Basic Work Cycle

Basic Work Cycle

Subversion has numerous features, options, bells, and whistles, but on a day-to-day basis, odds are that you will use only a few of them. In this section, we'll run through the most common things that you might find yourself doing with Subversion in the course of a day's work.

The typical work cycle looks like this:

  1. Update your working copy.

    • svn update

  2. Make changes.

    • svn add

    • svn delete

    • svn copy

    • svn move

  3. Examine your changes.

    • svn status

    • svn diff

  4. Possibly undo some changes.

    • svn revert

  5. Resolve conflicts (merge others' changes).

    • svn update

    • svn resolve

  6. Commit your changes.

    • svn commit

Update Your Working Copy

When working on a project with a team, you'll want to update your working copy to receive any changes other developers on the project have made since your last update. Use svn update to bring your working copy into sync with the latest revision in the repository:

$ svn update
U  foo.c
U  bar.c
Updated to revision 2.

In this case, it appears that someone checked in modifications to both foo.c and bar.c since the last time you updated, and Subversion has updated your working copy to include those changes.

When the server sends changes to your working copy via svn update, a letter code is displayed next to each item to let you know what actions Subversion performed to bring your working copy up to date. To find out what these letters mean, run svn help update.

Make Changes to Your Working Copy

Now you can get to work and make changes in your working copy. It's usually most convenient to decide on a discrete change (or set of changes) to make, such as writing a new feature, fixing a bug, and so on. The Subversion commands that you will use here are svn add, svn delete, svn copy, svn move, and svn mkdir. However, if you are merely editing files that are already in Subversion, you may not need to use any of these commands until you commit.

You can make two kinds of changes to your working copy: file changes and tree changes. You don't need to tell Subversion that you intend to change a file; just make your changes using your text editor, word processor, graphics program, or whatever tool you would normally use. Subversion automatically detects which files have been changed, and in addition, it handles binary files just as easily as it handles text files—and just as efficiently, too. For tree changes, you can ask Subversion to mark files and directories for scheduled removal, addition, copying, or moving. These changes may take place immediately in your working copy, but no additions or removals will happen in the repository until you commit them.

Here is an overview of the five Subversion subcommands that you'll use most often to make tree changes:

svn add foo

Schedule file, directory, or symbolic link foo to be added to the repository. When you next commit, foo will become a child of its parent directory. Note that if foo is a directory, everything underneath foo will be scheduled for addition. If you want only to add foo itself, pass the --depth empty option.

svn delete foo

Schedule file, directory, or symbolic link foo to be deleted from the repository. If foo is a file or link, it is immediately deleted from your working copy. If foo is a directory, it is not deleted, but Subversion schedules it for deletion. When you commit your changes, foo will be entirely removed from your working copy and the repository. [3]

svn copy foo bar

Create a new item bar as a duplicate of foo and automatically schedule bar for addition. When bar is added to the repository on the next commit, its copy history is recorded (as having originally come from foo). svn copy does not create intermediate directories unless you pass the --parents option.

svn move foo bar

This command is exactly the same as running svn copy foo bar; svn delete foo. That is, bar is scheduled for addition as a copy of foo, and foo is scheduled for removal. svn move does not create intermediate directories unless you pass the --parents option.

svn mkdir blort

This command is exactly the same as running mkdir blort; svn add blort. That is, a new directory named blort is created and scheduled for addition.

Examine Your Changes

Once you've finished making changes, you need to commit them to the repository, but before you do so, it's usually a good idea to take a look at exactly what you've changed. By examining your changes before you commit, you can make a more accurate log message. You may also discover that you've inadvertently changed a file, and this gives you a chance to revert those changes before committing. Additionally, this is a good opportunity to review and scrutinize changes before publishing them. You can see an overview of the changes you've made by using svn status, and dig into the details of those changes by using svn diff.

Subversion has been optimized to help you with this task, and it is able to do many things without communicating with the repository. In particular, your working copy contains a hidden cached pristine copy of each version-controlled file within the .svn area. Because of this, Subversion can quickly show you how your working files have changed or even allow you to undo your changes without contacting the repository.

See an overview of your changes

To get an overview of your changes, you'll use the svn status command. You'll probably use svn status more than any other Subversion command.

If you run svn status at the top of your working copy with no arguments, it will detect all file and tree changes you've made. Here are a few examples of the most common status codes that svn status can return. (Note that the text following # is not actually printed by svn status.)

?       scratch.c           # file is not under version control
A       stuff/loot/bloo.h   # file is scheduled for addition
C       stuff/loot/lump.c   # file has textual conflicts from an update
D       stuff/fish.c        # file is scheduled for deletion
M       bar.c               # the content in bar.c has local modifications

In this output format, svn status prints six columns of characters, followed by several whitespace characters, followed by a file or directory name. The first column tells the status of a file or directory and/or its contents. The codes we listed are:

A item

The file, directory, or symbolic link item has been scheduled for addition into the repository.

C item

The file item is in a state of conflict. That is, changes received from the server during an update overlap with local changes that you have in your working copy (and weren't resolved during the update). You must resolve this conflict before committing your changes to the repository.

D item

The file, directory, or symbolic link item has been scheduled for deletion from the repository.

M item

The contents of the file item have been modified.

If you pass a specific path to svn status, you get information about that item alone:

$ svn status stuff/fish.c
D      stuff/fish.c

svn status also has a --verbose (-v) option, which will show you the status of every item in your working copy, even if it has not been changed:

$ svn status -v
M               44        23    sally     README
                44        30    sally     INSTALL
M               44        20    harry     bar.c
                44        18    ira       stuff
                44        35    harry     stuff/trout.c
D               44        19    ira       stuff/fish.c
                44        21    sally     stuff/things
A                0         ?     ?        stuff/things/bloo.h
                44        36    harry     stuff/things/gloo.c

This is the long form output of svn status. The letters in the first column mean the same as before, but the second column shows the working revision of the item. The third and fourth columns show the revision in which the item last changed, and who changed it.

None of the prior invocations to svn status contact the repository—instead, they compare the metadata in the .svn directory with the working copy. Finally, there is the --show-updates (-u) option, which contacts the repository and adds information about things that are out of date:

$ svn status -u -v
M      *        44        23    sally     README
M               44        20    harry     bar.c
       *        44        35    harry     stuff/trout.c
D               44        19    ira       stuff/fish.c
A                0         ?     ?        stuff/things/bloo.h
Status against revision:   46

Notice the two asterisks: if you were to run svn update at this point, you would receive changes to README and trout.c. This tells you some very useful information—you'll need to update and get the server changes on README before you commit, or the repository will reject your commit for being out of date (more on this subject later).

svn status can display much more information about the files and directories in your working copy than we've shown here—for an exhaustive description of svn status and its output, see svn status.

Examine the details of your local modifications

Another way to examine your changes is with the svn diff command. You can find out exactly how you've modified things by running svn diff with no arguments, which prints out file changes in unified diff format:

$ svn diff
Index: bar.c
===================================================================
--- bar.c	(revision 3)
+++ bar.c	(working copy)
@@ -1,7 +1,12 @@
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <unistd.h>
+
+#include <stdio.h>

 int main(void) {
-  printf("Sixty-four slices of American Cheese...\n");
+  printf("Sixty-five slices of American Cheese...\n");
 return 0;
 }

Index: README
===================================================================
--- README	(revision 3)
+++ README	(working copy)
@@ -193,3 +193,4 @@
+Note to self:  pick up laundry.

Index: stuff/fish.c
===================================================================
--- stuff/fish.c	(revision 1)
+++ stuff/fish.c	(working copy)
-Welcome to the file known as 'fish'.
-Information on fish will be here soon.

Index: stuff/things/bloo.h
===================================================================
--- stuff/things/bloo.h	(revision 8)
+++ stuff/things/bloo.h	(working copy)
+Here is a new file to describe
+things about bloo.

The svn diff command produces this output by comparing your working files against the cached pristine copies within the .svn area. Files scheduled for addition are displayed as all added text, and files scheduled for deletion are displayed as all deleted text.

Output is displayed in unified diff format. That is, removed lines are prefaced with -, and added lines are prefaced with +. svn diff also prints filename and offset information useful to the patch program, so you can generate patches by redirecting the diff output to a file:

$ svn diff > patchfile

You could, for example, email the patch file to another developer for review or testing prior to a commit.

Subversion uses its internal diff engine, which produces unified diff format, by default. If you want diff output in a different format, specify an external diff program using --diff-cmd and pass any flags you'd like to it using the --extensions (-x) option. For example, to see local differences in file foo.c in context output format while ignoring case differences, you might run svn diff --diff-cmd /usr/bin/diff -x "-i" foo.c.

Undoing Working Changes

Suppose while viewing the output of svn diff you determine that all the changes you made to a particular file are mistakes. Maybe you shouldn't have changed the file at all, or perhaps it would be easier to make different changes starting from scratch.

This is a perfect opportunity to use svn revert:

$ svn revert README
Reverted 'README'

Subversion reverts the file to its premodified state by overwriting it with the cached pristine copy from the .svn area. But also note that svn revert can undo any scheduled operations—for example, you might decide that you don't want to add a new file after all:

$ svn status foo
?      foo

$ svn add foo
A         foo

$ svn revert foo
Reverted 'foo'

$ svn status foo
?      foo
[Note] Note

svn revert item has exactly the same effect as deleting item from your working copy and then running svn update -r BASE item. However, if you're reverting a file, svn revert has one very noticeable difference—it doesn't have to communicate with the repository to restore your file.

Or perhaps you mistakenly removed a file from version control:

$ svn status README

$ svn delete README
D         README

$ svn revert README
Reverted 'README'

$ svn status README

Resolve Conflicts (Merging Others' Changes)

We've already seen how svn status -u can predict conflicts. Suppose you run svn update and some interesting things occur:

$ svn update
U  INSTALL
G  README
Conflict discovered in 'bar.c'.
Select: (p) postpone, (df) diff-full, (e) edit,
        (h) help for more options:

The U and G codes are no cause for concern; those files cleanly absorbed changes from the repository. The files marked with U contained no local changes but were Updated with changes from the repository. The G stands for merGed, which means that the file had local changes to begin with, but the changes coming from the repository didn't overlap with the local changes.

But the next two lines are part of a feature (new in Subversion 1.5) called interactive conflict resolution. This means that the changes from the server overlapped with your own, and you have the opportunity to resolve this conflict. The most commonly used options are displayed, but you can see all of the options by typing h:

…
  (p)  postpone    - mark the conflict to be resolved later
  (df) diff-full   - show all changes made to merged file
  (e)  edit        - change merged file in an editor
  (r)  resolved    - accept merged version of file
  (mf) mine-full   - accept my version of entire file (ignore their changes)
  (tf) theirs-full - accept their version of entire file (lose my changes)
  (l)  launch      - launch external tool to resolve conflict
  (h)  help        - show this list

Let's briefly review each of these options before we go into detail on what each option means.

(p)ostpone

Leave the file in a conflicted state for you to resolve after your update is complete.

(d)iff-(f)ull

Display the differences between the base revision and the conflicted file itself in unified diff format.

(e)dit

Open the file in conflict with your favorite editor, as set in the environment variable EDITOR.

(r)esolved

After editing a file, tell svn that you've resolved the conflicts in the file and that it should accept the current contents—basically that you've resolved the conflict.

(m)ine-(f)ull

Discard the newly received changes from the server and use only your local changes for the file under review.

(t)heirs-(f)ull

Discard your local changes to the file under review and use only the newly received changes from the server.

(l)aunch

Launch an external program to perform the conflict resolution. This requires a bit of preparation beforehand.

(h)elp

Show the list of all possible commands you can use in interactive conflict resolution.

We'll cover these commands in more detail now, grouping them together by related functionality.

Viewing conflict differences interactively

Before deciding how to attack a conflict interactively, odds are that you'd like to see exactly what is in conflict, and the diff-full command (df) is what you'll use for this:

…
Select: (p) postpone, (df) diff-full, (e) edit,
        (h)elp for more options : df
--- .svn/text-base/sandwich.txt.svn-base      Tue Dec 11 21:33:57 2007
+++ .svn/tmp/tempfile.32.tmp     Tue Dec 11 21:34:33 2007
@@ -1 +1,5 @@
-Just buy a sandwich.
+<<<<<<< .mine
+Go pick up a cheesesteak.
+=======
+Bring me a taco!
+>>>>>>> .r32
…

The first line of the diff content shows the previous contents of the working copy (the BASE revision), the next content line is your change, and the last content line is the change that was just received from the server (usually the HEAD revision). With this information in hand, you're ready to move on to the next action.

Resolving conflict differences interactively

There are four different ways to resolve conflicts interactively—two of which allow you to selectively merge and edit changes, and two of which allow you to simply pick a version of the file and move along.

If you wish to choose some combination of your local changes, you can use the edit command (e) to manually edit the file with conflict markers in a text editor (determined by the EDITOR environment variable). Editing the file by hand in your favorite text editor is a somewhat low-tech way of remedying conflicts (see the section called “Merging conflicts by hand” for a walkthrough), so some people like to use fancy graphical merge tools instead.

To use a merge tool, you need to either set the SVN_MERGE environment variable or define the merge-tool-cmd option in your Subversion configuration file (see the section called “Configuration Options” for more details). Subversion will pass four arguments to the merge tool: the BASE revision of the file, the revision of the file received from the server as part of the update, the copy of the file containing your local edits, and the merged copy of the file (which contains conflict markers). If your merge tool is expecting arguments in a different order or format, you'll need to write a wrapper script for Subversion to invoke. After you've edited the file, if you're satisfied with the changes you've made, you can tell Subversion that the edited file is no longer in conflict by using the resolve command (r).

If you decide that you don't need to merge any changes, but just want to accept one version of the file or the other, you can either choose your changes (a.k.a. mine) by using the mine-full command (mf) or choose theirs by using the theirs-full command (tf).

Postponing conflict resolution

This may sound like an appropriate section for avoiding marital disagreements, but it's actually still about Subversion, so read on. If you're doing an update and encounter a conflict that you're not prepared to review or resolve, you can type p to postpone resolving a conflict on a file-by-file basis when you run svn update. If you're running an update and don't want to resolve any conflicts, you can pass the --non-interactive option to svn update, and any file in conflict will be marked with a C automatically.

The C stands for conflict. This means that the changes from the server overlapped with your own, and now you have to manually choose between them after the update has completed. When you postpone a conflict resolution, svn typically does three things to assist you in noticing and resolving that conflict:

  • Subversion prints a C during the update and remembers that the file is in a state of conflict.

  • If Subversion considers the file to be mergeable, it places conflict markers—special strings of text that delimit the sides of the conflict—into the file to visibly demonstrate the overlapping areas. (Subversion uses the svn:mime-type property to decide whether a file is capable of contextual, line-based merging. See the section called “File Content Type” to learn more.)

  • For every conflicted file, Subversion places three extra unversioned files in your working copy:

    filename.mine

    This is your file as it existed in your working copy before you updated your working copy—that is, without conflict markers. This file has only your latest changes in it. (If Subversion considers the file to be unmergeable, the .mine file isn't created, since it would be identical to the working file.)

    filename.rOLDREV

    This is the file that was the BASE revision before you updated your working copy. That is, the file that you checked out before you made your latest edits.

    filename.rNEWREV

    This is the file that your Subversion client just received from the server when you updated your working copy. This file corresponds to the HEAD revision of the repository.

    Here OLDREV is the revision number of the file in your .svn directory, and NEWREV is the revision number of the repository HEAD.

For example, Sally makes changes to the file sandwich.txt, but does not yet commit those changes. Meanwhile, Harry commits changes to that same file. Sally updates her working copy before committing and she gets a conflict, which she postpones:

$ svn update
Conflict discovered in 'sandwich.txt'.
Select: (p) postpone, (df) diff-full, (e) edit,
        (h)elp for more options : p
C  sandwich.txt
Updated to revision 2.
$ ls -1
sandwich.txt
sandwich.txt.mine
sandwich.txt.r1
sandwich.txt.r2

At this point, Subversion will not allow Sally to commit the file sandwich.txt until the three temporary files are removed:

$ svn commit -m "Add a few more things"
svn: Commit failed (details follow):
svn: Aborting commit: '/home/sally/svn-work/sandwich.txt' remains in conflict

If you've postponed a conflict, you need to resolve the conflict before Subversion will allow you to commit your changes. You'll do this with the svn resolve command and one of several arguments to the --accept option.

If you want to choose the version of the file that you last checked out before making your edits, choose the base argument.

If you want to choose the version that contains only your edits, choose the mine-full argument.

If you want to choose the version that your most recent update pulled from the server (and thus discarding your edits entirely), choose the theirs-full argument.

However, if you want to pick and choose from your changes and the changes that your update fetched from the server, merge the conflicted text by hand (by examining and editing the conflict markers within the file) and then choose the working argument.

svn resolve removes the three temporary files and accepts the version of the file that you specified with the --accept option, and Subversion no longer considers the file to be in a state of conflict:

$ svn resolve --accept working sandwich.txt
Resolved conflicted state of 'sandwich.txt'

Merging conflicts by hand

Merging conflicts by hand can be quite intimidating the first time you attempt it, but with a little practice, it can become as easy as falling off a bike.

Here's an example. Due to a miscommunication, you and Sally, your collaborator, both edit the file sandwich.txt at the same time. Sally commits her changes, and when you go to update your working copy, you get a conflict and you're going to have to edit sandwich.txt to resolve the conflict. First, let's take a look at the file:

$ cat sandwich.txt
Top piece of bread
Mayonnaise
Lettuce
Tomato
Provolone
<<<<<<< .mine
Salami
Mortadella
Prosciutto
=======
Sauerkraut
Grilled Chicken
>>>>>>> .r2
Creole Mustard
Bottom piece of bread

The strings of less-than signs, equals signs, and greater-than signs are conflict markers and are not part of the actual data in conflict. You generally want to ensure that those are removed from the file before your next commit. The text between the first two sets of markers is composed of the changes you made in the conflicting area:

<<<<<<< .mine
Salami
Mortadella
Prosciutto
=======

The text between the second and third sets of conflict markers is the text from Sally's commit:

=======
Sauerkraut
Grilled Chicken
>>>>>>> .r2

Usually you won't want to just delete the conflict markers and Sally's changes—she's going to be awfully surprised when the sandwich arrives and it's not what she wanted. This is where you pick up the phone or walk across the office and explain to Sally that you can't get sauerkraut from an Italian deli. [5] Once you've agreed on the changes you will commit, edit your file and remove the conflict markers:

Top piece of bread
Mayonnaise
Lettuce
Tomato
Provolone
Salami
Mortadella
Prosciutto
Creole Mustard
Bottom piece of bread

Now use svn resolve, and you're ready to commit your changes:

$ svn resolve --accept working sandwich.txt
Resolved conflicted state of 'sandwich.txt'
$ svn commit -m "Go ahead and use my sandwich, discarding Sally's edits."

Note that svn resolve, unlike most of the other commands we deal with in this chapter, requires that you explicitly list any filenames that you wish to resolve. In any case, you want to be careful and use svn resolve only when you're certain that you've fixed the conflict in your file—once the temporary files are removed, Subversion will let you commit the file even if it still contains conflict markers.

If you ever get confused while editing the conflicted file, you can always consult the three files that Subversion creates for you in your working copy—including your file as it was before you updated. You can even use a third-party interactive merging tool to examine those three files.

Discarding your changes in favor of a newly fetched revision

If you get a conflict and decide that you want to throw out your changes, you can run svn resolve --accept theirs-full CONFLICTED-PATH and Subversion will discard your edits and remove the temporary files:

$ svn update
Conflict discovered in 'sandwich.txt'.
Select: (p) postpone, (df) diff-full, (e) edit,
        (h) help for more options: p
C    sandwich.txt
Updated to revision 2.
$ ls sandwich.*
sandwich.txt  sandwich.txt.mine  sandwich.txt.r2  sandwich.txt.r1
$ svn resolve --accept theirs-full sandwich.txt
Resolved conflicted state of 'sandwich.txt'

Punting: Using svn revert

If you decide that you want to throw out your changes and start your edits again (whether this occurs after a conflict or anytime), just revert your changes:

$ svn revert sandwich.txt
Reverted 'sandwich.txt'
$ ls sandwich.*
sandwich.txt

Note that when you revert a conflicted file, you don't have to use svn resolve.

Commit Your Changes

Finally! Your edits are finished, you've merged all changes from the server, and you're ready to commit your changes to the repository.

The svn commit command sends all of your changes to the repository. When you commit a change, you need to supply a log message describing your change. Your log message will be attached to the new revision you create. If your log message is brief, you may wish to supply it on the command line using the --message (-m) option:

$ svn commit -m "Corrected number of cheese slices."
Sending        sandwich.txt
Transmitting file data .
Committed revision 3.

However, if you've been composing your log message as you work, you may want to tell Subversion to get the message from a file by passing the filename with the --file (-F) option:

$ svn commit -F logmsg
Sending        sandwich.txt
Transmitting file data .
Committed revision 4.

If you fail to specify either the --message (-m) or --file (-F) option, Subversion will automatically launch your favorite editor (see the information on editor-cmd in the section called “Config”) for composing a log message.

[Tip] Tip

If you're in your editor writing a commit message and decide that you want to cancel your commit, you can just quit your editor without saving changes. If you've already saved your commit message, simply delete the text, save again, and then abort:

$ svn commit
Waiting for Emacs...Done

Log message unchanged or not specified
(a)bort, (c)ontinue, (e)dit
a
$

The repository doesn't know or care whether your changes make any sense as a whole; it checks only to make sure nobody else has changed any of the same files that you did when you weren't looking. If somebody has done that, the entire commit will fail with a message informing you that one or more of your files are out of date:

$ svn commit -m "Add another rule"
Sending        rules.txt
svn: Commit failed (details follow):
svn: File '/sandwich.txt' is out of date
…

(The exact wording of this error message depends on the network protocol and server you're using, but the idea is the same in all cases.)

At this point, you need to run svn update, deal with any merges or conflicts that result, and attempt your commit again.

That covers the basic work cycle for using Subversion. Subversion offers many other features that you can use to manage your repository and working copy, but most of your day-to-day use of Subversion will involve only the commands that we've discussed so far in this chapter. We will, however, cover a few more commands that you'll use fairly often.



[3] Of course, nothing is ever totally deleted from the repository—just from the HEAD of the repository. You can get back anything you delete by checking out (or updating your working copy to) a revision earlier than the one in which you deleted it. Also see the section called “Resurrecting Deleted Items”.

[4] And you don't have a WLAN card. Thought you got us, huh?

[5] And if you ask them for it, they may very well ride you out of town on a rail.