29da7c | 957 | branch. Three commits are shown along with their log messages. |
914328ac JH |
958 | All of them have plus `+` characters in the first column, which |
959 | means they are now part of the `master` branch. Only the "Some | |
960 | work" commit has the plus `+` character in the second column, | |
961 | because `mybranch` has not been merged to incorporate these | |
6f60300b JH |
962 | commits from the master branch. The string inside brackets |
963 | before the commit log message is a short name you can use to | |
964 | name the commit. In the above example, 'master' and 'mybranch' | |
965 | are branch heads. 'master~1' is the first parent of 'master' | |
966 | branch head. Please see 'git-rev-parse' documentation if you | |
967 | see more complex cases. | |
2a29da7c JH |
968 | |
969 | Now, let's pretend you are the one who did all the work in | |
914328ac JH |
970 | `mybranch`, and the fruit of your hard work has finally been merged |
971 | to the `master` branch. Let's go back to `mybranch`, and run | |
2a29da7c JH |
972 | resolve to get the "upstream changes" back to your branch. |
973 | ||
2ae6c706 | 974 | ------------ |
f2416c27 JL |
975 | $ git checkout mybranch |
976 | $ git merge "Merge upstream changes." HEAD master | |
2ae6c706 | 977 | ------------ |
2a29da7c JH |
978 | |
979 | This outputs something like this (the actual commit object names | |
980 | would be different) | |
981 | ||
f2416c27 JL |
982 | ---------------- |
983 | Updating from ae3a2da... to a80b4aa.... | |
984 | example | 1 + | |
985 | hello | 1 + | |
986 | 2 files changed, 2 insertions(+), 0 deletions(-) | |
987 | ---------------- | |
2a29da7c JH |
988 | |
989 | Because your branch did not contain anything more than what are | |
914328ac | 990 | already merged into the `master` branch, the resolve operation did |
2a29da7c | 991 | not actually do a merge. Instead, it just updated the top of |
914328ac JH |
992 | the tree of your branch to that of the `master` branch. This is |
993 | often called 'fast forward' merge. | |
ed616049 | 994 | |
e1ccf53a | 995 | You can run `gitk \--all` again to see how the commit ancestry |
914328ac | 996 | looks like, or run `show-branch`, which tells you this. |
2a29da7c JH |
997 | |
998 | ------------------------------------------------ | |
999 | $ git show-branch master mybranch | |
0501c240 JH |
1000 | ! [master] Merge work in mybranch |
1001 | * [mybranch] Merge work in mybranch | |
2a29da7c | 1002 | -- |
0501c240 | 1003 | ++ [master] Merge work in mybranch |
2a29da7c JH |
1004 | ------------------------------------------------ |
1005 | ||
1006 | ||
1007 | Merging external work | |
1008 | --------------------- | |
ed616049 LT |
1009 | |
1010 | It's usually much more common that you merge with somebody else than | |
1011 | merging with your own branches, so it's worth pointing out that git | |
1012 | makes that very easy too, and in fact, it's not that different from | |
067744bd | 1013 | doing a `git merge`. In fact, a remote merge ends up being nothing |
ed616049 | 1014 | more than "fetch the work from a remote repository into a temporary tag" |
067744bd | 1015 | followed by a `git merge`. |
ed616049 | 1016 | |
914328ac JH |
1017 | Fetching from a remote repository is done by, unsurprisingly, |
1018 | `git fetch`: | |
ed616049 | 1019 | |
f2416c27 JL |
1020 | ---------------- |
1021 | $ git fetch <remote-repository> | |
1022 | ---------------- | |
ed616049 | 1023 | |
914328ac JH |
1024 | One of the following transports can be used to name the |
1025 | repository to download from: | |
ed616049 | 1026 | |
914328ac JH |
1027 | Rsync:: |
1028 | `rsync://remote.machine/path/to/repo.git/` | |
1029 | + | |
1030 | Rsync transport is usable for both uploading and downloading, | |
1031 | but is completely unaware of what git does, and can produce | |
1032 | unexpected results when you download from the public repository | |
1033 | while the repository owner is uploading into it via `rsync` | |
1034 | transport. Most notably, it could update the files under | |
1035 | `refs/` which holds the object name of the topmost commits | |
1036 | before uploading the files in `objects/` -- the downloader would | |
1037 | obtain head commit object name while that object itself is still | |
1038 | not available in the repository. For this reason, it is | |
1039 | considered deprecated. | |
1040 | ||
1041 | SSH:: | |
1042 | `remote.machine:/path/to/repo.git/` or | |
1043 | + | |
1044 | `ssh://remote.machine/path/to/repo.git/` | |
1045 | + | |
1046 | This transport can be used for both uploading and downloading, | |
1047 | and requires you to have a log-in privilege over `ssh` to the | |
1048 | remote machine. It finds out the set of objects the other side | |
1049 | lacks by exchanging the head commits both ends have and | |
1050 | transfers (close to) minimum set of objects. It is by far the | |
1051 | most efficient way to exchange git objects between repositories. | |
1052 | ||
1053 | Local directory:: | |
1054 | `/path/to/repo.git/` | |
1055 | + | |
1056 | This transport is the same as SSH transport but uses `sh` to run | |
1057 | both ends on the local machine instead of running other end on | |
1058 | the remote machine via `ssh`. | |
ed616049 | 1059 | |
72e9340c | 1060 | git Native:: |
914328ac JH |
1061 | `git://remote.machine/path/to/repo.git/` |
1062 | + | |
1063 | This transport was designed for anonymous downloading. Like SSH | |
1064 | transport, it finds out the set of objects the downstream side | |
1065 | lacks and transfers (close to) minimum set of objects. | |
ed616049 | 1066 | |
f2416c27 | 1067 | HTTP(S):: |
914328ac JH |
1068 | `http://remote.machine/path/to/repo.git/` |
1069 | + | |
1070 | HTTP and HTTPS transport are used only for downloading. They | |
1071 | first obtain the topmost commit object name from the remote site | |
1072 | by looking at `repo.git/info/refs` file, tries to obtain the | |
1073 | commit object by downloading from `repo.git/objects/xx/xxx\...` | |
1074 | using the object name of that commit object. Then it reads the | |
1075 | commit object to find out its parent commits and the associate | |
1076 | tree object; it repeats this process until it gets all the | |
1077 | necessary objects. Because of this behaviour, they are | |
1078 | sometimes also called 'commit walkers'. | |
1079 | + | |
1080 | The 'commit walkers' are sometimes also called 'dumb | |
72e9340c CM |
1081 | transports', because they do not require any git aware smart |
1082 | server like git Native transport does. Any stock HTTP server | |
914328ac JH |
1083 | would suffice. |
1084 | + | |
215a7ad1 | 1085 | There are (confusingly enough) `git-ssh-fetch` and `git-ssh-upload` |
914328ac | 1086 | programs, which are 'commit walkers'; they outlived their |
72e9340c | 1087 | usefulness when git Native and SSH transports were introduced, |
914328ac | 1088 | and not used by `git pull` or `git push` scripts. |
ed616049 | 1089 | |
914328ac JH |
1090 | Once you fetch from the remote repository, you `resolve` that |
1091 | with your current branch. | |
ade75a59 | 1092 | |
914328ac JH |
1093 | However -- it's such a common thing to `fetch` and then |
1094 | immediately `resolve`, that it's called `git pull`, and you can | |
1095 | simply do | |
ed616049 | 1096 | |
f2416c27 JL |
1097 | ---------------- |
1098 | $ git pull <remote-repository> | |
1099 | ---------------- | |
914328ac JH |
1100 | |
1101 | and optionally give a branch-name for the remote end as a second | |
1102 | argument. | |
ed616049 | 1103 | |
8db9307c JH |
1104 | [NOTE] |
1105 | You could do without using any branches at all, by | |
1106 | keeping as many local repositories as you would like to have | |
914328ac | 1107 | branches, and merging between them with `git pull`, just like |
8db9307c | 1108 | you merge between branches. The advantage of this approach is |
914328ac | 1109 | that it lets you keep set of files for each `branch` checked |
8db9307c JH |
1110 | out and you may find it easier to switch back and forth if you |
1111 | juggle multiple lines of development simultaneously. Of | |
1112 | course, you will pay the price of more disk usage to hold | |
1113 | multiple working trees, but disk space is cheap these days. | |
1114 | ||
1115 | [NOTE] | |
1116 | You could even pull from your own repository by | |
710c97db JH |
1117 | giving '.' as <remote-repository> parameter to `git pull`. This |
1118 | is useful when you want to merge a local branch (or more, if you | |
1119 | are making an Octopus) into the current branch. | |
ed616049 | 1120 | |
0c04094b | 1121 | It is likely that you will be pulling from the same remote |
2a29da7c | 1122 | repository from time to time. As a short hand, you can store |
c9517341 | 1123 | the remote repository URL in a file under .git/remotes/ |
0c04094b JH |
1124 | directory, like this: |
1125 | ||
c9517341 | 1126 | ------------------------------------------------ |
f2416c27 JL |
1127 | $ mkdir -p .git/remotes/ |
1128 | $ cat >.git/remotes/linus <<\EOF | |
c9517341 JH |
1129 | URL: http://www.kernel.org/pub/scm/git/git.git/ |
1130 | EOF | |
1131 | ------------------------------------------------ | |
0c04094b | 1132 | |
914328ac | 1133 | and use the filename to `git pull` instead of the full URL. |
c9517341 | 1134 | The URL specified in such file can even be a prefix |
0c04094b JH |
1135 | of a full URL, like this: |
1136 | ||
c9517341 | 1137 | ------------------------------------------------ |
f2416c27 | 1138 | $ cat >.git/remotes/jgarzik <<\EOF |
c9517341 JH |
1139 | URL: http://www.kernel.org/pub/scm/linux/git/jgarzik/ |
1140 | EOF | |
1141 | ------------------------------------------------ | |
1142 | ||
0c04094b JH |
1143 | |
1144 | Examples. | |
1145 | ||
914328ac JH |
1146 | . `git pull linus` |
1147 | . `git pull linus tag v0.99.1` | |
1148 | . `git pull jgarzik/netdev-2.6.git/ e100` | |
0c04094b JH |
1149 | |
1150 | the above are equivalent to: | |
1151 | ||
914328ac JH |
1152 | . `git pull http://www.kernel.org/pub/scm/git/git.git/ HEAD` |
1153 | . `git pull http://www.kernel.org/pub/scm/git/git.git/ tag v0.99.1` | |
1154 | . `git pull http://www.kernel.org/pub/.../jgarzik/netdev-2.6.git e100` | |
0c04094b | 1155 | |
e7c1ca42 | 1156 | |
44760f1d JH |
1157 | How does the merge work? |
1158 | ------------------------ | |
1159 | ||
1160 | We said this tutorial shows what plumbing does to help you cope | |
1161 | with the porcelain that isn't flushing, but we so far did not | |
1162 | talk about how the merge really works. If you are following | |
1163 | this tutorial the first time, I'd suggest to skip to "Publishing | |
1164 | your work" section and come back here later. | |
1165 | ||
1166 | OK, still with me? To give us an example to look at, let's go | |
1167 | back to the earlier repository with "hello" and "example" file, | |
1168 | and bring ourselves back to the pre-merge state: | |
1169 | ||
1170 | ------------ | |
1171 | $ git show-branch --more=3 master mybranch | |
1172 | ! [master] Merge work in mybranch | |
1173 | * [mybranch] Merge work in mybranch | |
1174 | -- | |
1175 | ++ [master] Merge work in mybranch | |
1176 | ++ [master^2] Some work. | |
1177 | ++ [master^] Some fun. | |
1178 | ------------ | |
1179 | ||
1180 | Remember, before running `git merge`, our `master` head was at | |
1181 | "Some fun." commit, while our `mybranch` head was at "Some | |
1182 | work." commit. | |
1183 | ||
1184 | ------------ | |
1185 | $ git checkout mybranch | |
1186 | $ git reset --hard master^2 | |
1187 | $ git checkout master | |
1188 | $ git reset --hard master^ | |
1189 | ------------ | |
1190 | ||
1191 | After rewinding, the commit structure should look like this: | |
1192 | ||
1193 | ------------ | |
1194 | $ git show-branch | |
1195 | * [master] Some fun. | |
1196 | ! [mybranch] Some work. | |
1197 | -- | |
1198 | + [mybranch] Some work. | |
1199 | + [master] Some fun. | |
1200 | ++ [mybranch^] New day. | |
1201 | ------------ | |
1202 | ||
1203 | Now we are ready to experiment with the merge by hand. | |
1204 | ||
1205 | `git merge` command, when merging two branches, uses 3-way merge | |
1206 | algorithm. First, it finds the common ancestor between them. | |
1207 | The command it uses is `git-merge-base`: | |
1208 | ||
1209 | ------------ | |
1210 | $ mb=$(git-merge-base HEAD mybranch) | |
1211 | ------------ | |
1212 | ||
1213 | The command writes the commit object name of the common ancestor | |
1214 | to the standard output, so we captured its output to a variable, | |
1215 | because we will be using it in the next step. BTW, the common | |
1216 | ancestor commit is the "New day." commit in this case. You can | |
1217 | tell it by: | |
1218 | ||
1219 | ------------ | |
1220 | $ git-name-rev $mb | |
1221 | my-first-tag | |
1222 | ------------ | |
1223 | ||
1224 | After finding out a common ancestor commit, the second step is | |
1225 | this: | |
1226 | ||
1227 | ------------ | |
1228 | $ git-read-tree -m -u $mb HEAD mybranch | |
1229 | ------------ | |
1230 | ||
1231 | This is the same `git-read-tree` command we have already seen, | |
1232 | but it takes three trees, unlike previous examples. This reads | |
1233 | the contents of each tree into different 'stage' in the index | |
1234 | file (the first tree goes to stage 1, the second stage 2, | |
1235 | etc.). After reading three trees into three stages, the paths | |
1236 | that are the same in all three stages are 'collapsed' into stage | |
1237 | 0. Also paths that are the same in two of three stages are | |
1238 | collapsed into stage 0, taking the SHA1 from either stage 2 or | |
1239 | stage 3, whichever is different from stage 1 (i.e. only one side | |
1240 | changed from the common ancestor). | |
1241 | ||
1242 | After 'collapsing' operation, paths that are different in three | |
1243 | trees are left in non-zero stages. At this point, you can | |
1244 | inspect the index file with this command: | |
1245 | ||
1246 | ------------ | |
1247 | $ git-ls-files --stage | |
1248 | 100644 7f8b141b65fdcee47321e399a2598a235a032422 0 example | |
1249 | 100644 263414f423d0e4d70dae8fe53fa34614ff3e2860 1 hello | |
1250 | 100644 06fa6a24256dc7e560efa5687fa84b51f0263c3a 2 hello | |
1251 | 100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello | |
1252 | ------------ | |
1253 | ||
1254 | In our example of only two files, we did not have unchanged | |
1255 | files so only 'example' resulted in collapsing, but in real-life | |
1256 | large projects, only small number of files change in one commit, | |
1257 | and this 'collapsing' tends to trivially merge most of the paths | |
067744bd JH |
1258 | fairly quickly, leaving only a handful the real changes in non-zero |
1259 | stages. | |
44760f1d JH |
1260 | |
1261 | To look at only non-zero stages, use `\--unmerged` flag: | |
1262 | ||
1263 | ------------ | |
1264 | $ git-ls-files --unmerged | |
1265 | 100644 263414f423d0e4d70dae8fe53fa34614ff3e2860 1 hello | |
1266 | 100644 06fa6a24256dc7e560efa5687fa84b51f0263c3a 2 hello | |
1267 | 100644 cc44c73eb783565da5831b4d820c962954019b69 3 hello | |
1268 | ------------ | |
1269 | ||
1270 | The next step of merging is to merge these three versions of the | |
1271 | file, using 3-way merge. This is done by giving | |
1272 | `git-merge-one-file` command as one of the arguments to | |
1273 | `git-merge-index` command: | |
1274 | ||
1275 | ------------ | |
1276 | $ git-merge-index git-merge-one-file hello | |
1277 | Auto-merging hello. | |
1278 | merge: warning: conflicts during merge | |
1279 | ERROR: Merge conflict in hello. | |
1280 | fatal: merge program failed | |
1281 | ------------ | |
1282 | ||
1283 | `git-merge-one-file` script is called with parameters to | |
1284 | describe those three versions, and is responsible to leave the | |
1285 | merge results in the working tree and register it in the index | |
1286 | file. It is a fairly straightforward shell script, and | |
1287 | eventually calls `merge` program from RCS suite to perform the | |
1288 | file-level 3-way merge. In this case, `merge` detects | |
1289 | conflicts, and the merge result with conflict marks is left in | |
1290 | the working tree, while the index file is updated with the | |
1291 | version from the current branch (this is to make `git diff` | |
1292 | useful after this step). This can be seen if you run `ls-files | |
1293 | --stage` again at this point: | |
1294 | ||
1295 | ------------ | |
1296 | $ git-ls-files --stage | |
1297 | 100644 7f8b141b65fdcee47321e399a2598a235a032422 0 example | |
1298 | 100644 06fa6a24256dc7e560efa5687fa84b51f0263c3a 0 hello | |
1299 | ------------ | |
1300 | ||
1301 | As you can see, there is no unmerged paths in the index file. | |
1302 | This is the state of the index file and the working file after | |
1303 | `git merge` returns control back to you, leaving the conflicting | |
1304 | merge for you to resolve. | |
1305 | ||
1306 | ||
2a29da7c JH |
1307 | Publishing your work |
1308 | -------------------- | |
e7c1ca42 | 1309 | |
3eb5128a | 1310 | So we can use somebody else's work from a remote repository; but |
914328ac | 1311 | how can *you* prepare a repository to let other people pull from |
3eb5128a | 1312 | it? |
e7c1ca42 | 1313 | |
2a29da7c | 1314 | Your do your real work in your working tree that has your |
914328ac JH |
1315 | primary repository hanging under it as its `.git` subdirectory. |
1316 | You *could* make that repository accessible remotely and ask | |
3eb5128a | 1317 | people to pull from it, but in practice that is not the way |
2a29da7c | 1318 | things are usually done. A recommended way is to have a public |
3eb5128a | 1319 | repository, make it reachable by other people, and when the |
2a29da7c JH |
1320 | changes you made in your primary working tree are in good shape, |
1321 | update the public repository from it. This is often called | |
914328ac | 1322 | 'pushing'. |
e7c1ca42 | 1323 | |
8db9307c | 1324 | [NOTE] |
914328ac JH |
1325 | This public repository could further be mirrored, and that is |
1326 | how git repositories at `kernel.org` are managed. | |
e7c1ca42 | 1327 | |
3eb5128a JH |
1328 | Publishing the changes from your local (private) repository to |
1329 | your remote (public) repository requires a write privilege on | |
2a29da7c | 1330 | the remote machine. You need to have an SSH account there to |
914328ac | 1331 | run a single command, `git-receive-pack`. |
e7c1ca42 | 1332 | |
3eb5128a | 1333 | First, you need to create an empty repository on the remote |
2a29da7c | 1334 | machine that will house your public repository. This empty |
3eb5128a | 1335 | repository will be populated and be kept up-to-date by pushing |
2a29da7c | 1336 | into it later. Obviously, this repository creation needs to be |
e7c1ca42 JH |
1337 | done only once. |
1338 | ||
8db9307c | 1339 | [NOTE] |
914328ac JH |
1340 | `git push` uses a pair of programs, |
1341 | `git-send-pack` on your local machine, and `git-receive-pack` | |
8db9307c JH |
1342 | on the remote machine. The communication between the two over |
1343 | the network internally uses an SSH connection. | |
3eb5128a | 1344 | |
72e9340c | 1345 | Your private repository's git directory is usually `.git`, but |
3eb5128a | 1346 | your public repository is often named after the project name, |
914328ac JH |
1347 | i.e. `<project>.git`. Let's create such a public repository for |
1348 | project `my-git`. After logging into the remote machine, create | |
3eb5128a | 1349 | an empty directory: |
e7c1ca42 | 1350 | |
2ae6c706 | 1351 | ------------ |
f2416c27 | 1352 | $ mkdir my-git.git |
2ae6c706 | 1353 | ------------ |
e7c1ca42 | 1354 | |
72e9340c | 1355 | Then, make that directory into a git repository by running |
914328ac JH |
1356 | `git init-db`, but this time, since its name is not the usual |
1357 | `.git`, we do things slightly differently: | |
e7c1ca42 | 1358 | |
2ae6c706 | 1359 | ------------ |
f2416c27 | 1360 | $ GIT_DIR=my-git.git git-init-db |
2ae6c706 | 1361 | ------------ |
e7c1ca42 JH |
1362 | |
1363 | Make sure this directory is available for others you want your | |
2a29da7c | 1364 | changes to be pulled by via the transport of your choice. Also |
914328ac JH |
1365 | you need to make sure that you have the `git-receive-pack` |
1366 | program on the `$PATH`. | |
3eb5128a | 1367 | |
8db9307c | 1368 | [NOTE] |
914328ac JH |
1369 | Many installations of sshd do not invoke your shell as the login |
1370 | shell when you directly run programs; what this means is that if | |
1371 | your login shell is `bash`, only `.bashrc` is read and not | |
1372 | `.bash_profile`. As a workaround, make sure `.bashrc` sets up | |
1373 | `$PATH` so that you can run `git-receive-pack` program. | |
3eb5128a | 1374 | |
953e5842 JH |
1375 | [NOTE] |
1376 | If you plan to publish this repository to be accessed over http, | |
1377 | you should do `chmod +x my-git.git/hooks/post-update` at this | |
1378 | point. This makes sure that every time you push into this | |
1379 | repository, `git-update-server-info` is run. | |
1380 | ||
3eb5128a | 1381 | Your "public repository" is now ready to accept your changes. |
2a29da7c | 1382 | Come back to the machine you have your private repository. From |
e7c1ca42 JH |
1383 | there, run this command: |
1384 | ||
2ae6c706 | 1385 | ------------ |
f2416c27 | 1386 | $ git push <public-host>:/path/to/my-git.git master |
2ae6c706 | 1387 | ------------ |
e7c1ca42 JH |
1388 | |
1389 | This synchronizes your public repository to match the named | |
914328ac | 1390 | branch head (i.e. `master` in this case) and objects reachable |
3eb5128a | 1391 | from them in your current repository. |
e7c1ca42 JH |
1392 | |
1393 | As a real example, this is how I update my public git | |
2a29da7c | 1394 | repository. Kernel.org mirror network takes care of the |
3eb5128a | 1395 | propagation to other publicly visible machines: |
e7c1ca42 | 1396 | |
2ae6c706 | 1397 | ------------ |
f2416c27 | 1398 | $ git push master.kernel.org:/pub/scm/git/git.git/ |
2ae6c706 | 1399 | ------------ |
e7c1ca42 JH |
1400 | |
1401 | ||
2a29da7c JH |
1402 | Packing your repository |
1403 | ----------------------- | |
3eb5128a | 1404 | |
914328ac | 1405 | Earlier, we saw that one file under `.git/objects/??/` directory |
2a29da7c | 1406 | is stored for each git object you create. This representation |
914328ac | 1407 | is efficient to create atomically and safely, but |
2a29da7c | 1408 | not so convenient to transport over the network. Since git objects are |
3eb5128a | 1409 | immutable once they are created, there is a way to optimize the |
2a29da7c | 1410 | storage by "packing them together". The command |
3eb5128a | 1411 | |
2ae6c706 | 1412 | ------------ |
f2416c27 | 1413 | $ git repack |
2ae6c706 | 1414 | ------------ |
3eb5128a | 1415 | |
2a29da7c | 1416 | will do it for you. If you followed the tutorial examples, you |
914328ac JH |
1417 | would have accumulated about 17 objects in `.git/objects/??/` |
1418 | directories by now. `git repack` tells you how many objects it | |
1419 | packed, and stores the packed file in `.git/objects/pack` | |
3eb5128a JH |
1420 | directory. |
1421 | ||
8db9307c | 1422 | [NOTE] |
914328ac JH |
1423 | You will see two files, `pack-\*.pack` and `pack-\*.idx`, |
1424 | in `.git/objects/pack` directory. They are closely related to | |
8db9307c JH |
1425 | each other, and if you ever copy them by hand to a different |
1426 | repository for whatever reason, you should make sure you copy | |
1427 | them together. The former holds all the data from the objects | |
1428 | in the pack, and the latter holds the index for random | |
1429 | access. | |
3eb5128a | 1430 | |
914328ac | 1431 | If you are paranoid, running `git-verify-pack` command would |
3eb5128a JH |
1432 | detect if you have a corrupt pack, but do not worry too much. |
1433 | Our programs are always perfect ;-). | |
1434 | ||
1435 | Once you have packed objects, you do not need to leave the | |
1436 | unpacked objects that are contained in the pack file anymore. | |
1437 | ||
2ae6c706 | 1438 | ------------ |
f2416c27 | 1439 | $ git prune-packed |
2ae6c706 | 1440 | ------------ |
3eb5128a JH |
1441 | |
1442 | would remove them for you. | |
1443 | ||
914328ac JH |
1444 | You can try running `find .git/objects -type f` before and after |
1445 | you run `git prune-packed` if you are curious. Also `git | |
1446 | count-objects` would tell you how many unpacked objects are in | |
1447 | your repository and how much space they are consuming. | |
3eb5128a | 1448 | |
8db9307c | 1449 | [NOTE] |
914328ac JH |
1450 | `git pull` is slightly cumbersome for HTTP transport, as a |
1451 | packed repository may contain relatively few objects in a | |
8db9307c JH |
1452 | relatively large pack. If you expect many HTTP pulls from your |
1453 | public repository you might want to repack & prune often, or | |
1454 | never. | |
3eb5128a | 1455 | |
914328ac | 1456 | If you run `git repack` again at this point, it will say |
2a29da7c | 1457 | "Nothing to pack". Once you continue your development and |
914328ac | 1458 | accumulate the changes, running `git repack` again will create a |
3eb5128a | 1459 | new pack, that contains objects created since you packed your |
2a29da7c | 1460 | repository the last time. We recommend that you pack your project |
3eb5128a | 1461 | soon after the initial import (unless you are starting your |
914328ac | 1462 | project from scratch), and then run `git repack` every once in a |
3eb5128a JH |
1463 | while, depending on how active your project is. |
1464 | ||
914328ac | 1465 | When a repository is synchronized via `git push` and `git pull` |
ade75a59 | 1466 | objects packed in the source repository are usually stored |
3eb5128a | 1467 | unpacked in the destination, unless rsync transport is used. |
914328ac JH |
1468 | While this allows you to use different packing strategies on |
1469 | both ends, it also means you may need to repack both | |
1470 | repositories every once in a while. | |
3eb5128a JH |
1471 | |
1472 | ||
2a29da7c JH |
1473 | Working with Others |
1474 | ------------------- | |
3eb5128a | 1475 | |
a232a132 JH |
1476 | Although git is a truly distributed system, it is often |
1477 | convenient to organize your project with an informal hierarchy | |
2a29da7c | 1478 | of developers. Linux kernel development is run this way. There |
a232a132 | 1479 | is a nice illustration (page 17, "Merges to Mainline") in Randy |
914328ac | 1480 | Dunlap's presentation (`http://tinyurl.com/a2jdg`). |
a232a132 | 1481 | |
914328ac | 1482 | It should be stressed that this hierarchy is purely *informal*. |
a232a132 | 1483 | There is nothing fundamental in git that enforces the "chain of |
2a29da7c | 1484 | patch flow" this hierarchy implies. You do not have to pull |
a232a132 JH |
1485 | from only one remote repository. |
1486 | ||
a232a132 | 1487 | A recommended workflow for a "project lead" goes like this: |
3eb5128a | 1488 | |
8db9307c JH |
1489 | 1. Prepare your primary repository on your local machine. Your |
1490 | work is done there. | |
3eb5128a | 1491 | |
8db9307c | 1492 | 2. Prepare a public repository accessible to others. |
914328ac JH |
1493 | + |
1494 | If other people are pulling from your repository over dumb | |
1495 | transport protocols, you need to keep this repository 'dumb | |
1496 | transport friendly'. After `git init-db`, | |
1497 | `$GIT_DIR/hooks/post-update` copied from the standard templates | |
1498 | would contain a call to `git-update-server-info` but the | |
1499 | `post-update` hook itself is disabled by default -- enable it | |
1500 | with `chmod +x post-update`. | |
3eb5128a | 1501 | |
8db9307c JH |
1502 | 3. Push into the public repository from your primary |
1503 | repository. | |
3eb5128a | 1504 | |
914328ac | 1505 | 4. `git repack` the public repository. This establishes a big |
8db9307c | 1506 | pack that contains the initial set of objects as the |
914328ac | 1507 | baseline, and possibly `git prune` if the transport |
8db9307c JH |
1508 | used for pulling from your repository supports packed |
1509 | repositories. | |
3eb5128a | 1510 | |
8db9307c JH |
1511 | 5. Keep working in your primary repository. Your changes |
1512 | include modifications of your own, patches you receive via | |
1513 | e-mails, and merges resulting from pulling the "public" | |
1514 | repositories of your "subsystem maintainers". | |
1515 | + | |
1516 | You can repack this private repository whenever you feel like. | |
3eb5128a | 1517 | |
8db9307c JH |
1518 | 6. Push your changes to the public repository, and announce it |
1519 | to the public. | |
3eb5128a | 1520 | |
8db9307c JH |
1521 | 7. Every once in a while, "git repack" the public repository. |
1522 | Go back to step 5. and continue working. | |
3eb5128a | 1523 | |
a232a132 | 1524 | |
ade75a59 JS |
1525 | A recommended work cycle for a "subsystem maintainer" who works |
1526 | on that project and has an own "public repository" goes like this: | |
3eb5128a | 1527 | |
914328ac | 1528 | 1. Prepare your work repository, by `git clone` the public |
8db9307c | 1529 | repository of the "project lead". The URL used for the |
914328ac | 1530 | initial cloning is stored in `.git/remotes/origin`. |
3eb5128a | 1531 | |
914328ac JH |
1532 | 2. Prepare a public repository accessible to others, just like |
1533 | the "project lead" person does. | |
3eb5128a | 1534 | |
8db9307c | 1535 | 3. Copy over the packed files from "project lead" public |
aa7f412a JH |
1536 | repository to your public repository, unless the "project |
1537 | lead" repository lives on the same machine as yours. In the | |
1538 | latter case, you can use `objects/info/alternates` file to | |
1539 | point at the repository you are borrowing from. | |
3eb5128a | 1540 | |
8db9307c | 1541 | 4. Push into the public repository from your primary |
914328ac JH |
1542 | repository. Run `git repack`, and possibly `git prune` if the |
1543 | transport used for pulling from your repository supports | |
1544 | packed repositories. | |
3eb5128a | 1545 | |
8db9307c JH |
1546 | 5. Keep working in your primary repository. Your changes |
1547 | include modifications of your own, patches you receive via | |
1548 | e-mails, and merges resulting from pulling the "public" | |
1549 | repositories of your "project lead" and possibly your | |
1550 | "sub-subsystem maintainers". | |
1551 | + | |
1552 | You can repack this private repository whenever you feel | |
1553 | like. | |
3eb5128a | 1554 | |
8db9307c JH |
1555 | 6. Push your changes to your public repository, and ask your |
1556 | "project lead" and possibly your "sub-subsystem | |
1557 | maintainers" to pull from it. | |
a232a132 | 1558 | |
914328ac | 1559 | 7. Every once in a while, `git repack` the public repository. |
8db9307c | 1560 | Go back to step 5. and continue working. |
3eb5128a | 1561 | |
a232a132 | 1562 | |
3eb5128a | 1563 | A recommended work cycle for an "individual developer" who does |
2a29da7c | 1564 | not have a "public" repository is somewhat different. It goes |
3eb5128a JH |
1565 | like this: |
1566 | ||
914328ac | 1567 | 1. Prepare your work repository, by `git clone` the public |
8db9307c JH |
1568 | repository of the "project lead" (or a "subsystem |
1569 | maintainer", if you work on a subsystem). The URL used for | |
914328ac | 1570 | the initial cloning is stored in `.git/remotes/origin`. |
3eb5128a | 1571 | |
914328ac | 1572 | 2. Do your work in your repository on 'master' branch. |
3eb5128a | 1573 | |
914328ac | 1574 | 3. Run `git fetch origin` from the public repository of your |
8db9307c | 1575 | upstream every once in a while. This does only the first |
914328ac JH |
1576 | half of `git pull` but does not merge. The head of the |
1577 | public repository is stored in `.git/refs/heads/origin`. | |
3eb5128a | 1578 | |
914328ac JH |
1579 | 4. Use `git cherry origin` to see which ones of your patches |
1580 | were accepted, and/or use `git rebase origin` to port your | |
8db9307c | 1581 | unmerged changes forward to the updated upstream. |
3eb5128a | 1582 | |
914328ac | 1583 | 5. Use `git format-patch origin` to prepare patches for e-mail |
8db9307c | 1584 | submission to your upstream and send it out. Go back to |
914328ac | 1585 | step 2. and continue. |
3eb5128a JH |
1586 | |
1587 | ||
2a29da7c JH |
1588 | Working with Others, Shared Repository Style |
1589 | -------------------------------------------- | |
1590 | ||
1591 | If you are coming from CVS background, the style of cooperation | |
1592 | suggested in the previous section may be new to you. You do not | |
1593 | have to worry. git supports "shared public repository" style of | |
c9517341 JH |
1594 | cooperation you are probably more familiar with as well. |
1595 | ||
1596 | For this, set up a public repository on a machine that is | |
1597 | reachable via SSH by people with "commit privileges". Put the | |
1598 | committers in the same user group and make the repository | |
1599 | writable by that group. | |
1600 | ||
914328ac | 1601 | You, as an individual committer, then: |
c9517341 | 1602 | |
914328ac | 1603 | - First clone the shared repository to a local repository: |
c9517341 JH |
1604 | ------------------------------------------------ |
1605 | $ git clone repo.shared.xz:/pub/scm/project.git/ my-project | |
1606 | $ cd my-project | |
1607 | $ hack away | |
1608 | ------------------------------------------------ | |
1609 | ||
914328ac JH |
1610 | - Merge the work others might have done while you were hacking |
1611 | away: | |
c9517341 JH |
1612 | ------------------------------------------------ |
1613 | $ git pull origin | |
1614 | $ test the merge result | |
1615 | ------------------------------------------------ | |
914328ac JH |
1616 | [NOTE] |
1617 | ================================ | |
1618 | The first `git clone` would have placed the following in | |
1619 | `my-project/.git/remotes/origin` file, and that's why this and | |
1620 | the next step work. | |
1621 | ------------ | |
1622 | URL: repo.shared.xz:/pub/scm/project.git/ my-project | |
1623 | Pull: master:origin | |
1624 | ------------ | |
1625 | ================================ | |
c9517341 | 1626 | |
914328ac JH |
1627 | - push your work as the new head of the shared |
1628 | repository. | |
c9517341 JH |
1629 | ------------------------------------------------ |
1630 | $ git push origin master | |
1631 | ------------------------------------------------ | |
c9517341 | 1632 | If somebody else pushed into the same shared repository while |
914328ac JH |
1633 | you were working locally, `git push` in the last step would |
1634 | complain, telling you that the remote `master` head does not | |
c9517341 JH |
1635 | fast forward. You need to pull and merge those other changes |
1636 | back before you push your work when it happens. | |
2a29da7c JH |
1637 | |
1638 | ||
6f60300b JH |
1639 | Bundling your work together |
1640 | --------------------------- | |
1641 | ||
1642 | It is likely that you will be working on more than one thing at | |
1643 | a time. It is easy to use those more-or-less independent tasks | |
1644 | using branches with git. | |
1645 | ||
1646 | We have already seen how branches work in a previous example, | |
1647 | with "fun and work" example using two branches. The idea is the | |
1648 | same if there are more than two branches. Let's say you started | |
1649 | out from "master" head, and have some new code in the "master" | |
1650 | branch, and two independent fixes in the "commit-fix" and | |
1651 | "diff-fix" branches: | |
1652 | ||
1653 | ------------ | |
1654 | $ git show-branch | |
1655 | ! [commit-fix] Fix commit message normalization. | |
1656 | ! [diff-fix] Fix rename detection. | |
1657 | * [master] Release candidate #1 | |
1658 | --- | |
1659 | + [diff-fix] Fix rename detection. | |
1660 | + [diff-fix~1] Better common substring algorithm. | |
1661 | + [commit-fix] Fix commit message normalization. | |
1662 | + [master] Release candidate #1 | |
1663 | +++ [diff-fix~2] Pretty-print messages. | |
1664 | ------------ | |
1665 | ||
1666 | Both fixes are tested well, and at this point, you want to merge | |
1667 | in both of them. You could merge in 'diff-fix' first and then | |
1668 | 'commit-fix' next, like this: | |
1669 | ||
1670 | ------------ | |
067744bd JH |
1671 | $ git merge 'Merge fix in diff-fix' master diff-fix |
1672 | $ git merge 'Merge fix in commit-fix' master commit-fix | |
6f60300b JH |
1673 | ------------ |
1674 | ||
1675 | Which would result in: | |
1676 | ||
1677 | ------------ | |
1678 | $ git show-branch | |
1679 | ! [commit-fix] Fix commit message normalization. | |
1680 | ! [diff-fix] Fix rename detection. | |
1681 | * [master] Merge fix in commit-fix | |
1682 | --- | |
1683 | + [master] Merge fix in commit-fix | |
1684 | + + [commit-fix] Fix commit message normalization. | |
1685 | + [master~1] Merge fix in diff-fix | |
1686 | ++ [diff-fix] Fix rename detection. | |
1687 | ++ [diff-fix~1] Better common substring algorithm. | |
1688 | + [master~2] Release candidate #1 | |
1689 | +++ [master~3] Pretty-print messages. | |
1690 | ------------ | |
1691 | ||
1692 | However, there is no particular reason to merge in one branch | |
1693 | first and the other next, when what you have are a set of truly | |
1694 | independent changes (if the order mattered, then they are not | |
1695 | independent by definition). You could instead merge those two | |
1696 | branches into the current branch at once. First let's undo what | |
1697 | we just did and start over. We would want to get the master | |
1698 | branch before these two merges by resetting it to 'master~2': | |
1699 | ||
1700 | ------------ | |
1701 | $ git reset --hard master~2 | |
1702 | ------------ | |
1703 | ||
1704 | You can make sure 'git show-branch' matches the state before | |
067744bd JH |
1705 | those two 'git merge' you just did. Then, instead of running |
1706 | two 'git merge' commands in a row, you would pull these two | |
6f60300b JH |
1707 | branch heads (this is known as 'making an Octopus'): |
1708 | ||
1709 | ------------ | |
1710 | $ git pull . commit-fix diff-fix | |
1711 | $ git show-branch | |
1712 | ! [commit-fix] Fix commit message normalization. | |
1713 | ! [diff-fix] Fix rename detection. | |
1714 | * [master] Octopus merge of branches 'diff-fix' and 'commit-fix' | |
1715 | --- | |
1716 | + [master] Octopus merge of branches 'diff-fix' and 'commit-fix' | |
1717 | + + [commit-fix] Fix commit message normalization. | |
1718 | ++ [diff-fix] Fix rename detection. | |
1719 | ++ [diff-fix~1] Better common substring algorithm. | |
1720 | + [master~1] Release candidate #1 | |
1721 | +++ [master~2] Pretty-print messages. | |
1722 | ------------ | |
1723 | ||
1724 | Note that you should not do Octopus because you can. An octopus | |
1725 | is a valid thing to do and often makes it easier to view the | |
1726 | commit history if you are pulling more than two independent | |
1727 | changes at the same time. However, if you have merge conflicts | |
1728 | with any of the branches you are merging in and need to hand | |
1729 | resolve, that is an indication that the development happened in | |
1730 | those branches were not independent after all, and you should | |
1731 | merge two at a time, documenting how you resolved the conflicts, | |
1732 | and the reason why you preferred changes made in one side over | |
1733 | the other. Otherwise it would make the project history harder | |
1734 | to follow, not easier. | |
1735 | ||
3eb5128a | 1736 | [ to be continued.. cvsimports ] |