Darcs is a revision control system, along the lines of CVS or arch. That means that it keeps track of various revisions and branches of your project, allows for changes to propagate from one branch to another. Darcs is intended to be an ``advanced'' revision control system. Darcs has two particularly distinctive features which differ from other revision control systems: 1) each copy of the source is a fully functional branch, and 2) underlying darcs is a consistent and powerful theory of patches.
![[*]](./crossref.png)
.
Really, only the first of these three layers is of particular interest to
me, so the other two are done as simply as possible. At the database
layer, darcs just has an ordered list of patches along with the patches
themselves, each stored as an individual file. Darcs' distribution system
is strongly inspired by that of arch. Like arch, darcs uses a dumb server,
typically apache or just a local or network file system when pulling
patches. darcs has built-in support for using ssh to write to a remote file
system. A darcs executable is called on the remote system to apply the patches.
Arbitrary other transport protocols are supported, through an environment
variable describing a command that will run darcs on the remote system.
See the documentation for DARCS_APPLY_FOO in Chapter
for details.
The recommended method is to send patches through gpg-signed email messages, which has the advantage of being mostly asynchronous.
In the last paragraph, I explained revision control systems in terms of three layers. One can also look at them as having two distinct uses. One is to provide a history of previous versions. The other is to keep track of changes that are made to the repository, and to allow these changes to be merged and moved from one repository to another. These two uses are distinct, and almost orthogonal, in the sense that a tool can support one of the two uses optimally while providing no support for the other. Darcs is not intended to maintain a history of versions, although it is possible to kludge together such a revision history, either by making each new patch depend on all previous patches, or by tagging regularly. In a sense, this is what the tag feature is for, but the intention is that tagging will be used only to mark particularly notable versions (e.g. released versions, or perhaps versions that pass a time consuming test suite).
Other revision control systems are centered upon the job of keeping track of a history of versions, with the ability to merge changes being added as it was seen that this would be desirable. But the fundamental object remained the versions themselves.
In such a system, a patch (I am using patch here to mean an encapsulated set of changes) is uniquely determined by two trees. Merging changes that are in two trees consists of finding a common parent tree, computing the diffs of each tree with their parent, and then cleverly combining those two diffs and applying the combined diff to the parent tree, possibly at some point in the process allowing human intervention, to allow for fixing up problems in the merge such as conflicts.
In the world of darcs, the source tree is not the fundamental object, but rather the patch is the fundamental object. Rather than a patch being defined in terms of the difference between two trees, a tree is defined as the result of applying a given set of patches to an empty tree. Moreover, these patches may be reordered (unless there are dependencies between the patches involved) without changing the tree. As a result, there is no need to find a common parent when performing a merge. Or, if you like, their common parent is defined by the set of common patches, and may not correspond to any version in the version history.
One useful consequence of darcs' patch-oriented philosophy is that since a
patch need not be uniquely defined by a pair of trees (old and new), we can
have several ways of representing the same change, which differ only in how
they commute and what the result of merging them is. Of course, creating
such a patch will require some sort of user input. This is a Good Thing,
since the user creating the patch should be the one forced to think
about what he really wants to change, rather than the users merging the
patch. An example of this is the token replace patch (See
Section ). This feature makes it possible to create a
patch, for example, which changes every instance of the variable
``stupidly_named_var'' to ``better_var_name'', while leaving
``other_stupidly_named_var'' untouched. When this patch is merged with
any other patch involving the ``stupidly_named_var'', that instance will
also be modified to ``better_var_name''. This is in contrast to a more
conventional merging method which would not only fail to change new
instances of the variable, but would also involve conflicts when merging
with any patch that modifies lines containing the variable. By more using
additional information about the programmer's intent, darcs is thus able to
make the process of changing a variable name the trivial task that it
really is, which is really just a trivial search and replace, modulo
tokenizing the code appropriately.
The patch formalism discussed in Appendix is what makes darcs'
approach possible. In order for a tree to consist of a set of patches,
there must be a deterministic merge of any set of patches, regardless of the
order in which they must be merged. This requires that one be able to
reorder patches. While I don't know that the patches are required to be
invertible as well, my implementation certainly requires invertibility. In
particular, invertibility is required to make use of
Theorem , which is used extensively in the manipulation of
merges.
Darcs is refreshingly different from CVS.
CVS keeps version controlled data in a central repository, and requires that users check out a working directory whenever they wish to access the version-controlled sources. In order to modify the central repository, a user needs to have write access to the central repository; if he doesn't, CVS merely becomes a tool to get the latest sources.
In darcs there is no distinction between working directories and repositories. In order to work on a project, a user makes a local copy of the repository he wants to work in; he may then harness the full power of version control locally. In order to distribute his changes, a user who has write access can push them to the remote repository; one who doesn't can simply send them by e-mail in a format that makes them easy to apply on the remote system.
--dry-run
(summarize remote changes)
--summary
(summarize local changes)
^a
(list potential files to add)
Tools and instructions for migrating CVS repositories to darcs are provided on the darcs community website: http://darcs.net/DarcsWiki/ConvertingFromCvs
Although arch, like darcs, is a distributed system, and the two systems have many similarities (both require no special server, for example), their essential organization is very different.
Like CVS, arch keeps data in two types of data structures:
repositories (called ``archives'') and working directories. In order
to modify a repository, one must first check out a corresponding
working directory. This requires that users remember a number of
different ways of pushing data around -- tla get,
update, commit, archive-mirror and so on.
In darcs, on the other hand, there is no distinction between working
directories and repositories, and just checking out your sources
creates a local copy of a repository. This allows you to harness the
full power of version control in any scratch copy of your sources, and
also means that there are just two ways to push data around:
darcs record, which stores edits into your local
repository, and pull, which moves data between repositories.
(darcs push is merely the opposite of pull;
send and apply are just the two halves of push).
Because of the different models used by arch and darcs, it is
difficult to provide a complete equivalence between arch and darcs.
A rough correspondence for the everyday commands follows:
Tools and instructions for migrating arch repositories to darcs are provided on the darcs community website: http://darcs.net/DarcsWiki/ConvertingFromArch
This chapter should walk you through the steps necessary to build darcs for yourself. There are in general two ways to build darcs. One is for building released versions from tarballs, and the other is to build the latest and greatest darcs, from the darcs repo itself.
Please let me know if you have any problems building darcs, or don't have problems described in this chapter and think there's something obsolete here, so I can keep this page up-to-date.
It is a good idea (but not required) to have software installed that provide darcs network access. The libwww-dev, libwww-ssl-dev or libcurl packages newer than than 7.18.0 are recommended because they provide pipelining support speed up HTTP access. They have to be explicitly chosen with -with-libwww or -with-curl-pipelining. Otherwise, darcs will automatically look for one of libcurl, curl or wget. You also might want to have scp available if you want to grab your repos over ssh...
To use the diff command of darcs, a diff program supporting
options -r (recursive diff) and -N (show new files as
differences against an empty file) is required. The configure
script will look for gdiff, gnudiff and diff in this
order. You can force the use of another program by setting the DIFF
environment variable before running configure.
To rebuild the documentation (which should not be necessary since it is included in html form with the tarballs), you will need to have latex installed, as well as latex2html if you want to build it in html form.
Copy the zlib and curl libraries and headers to both GHC and MinGW. GHC
stores C headers in <ghc-dir>/gcc-lib/include and libraries in
<ghc-dir>/gcc-lib. MinGW stores headers in
<mingw-dir>/include and libraries in <mingw-dir>/lib.
Set PATH to include the <msys-dir>/bin, <mingw-dir>/bin,
<curl-dir>, and a directory containing a pre-built darcs.exe if you
want the build's patch context stored for `darcs --exact-version'.
C:\darcs> cd <darcs-source-dir> C:\darcs> sh $ export GHC=/c/<ghc-dir>/bin/ghc.exe $ autoconf $ ./configure --target=mingw $ make
% ./configure % make # Optional, but recommended % make test % make install
There are options to configure that you may want to check out with
% ./configure --help
If your header files are installed in a non-standard location, you may need
to define the CFLAGS and CPPFLAGS environment variables to
include the path to the headers. e.g. on NetBSD, you may need to run
% CFLAGS=-I/usr/pkg/include CPPFLAGS=-I/usr/pkg/include ./configure
% darcs get -v http://darcs.net/and once you have the darcs repository you can bring it up to date with a
% darcs pull
The repository doesn't hold automatically generated files, which include
the configure script and the HTML documentation, so you need to run
autoconf first.
You'll need autoconf 2.50 or higher. Some systems have more than one
version of autoconf installed. For example, autoconf may point to
version 2.13, while autoconf259 runs version 2.59.
Also note that make is really "GNU make". On some systems, such as
the *BSDs, you may need to type gmake instead of make for this to work.
If you want to create readable documentation you'll need to have latex installed.
% autoconf % ./configure % make % make install
If you want to tweak the configure options, you'll need to run ./configure yourself after the make, and then run make again.
building_darcs.tex) based
on your experience building darcs...
To do so, just record your changes (which you made in the darcs repository)
% darcs record --no-testmaking sure to give the patch a nice descriptive name. The
--no-test options keeps darcs from trying to run the unit tests,
which can be rather time-consuming. Then you can send the patch to the
darcs-devel mailing list by email by
% darcs send -uThe darcs repository stores the email address to which patches should be sent by default. The email address you see is actually my own, but when darcs notices that you haven't signed the patch with my GPG key, it will forward the message to darcs-devel.
This chapter will lead you through an example use of darcs, which hopefully will allow you to get started using darcs with your project.
Creating your repository in the first place just involves telling darcs to create the special directory (called _darcs) in your project tree, which will hold the revision information. This is done by simply calling from the root directory of your project:
% cd my_project/ % darcs initialize --darcs-2This creates the
_darcs directory and populates it with whatever
files and directories are needed to describe an empty project. You now
need to tell darcs what files and directories in your project should be
under revision control. You do this using the command darcs add3.1:
% darcs add *.c Makefile.am configure.acWhen you have added all your files (or at least, think you have), you will want to record your changes. ``Recording'' always includes adding a note as to why the change was made, or what it does. In this case, we'll just note that this is the initial version.
% darcs record --all What is the patch name? Initial revision.Note that since we didn't specify a patch name on the command line we were prompted for one. If the environment variable `EMAIL' isn't set, you will also be prompted for your email address. Each patch that is recorded is given a unique identifier consisting of the patch name, its creator's email address, and the date when it was created.
Now that we have created our repository, make a change to one or more of your files. After making the modification run:
% darcs whatsnewThis should show you the modifications that you just made, in the darcs patch format. If you prefer to see your changes in a different format, read Section
, which describes the whatsnew command in
detail.
Let's say you have now made a change to your project. The next thing to do is to record a patch. Recording a patch consists of grouping together a set of related changes, and giving them a name. It also tags the patch with the date it was recorded and your email address.
To record a patch simply type:
% darcs recorddarcs will then prompt you with all the changes that you have made that have not yet been recorded, asking you which ones you want to include in the new patch. Finally, darcs will ask you for a name for the patch.
You can now rerun whatsnew, and see that indeed the changes you have recorded are no longer marked as new.
% cd /var/www/repos % ln -s /home/username/myproject .
As long as you're running a web server and making your repository available to
the world, you may as well make it easy for people to see what changes
you've made. You can do this by running make installserver, which
installs the program darcs_cgi at /usr/lib/cgi-bin/darcs. You
also will need to create a cache directory named
/var/cache/darcs_cgi, and make sure the owner of that directory is
the same user that your web server runs its cgi scripts as. For me,
this is www-data. Now your friends and enemies should be able to easily
browse your repositories by pointing their web browsers at
http://your.server.org/cgi-bin/darcs.
cd into my repository,
and there type:
% darcs pull http://your.server.org/repos/yourprojectDarcs will check to see if you have recorded any changes that aren't in my current repository. If so, it'll prompt me for each one, to see which ones I want to add to my repository. Note that you may see a different series of prompts depending on your answers, since sometimes one patch depends on another, so if you answer yes to the first one, you won't be prompted for the second if the first depends on it.
Of course, maybe I don't even have a copy of your repository. In that case I'd want to do a
% darcs get http://your.server.org/repos/yourprojectwhich gets the whole repository.
I could instead create an empty repository and fetch all of your patches with pull. Get is just a more efficient way to clone a whole repository.
Get, pull and push also work over ssh. Ssh-paths are of the same form
accepted by scp, namely [username@]host:/path/to/repository.
Darcs is flexible as to how you move patches from one repository to another. This section will introduce all the ways you can get patches from one place to another, starting with the simplest and moving to the most complicated.
The simplest method is the ``all-pull'' method. This involves making each
repository readable (by http, ftp, nfs-mounted disk, whatever), and you
run darcs pull in the repository you want to move the patch to. This is nice,
as it doesn't require you to give write access to anyone else, and is
reasonably simple.
Sometimes you have a machine on which it is not convenient to set up a web
server, perhaps because it's behind a firewall or perhaps for security
reasons, or because it is often turned off. In this case you can use
darcs send
from that computer to generate a patch bundle destined for another
repository. You can either let darcs email the patch for you, or save it
as a file and transfer it by hand. Then in the destination repository you
(or the owner of that repository) run darcs apply to apply the patches contained
in the bundle. This is also quite a simple method since, like the all-pull
method, it doesn't require that you give anyone write access to your
repository. But it's less convenient, since you have to keep track of the
patch bundle (in the email, or whatever).
If you use the send and apply method with email, you'll probably want to
create a _darcs/prefs/email file containing your email address.
This way anyone who sends to your repository will automatically send the
patch bundle to your email address.
If you receive many patches by email, you probably will benefit by running
darcs apply directly from your mail program. I have in my .muttrc
the following:
auto_view text/x-patch text/x-darcs-patch
macro pager A "<pipe-entry>darcs apply --verbose --mark-conflicts \
--reply droundy@abridgegame.org --repodir ~/darcs"
which allows me to view a sent patch, and then apply the patch directly from mutt, sending a
confirmation email to the person who sent me the patch. The autoview line relies on on the following
lines, or something like them, being present in one's .mailcap:
text/x-patch; cat; copiousoutput text/x-darcs-patch; cat; copiousoutput
If you use ssh (and preferably also ssh-agent, so you won't have to keep retyping your password), you can use the push method to transfer changes (using the scp protocol for communication). This method is again not very complicated, since you presumably already have the ssh permissions set up. Push can also be used when the target repository is local, in which case ssh isn't needed. On the other hand, in this situation you could as easily run a pull, so there isn't much benefit.
Note that you can use push to administer a multiple-user repository. You
just need to create a user for the repository (or repositories), and give
everyone with write access ssh access, perhaps using
.ssh/authorized_keys. Then they run
% darcs push repouser@repo.server:repo/directory
Now we get more subtle. If you like the idea in the previous paragraph about creating a repository user to own a repository which is writable by a number of users, you have one other option.
Push --apply-as can run on either a local repository or one accessed
with ssh, but uses sudo to run a darcs apply command (having created
a patch bundle as in send) as another user. You can add the following line
in your sudoers file to allow the users to apply their patches to a
centralized repository:
ALL ALL = (repo-user) NOPASSWD: /usr/bin/darcs apply --all --repodir /repo/path*This method is ideal for a centralized repository when all the users have accounts on the same computer, if you don't want your users to be able to run arbitrary commands as repo-user.
Most of the previous methods are a bit clumsy if you don't want to give each person with write access to a repository an account on your server. Darcs send can be configured to send a cryptographically signed patch by email. You can then set up your mail system to have darcs verify that patches were signed by an authorized user and apply them when a patch is received by email. The results of the apply can be returned to the user by email. Unsigned patches (or patches signed by unauthorized users) will be forwarded to the repository owner (or whoever you configure them to be forwarded to...).
This method is especially nice when combined with the --test option
of darcs apply, since it allows you to run the test suite (assuming you
have one) and reject patches that fail--and it's all done on the server,
so you can happily go on working on your development machine without
slowdown while the server runs the tests.
Setting up darcs to run automatically in response to email is by far the most complicated way to get patches from one repository to another... so it'll take a few sections to explain how to go about it.
When you set up darcs to run apply on signed patches, you should assume
that a user with write access can write to any file or directory that is
writable by the user under which the apply process runs. Unless you
specify the --no-test flag to darcs apply (and this is not
the default), you are also allowing anyone with write access to that
repository to run arbitrary code on your machine (since they can run a test
suite--which they can modify however they like). This is quite a
potential security hole.
For these reasons, if you don't implicitly trust your users, it is recommended that you create a user for each repository to limit the damage an attacker can do with access to your repository. When considering who to trust, keep in mind that a security breach on any developer's machine could give an attacker access to their private key and passphrase, and thus to your repository.
You also must install the following programs: gnupg, a mailer configured to receive mail (e.g. exim, sendmail or postfix), and a web server (usually apache). If you want to be able to browse your repository on the web you must also configure your web server to run cgi scripts and make sure the darcs cgi script was properly installed (by either a darcs-server package, or `make install-server').
You create your gpg key by running (as your normal user):
% gpg --gen-keyYou will be prompted for your name and email address, among other options. Of course, you can skip this step if you already have a gpg key you wish to use.
You now need to export the public key so we can tell the patcher about it. You can do this with the following command (again as your normal user):
% gpg --export "email@address" > /tmp/exported_keyAnd now we can add your key to the
allowed_keys:
(as root)> gpg --keyring /var/lib/darcs/repos/myproject/allowed_keys \
--no-default-keyring --import /tmp/exported_key
You can repeat this process any number of times to authorize multiple users
to send patches to the repository.
You should now be able to send a patch to the repository by running as your normal user, in a working copy of the repository:
% darcs send --sign http://your.computer/repos/myprojectYou may want to add ``send sign'' to the file
_darcs/prefs/defaults
so that you won't need to type --sign every time you want to
send...
If your gpg key is protected by a passphrase, then executing send
with the --sign option might give you the following error:
darcs failed: Error running external program 'gpg'The most likely cause of this error is that you have a misconfigured gpg that tries to automatically use a non-existent gpg-agent program. GnuPG will still work without gpg-agent when you try to sign or encrypt your data with a passphrase protected key. However, it will exit with an error code 2 (
ENOENT) causing darcs to
fail. To fix this, you will need to edit your ~/.gnupg/gpg.conf
file and comment out or remove the line that says:
use-agentIf after commenting out or removing the
use-agent line in your
gpg configuration file you still get the same error, then you probably
have a modified GnuPG with use-agent as a hard-coded option. In that
case, you should change use-agent to no-use-agent to
disable it explicitly.
To begin with, you must configure your repository so that a darcs send to
your repository will know where to send the email. Do this by creating a
file in /path/to/your/repo/_darcs/prefs called email
containing your email address. As a trick (to be explained below), we will
create the email address with ``darcs repo'' as your name, in an email
address of the form ``David Roundy 
droundy@abridgegame.org
.''
% echo 'my darcs repo <user@host.com>' \
> /path/to/your/repo/_darcs/prefs/email
The next step is to set up a gnupg keyring containing the public keys of people authorized to send to your repository. Here I'll give a second way of going about this (see above for the first). This time I'll assume you want to give me write access to your repository. You can do this by:
gpg --no-default-keyring \
--keyring /path/to/the/allowed_keys --recv-keys D3D5BCEC
This works because ``D3D5BCEC'' is the ID of my gpg key, and I have
uploaded my key to the gpg keyservers. Actually, this also requires that
you have configured gpg to access a valid keyserver. You can, of course,
repeat this command for all keys you want to allow access to.
Finally, we add a few lines to your .procmailrc:
:0
* ^TOmy darcs repo
|(umask 022; darcs apply --reply user@host.com \
--repodir /path/to/your/repo --verify /path/to/the/allowed_keys)
The purpose for the ``my darcs repo'' trick is partially to make it easier
to recognize patches sent to the repository, but is even more crucial to
avoid nasty bounce loops by making the --reply option have an email
address that won't go back to the repository. This means that unsigned
patches that are sent to your repository will be forwarded to your ordinary
email.
Like most mail-processing programs, Procmail by default sets a tight umask.
However, this will prevent the repository from remaining world-readable;
thus, the ``umask 022'' is required to relax the umask.
(Alternatively, you could set Procmail's global UMASK variable
to a more suitable value.)
After sending a patch with darcs send, you may not receive any feedback,
even if the patch is applied. You can confirm whether or not your patch was applied
to the remote repository by pointing darcs changes at a remote repository:
darcs changes --last=10 --repo=http://darcs.net/
That shows you the last 10 changes in the remote repository. You can adjust the options given
to changes if a more advanced query is needed.
A Darcs repository contains the patches that Darcs uses to store history, the working directory, and a pristine tree (a copy of the working directory files with no local modifications). For large repositories, this can add up to a fair amount of disk usage.
There are two techniques that can be used to reduce the amount of space used by Darcs repositories: linking and using no pristine tree. The former can be used on any repository; the latter is only suitable in special circumstances, as it makes some operations much slower.
A number of filesystems support linking files, sharing a single file data between different directories. Under some circumstances, when repositories are very similar (typically because they represent different branches of the same piece of software), Darcs will use linking to avoid storing the same file multiple times.
Whenever you invoke darcs get to copy a repository from a local
filesystem onto the same filesystem, Darcs will link patches whenever
possible.
In order to save time, darcs get does not link pristine trees
even when individual files are identical. Additionally, as you pull
patches into trees, patches will become unlinked. This will result in
a lot of wasted space if two repositories have been living for a long
time but are similar. In such a case, you should relink files
between the two repositories.
Relinking is an asymmetric operation: you relink one repository (to
which you must have write access) to another repository, called the
sibling. This is done with darcs optimize --relink, with
-the --sibling flag specifying the sibling.
$ cd /var/repos/darcs.net $ darcs optimize --relink --sibling /var/repos/darcsThe
--sibling flag can be repeated multiple times, in which
case Darcs will try to find a file to link to in all of the siblings.
If a default repository is defined, Darcs will try, as a last resort,
to link against the default repository.
Additional space savings can be achieved by relinking files in the
pristine tree (see below) by using the --relink-pristine flag.
However, doing this prevents Darcs from having precise timestamps on
the pristine files, which carries a moderate performance penalty.
By default, every Darcs repository contains a complete copy of the pristine tree, the working tree as it would be if there were no local edits. By avoiding the need to consult a possibly large number of patches just to find out if a file is modified, the pristine tree makes a lot of operations much faster than they would otherwise be.
Under some circumstances, keeping a whole pristine tree is not desirable. This is the case when preparing a repository to back up, when publishing a repository on a public web server with limited space, or when storing a repository on floppies or small USB keys. In such cases, it is possible to use a repository with no pristine tree.
Darcs automatically recognizes a repository with no pristine
tree. In order to create such a tree, specify the
--no-pristine-tree flag to darcs initialize or
darcs get. There is currently no way to switch an existing
repository to use no pristine tree.
The support for --no-pristine-tree repositories is fairly new,
and has not been extensively optimized yet. Please let us know if you
use this functionality, and which operations you find are too slow.
There are several ways you can adjust darcs' behavior to suit your needs.
The first is to edit files in the _darcs/prefs/ directory of a
repository. Such configuration only applies when working with that
repository. To configure darcs on a per-user rather than per-repository
basis (but with essentially the same methods), you can edit (or create)
files in the ~/.darcs/ directory. Finally, the behavior of some
darcs commands can be modified by setting appropriate environment
variables.
The _darcs directory contains a prefs directory. This
directory exists simply to hold user configuration settings specific to
this repository. The contents of this directory are intended to be
modifiable by the user, although in some cases a mistake in such a
modification may cause darcs to behave strangely.
Default values for darcs commands can be configured on a per-repository
basis by editing (and possibly creating) the _darcs/prefs/defaults
file. Each line of this file has the following form:
COMMAND FLAG VALUEwhere
COMMAND is either the name of the command to which the default
applies, or ALL to indicate that the default applies to all commands
accepting that flag. The FLAG term is the name of the long argument
option without the ``--'', i.e. verbose rather than
--verbose. Finally, the VALUE option can be omitted if the
flag is one such as verbose that doesn't involve a value.
If the value has spaces in it, use single quotes, not double quotes, to surround it.
Each line only takes one flag. To set multiple defaults for the same
command (or for ALL commands), use multiple lines.
Note that the use of ALL easily can have unpredicted consequences,
especially if commands in newer versions of darcs accepts flags that they
didn't in previous versions. A command like obliterate could be
devastating with the ``wrong'' flags (for example -all). Only use safe
flags with ALL.
~/.darcs/defaults |
provides defaults for this user account |
repo/_darcs/prefs/defaults |
provides defaults for one project, |
| overrules changes per user |
For example, if your system clock is bizarre, you could instruct darcs to
always ignore the file modification times by adding the following line to
your _darcs/prefs/defaults file. (Note that this would have to be
done for each repository!)
ALL ignore-times
If you never want to run a test when recording to a particular repository
(but still want to do so when running
check on that repository), and like to name
all your patches ``Stupid patch'', you could use the following:
record no-test record patch-name Stupid patch
If you would like a command to be run every time patches are recorded in a particular repository (for example if you have one central repository, that all developers contribute to), then you can set apply to always run a command when apply is successful. For example, if you need to make sure that the files in the repository have the correct access rights you might use the following. There are two things to note about using darcs this way:
darcs add; doing so would
allow people to modify that file and then run arbitrary scripts on
your main repository, possibly damaging or violating security.
apply posthook chmod -R a+r * apply run-posthook
Similarly, if you need a command to run automatically before darcs performs an action you can use a prehook. Using prehooks it could be possible to canonicalize line endings before recording patches.
There are some options which are meant specifically for use in
_darcs/prefs/defaults. One of them is --disable. As the name
suggests, this option will disable every command that got it as argument. So,
if you are afraid that you could damage your repositories by inadvertent use of
a command like amend-record, add the following line to
_darcs/prefs/defaults:
amend-record disable
Also, a global preferences file can be created with the name
.darcs/defaults in your home directory. Options present there will
be added to the repository-specific preferences.
If they conflict with repository-specific options, the repository-specific
ones will take precedence.
_darcs/prefs/repos file contains a list of repositories you have
pulled from or pushed to, and is used for autocompletion of pull and push
commands in bash. Feel free to delete any lines from this list that might
get in there, or to delete the file as a whole.
_darcs/prefs/author file contains the email address (or name) to
be used as the author when patches are recorded in this repository,
e.g. David Roundy <droundy@abridgegame.org>. This
file overrides the contents of the environment variables
$DARCS_EMAIL and $EMAIL.
_darcs/prefs/boring file may contain a list of regular
expressions describing files, such as object files, that you do not expect
to add to your project. As an example, the boring file that I use with
my darcs repository is:
\.hi$ \.o$ ^\.[^/] ^_ ~$ (^|/)CVS($|/)A newly created repository has a longer boring file that includes many common source control, backup, temporary, and compiled files.
You may want to have the boring file under version
control. To do this you can use darcs setpref to set the value
``boringfile'' to the name of your desired boring file
(e.g. darcs setpref boringfile .boring, where .boring
is the repository path of a file
that has been
darcs added to your repository). The boringfile preference overrides
_darcs/prefs/boring, so be sure to copy that file to the boringfile.
You can also set up a ``boring'' regexps
file in your home directory, named ~/.darcs/boring, which will be
used with all of your darcs repositories.
Any file not already managed by darcs and whose repository path (such
as manual/index.html) matches any of
the boring regular expressions is considered boring. The boring file is
used to filter the files provided to darcs add, to allow you to use a
simple darcs add newdir newdir/*
without accidentally adding a bunch of
object files. It is also used when the --look-for-adds flag is
given to whatsnew or record.
Note that once a file has been added to darcs, it is not considered
boring, even if it matches the boring file filter.
_darcs/prefs/binaries file may contain a list of regular
expressions describing files that should be treated as binary files rather
than text files. Darcs automatically treats files containing
^Z\ or '\0' within the first 4096 bytes as being binary files.
You probably will want to have the binaries file under
version control. To do this you can use darcs setpref to set the value
``binariesfile'' to the name of your desired binaries file
(e.g. darcs setpref binariesfile ./.binaries, where
.binaries is a file that has been
darcs added to your repository). As with the boring file, you can also set
up a ~/.darcs/binaries file if you like.
_darcs/prefs/email file is used to provide the e-mail address for your
repository that others will use when they darcs send a patch back to you.
The contents of the file should simply be an e-mail address.
_darcs/prefs/sources file is used to indicate alternative
locations from which to download patches when using a ``hashed''
repository. This file contains lines such as:
cache:/home/droundy/.darcs/cache readonly:/home/otheruser/.darcs/cache repo:http://darcs.netThis would indicate that darcs should first look in
/home/droundy/.darcs/cache for patches that might be missing, and if
the patch isn't there, it should save a copy there for future use. In that
case, darcs will look in /home/otheruser/.darcs/cache to see if that
user might have downloaded a copy, but won't try to save a copy there, of
course. Finally, it will look in http://darcs.net. Note that the
sources file can also exist in ~/.darcs/. Also note that the
sources mentioned in your sources file will be tried before
the repository you are pulling from. This can be useful in avoiding
downloading patches multiple times when you pull from a remote repository
to more than one local repository.
We strongly advise that you enable a global cache directory, which will allow darcs to avoid re-downloading patches (for example, when doing a second darcs get of the same repository), and also allows darcs to use hard links to reduce disk usage. To do this, simply
mkdir -p $HOME/.darcs/cache echo cache:$HOME/.darcs/cache > $HOME/.darcs/sourcesNote that the cache directory should reside on the same filesystem as your repositories, so you may need to vary this. You can also use multiple cache directories on different filesystems, if you have several filesystems on which you use darcs.
_darcs/prefs/motd file may contain a ``message of the day''
which will be displayed to users who get or pull from the repository without the
--quiet option.
There are a few environment variables whose contents affect darcs' behavior. Here is a quick list of all the variables and their documentation in the rest of the manual:
| Variable | Section |
|---|---|
| DARCS_EDITOR, EDITOR, VISUAL | |
| DARCS_PAGER, PAGER | |
| HOME | |
| TERM | |
| DARCS_EMAIL, EMAIL | |
| DARCS_APPLY_FOO | |
| DARCS_GET_FOO | |
| DARCS_MGET_FOO | |
| DARCS_MGETMAX | |
| DARCS_PROXYUSERPWD | |
| DARCS_WGET | |
| DARCS_SSH | |
| DARCS_SCP | |
| DARCS_SFTP | |
| SSH_PORT | |
| DARCS_ALTERNATIVE_COLOR | |
| DARCS_ALWAYS_COLOR | |
| DARCS_DO_COLOR_LINES | |
| DARCS_DONT_COLOR | |
| DARCS_DONT_ESCAPE_TRAILING_CR | |
| DARCS_DONT_ESCAPE_TRAILING_SPACES | |
| DARCS_DONT_ESCAPE_8BIT | |
| DARCS_DONT_ESCAPE_ANYTHING | |
| DARCS_DONT_ESCAPE_ISPRINT | |
| DARCS_ESCAPE_EXTRA | |
| DARCS_DONT_ESCAPE_EXTRA | |
vi, emacs, emacs -nw and
nano in that order.
more.
/tmp and if /tmp doesn't exist,
it'll put the temporaries in _darcs.
This is very helpful, for example, when recording with a test suite that
uses MPI, in which case using /tmp to hold the test copy is no good,
as /tmp isn't shared over NFS and thus the mpirun call will
fail, since the binary isn't present on the compute nodes.
$HOME/.darcs.
user@foo.org:foo or foo.org:foo. Darcs will use
scp to copy single files (e.g. repository meta-information), and sftp
to copy multiple files in batches (e.g. patches). These commands are
not interpreted by a shell, so you cannot use shell
metacharacters, and the first word in the command must be the name of
an executable located in your path. By default, scp and sftp
are used. When you can use sftp, but not scp (e.g. some ISP web sites), it
works to set DARCS_SCP to `sftp'. The other way around does not work, i.e.
DARCS_FTP must reference an sftp program, not scp.
This method overrides all other ways of getting foo://xxx URLs.
Note that each command should be constructed so that it sends the downloaded content to STDOUT, and the next argument to it should be the URL. Here are some examples that should work for DARCS_GET_HTTP:
fetch -q -o - curl -s -f lynx -source wget -q -O -
If set, DARCS_MGET_FOO will be used to fetch many files from a single repository simultaneously. Replace FOO and foo as appropriate to handle other URL schemes. These commands are not interpreted by a shell, so you cannot use shell metacharacters, and the first word in the command must be the name of an executable located in your path. The GET command will be called with a URL for each file. The MGET command will be invoked with a number of URLs and is expected to download the files to the current directory, preserving the file name but not the path. The APPLY command will be called with a darcs patchfile piped into its standard input. Example:
wget -q
The DARCS_WGET environment variable defines the command that darcs will use to fetch all URLs for remote repositories. The first word in the command must be the name of an executable located in your path. Extra arguments can be included as well, such as:
wget -q
Darcs will append -i to the argument list, which it uses to provide a
list of URLS to download. This allows wget to download multiple patches at the
same time. It's possible to use another command besides wget with this
environment variable, but it must support the -i option in the same way.
These commands are not interpreted by a shell, so you cannot use shell meta-characters.
If the terminal understands ANSI color escape sequences,
darcs will highlight certain keywords and delimiters when printing patches.
This can be turned off by setting the environment variable DARCS_DONT_COLOR to 1.
If you use a pager that happens to understand ANSI colors, like less -R,
darcs can be forced always to highlight the output
by setting DARCS_ALWAYS_COLOR to 1.
If you can't see colors you can set DARCS_ALTERNATIVE_COLOR to 1,
and darcs will use ANSI codes for bold and reverse video instead of colors.
In addition, there is an extra-colorful mode, which is not enabled by
default, which can be activated with DARCS_DO_COLOR_LINES.
By default darcs will escape (by highlighting if possible) any kind of spaces at the end of lines when showing patch contents. If you don't want this you can turn it off by setting DARCS_DONT_ESCAPE_TRAILING_SPACES to 1. A special case exists for only carriage returns: DARCS_DONT_ESCAPE_TRAILING_CR.
Darcs needs to escape certain characters when printing patch contents to a terminal. Characters like backspace can otherwise hide patch content from the user, and other character sequences can even in some cases redirect commands to the shell if the terminal allows it.
By default darcs will only allow printable 7-bit ASCII characters (including space),
and the two control characters tab and newline.
(See the last paragraph in this section for a way to tailor this behavior.)
All other octets are printed in quoted form (as ^<control letter> or
\<hex code>).
Darcs has some limited support for locales. If the system's locale is a single-byte character encoding, like the Latin encodings, you can set the environment variable DARCS_DONT_ESCAPE_ISPRINT to 1 and darcs will display all the printables in the current system locale instead of just the ASCII ones. NOTE: This curently does not work on some architectures if darcs is compiled with GHC 6.4. Some non-ASCII control characters might be printed and can possibly spoof the terminal.
For multi-byte character encodings things are less smooth. UTF-8 will work if you set DARCS_DONT_ESCAPE_8BIT to 1, but non-printables outside the 7-bit ASCII range are no longer escaped. E.g., the extra control characters from Latin-1 might leave your terminal at the mercy of the patch contents. Space characters outside the 7-bit ASCII range are no longer recognized and will not be properly escaped at line endings.
As a last resort you can set DARCS_DONT_ESCAPE_ANYTHING to 1. Then everything that doesn't flip code sets should work, and so will all the bells and whistles in your terminal. This environment variable can also be handy if you pipe the output to a pager or external filter that knows better than darcs how to handle your encoding. Note that all escaping, including the special escaping of any line ending spaces, will be turned off by this setting.
There are two environment variables you can set to explicitly tell darcs to not escape or escape octets. They are DARCS_DONT_ESCAPE_EXTRA and DARCS_ESCAPE_EXTRA. Their values should be strings consisting of the verbatim octets in question. The do-escapes take precedence over the dont-escapes. Space characters are still escaped at line endings though. The special environment variable DARCS_DONT_ESCAPE_TRAILING_CR turns off escaping of carriage return last on the line (DOS style).
This chapter is intended to review various scenarios and describe in each case effective ways of using darcs. There is no one ``best practice'', and darcs is a sufficiently low-level tool that there are many high-level ways one can use it, which can be confusing to new users. The plan (and hope) is that various users will contribute here describing how they use darcs in different environments. However, this is not a wiki, and contributions will be edited and reviewed for consistency and wisdom.
This section will lay down the concepts around patch creation. The aim is to develop a way of thinking that corresponds well to how darcs is behaving -- even in complicated situations.
In a single darcs repository you can think of two ``versions'' of the source tree.
They are called the working and pristine trees.
Working is your normal source tree, with or without darcs alongside.
The only thing that makes it part of a darcs repository
is the _darcs directory in its root.
Pristine is the recorded state of the source tree.
The pristine tree is constructed from groups of changes,
called patches (some other version control systems use the
term changeset instead of patch).5.1 Darcs will create and store these patches
based on the changes you make in working.
whatsnew (as well as diff) can show
the difference between working and pristine to you.
It will be shown as a difference in working.
In advanced cases it need not be working that has changed;
it can just as well have been pristine, or both.
The important thing is the difference and what darcs can do with it.
record it to keep it, or revert it to lose the changes.5.2
If you have a difference between working and pristine--for example after editing some files in working--whatsnew will show some ``unrecorded changes''.
To save these changes, use record.
It will create a new patch in pristine with the same changes,
so working and pristine are no longer different.
To instead undo the changes in working, use revert.
It will modify the files in working to be the same as in pristine
(where the changes do not exist).
unrecord is a command meant to be run only in private
repositories. Its intended purpose is to allow developers the flexibility
to undo patches that haven't been distributed yet.
However, darcs does not prevent you from unrecording a patch that has been copied to another repository. Be aware of this danger!
If you unrecord a patch, that patch will be deleted from pristine.
This will cause working to be different from pristine,
and whatsnew to report unrecorded changes.
The difference will be the same as just before that patch was recorded.
Think about it.
record examines what's different with working
and constructs a patch with the same changes in pristine
so they are no longer different.
unrecord deletes this patch;
the changes in pristine disappear and the difference is back.
If the recorded changes included an error, the resulting flawed patch can be unrecorded. When the changes have been fixed, they can be recorded again as a new--hopefully flawless--patch.
If the whole change was wrong it can be discarded from working too,
with revert.
revert will update working to the state of pristine,
in which the changes do no longer exist after the patch was deleted.
Keep in mind that the patches are your history,
so deleting them with unrecord makes it impossible to track
what changes you really made.
Redoing the patches is how you ``cover the tracks''.
On the other hand,
it can be a very convenient way to manage and organize changes
while you try them out in your private repository.
When all is ready for shipping,
the changes can be reorganized in what seems as useful and impressive patches.
Use it with care.
All patches are global, so don't ever replace an already ``shipped'' patch in this way! If an erroneous patch is deleted and replaced with a better one, you have to replace it in all repositories that have a copy of it. This may not be feasible, unless it's all private repositories. If other developers have already made patches or tags in their repositories that depend on the old patch, things will get complicated.
The patches described in the previous sections have mostly been hunks. A hunk is one of darcs' primitive patch types, and it is used to remove old lines and/or insert new lines. There are other types of primitive patches, such as adddir and addfile which add new directories and files, and replace which does a search-and-replace on tokens in files.
Hunks are always calculated in place with a diff algorithm
just before whatsnew or record.
But other types of primitive patches need to be explicitly created
with a darcs command.
They are kept in pending5.3until they are either recorded or reverted.
Pending can be thought of as a special extension of working.
When you issue, e.g., a darcs replace command,
the replace is performed on the files in working
and at the same time a replace patch is put in pending.
Patches in pending describe special changes made in working.
The diff algorithm will fictively apply these changes to pristine
before it compares it to working,
so all lines in working that are changed by a replace command
will also be changed in pending
pristine
when the hunks are calculated.
That's why no hunks with the replaced lines will be shown by whatsnew;
it only shows the replace patch in pending responsible for the change.
If a special patch is recorded, it will simply be moved to pristine. If it is instead reverted, it will be deleted from pending and the accompanying change will be removed from working.
Note that reverting a patch in pending is not the same as
simply removing it from pending.
It actually applies the inverse of the change to working.
Most notable is that reverting an addfile patch
will delete the file in working (the inverse of adding it).
So if you add the wrong file to darcs by mistake,
don't revert the addfile.
Instead use remove, which cancels out the addfile in pending.
This section will lay down the concepts around patch distribution and branches. The aim is to develop a way of thinking that corresponds well to how darcs is behaving -- even in complicated situations.
A repository is a collection of patches. Patches have no defined order, but patches can have dependencies on other patches. Patches can be added to a repository in any order as long as all patches depended upon are there. Patches can be removed from a repository in any order, as long as no remaining patches depend on them.
Repositories can be cloned to create branches. Patches created in different branches may conflict. A conflict is a valid state of a repository. A conflict makes the working tree ambiguous until the conflict is resolved.
There are two kinds of dependencies: implicit dependencies and explicit dependencies.
Implicit dependencies is the far most common kind. These are calculated automatically by darcs. If a patch removes a file or a line of code, it will have to depend on the patch that added that file or line of code.5.4If a patch adds a line of code, it will usually have to depend on the patch or patches that added the adjacent lines.
Explicit dependencies can be created if you give the --ask-deps option to darcs record.
This is good for assuring that logical dependencies hold between patches.
It can also be used to group patches--a patch with explicit dependencies doesn't need to change anything--and pulling the patch also pulls all patches it was made to depend on.
Darcs does not have branches--it doesn't need to. Every repository can be used as a branch. This means that any two repositories are ``branches'' in darcs, but it is not of much use unless they have a large portion of patches in common. If they are different projects they will have nothing in common, but darcs may still very well be able to merge them, although the result probably is nonsense. Therefore the word ``branch'' isn't a technical term in darcs; it's just the way we think of one repository in relation to another.
Branches are very useful in darcs.
They are in fact necessary if you want to do more than only simple work.
When you get someone's repository from the Internet,
you are actually creating a branch of it.
It may first seem inefficient (or if you come from CVS--frightening),
not to say plain awkward.
But darcs is designed this way, and it has means to make it efficient.
The answer to many questions about how to do a thing with darcs is: ``use a branch''.
It is a simple and elegant solution with great power and flexibility,
which contributes to darcs' uncomplicated user interface.
You create new branches (i.e., clone repositories)
with the get and put commands.
Patches are global, and a copy of a patch either is or is not present in a branch. This way you can rig a branch almost any way you like, as long as dependencies are fulfilled--darcs won't let you break dependencies. If you suspect a certain feature from some time ago introduced a bug, you can remove the patch/patches that adds the feature, and try without it.5.5
Patches are added to a repository with pull
and removed from the repositories with obliterate.
Don't confuse these two commands with record and unrecord,
which constructs and deconstructs patches.
It is important not to lose patches when (re)moving them around.
pull needs a source repository to copy the patch from,
whereas obliterate just erases the patch.
Beware that if you obliterate all copies of a patch
it is completely lost--forever.
Therefore you should work with branches when you obliterate patches.
The obliterate command can wisely be disabled in a dedicated main repository
by adding obliterate disable to the repository's defaults file.
For convenience, there is a push command.
It works like pull but in the other direction.
It also differs from pull in an important way:
it starts a second instance of darcs to apply the patch in the target repository,
even if it's on the same computer.
It can cause surprises if you have a ``wrong'' darcs in your PATH.
While pull and obliterate can be used to
construct different ``versions'' in a repository,
it is often desirable to name specific configurations of patches
so they can be identified and retrieved easily later.
This is how darcs implements what is usually known as versions.
The command for this is tag,
and it records a tag in the current repository.
A tag is just a patch, but it only contains explicit dependencies. It will depend on all the patches in the current repository.5.6Darcs can recognize if a patch is as a tag; tags are sometimes treated specially by darcs commands.
While traditional revision control systems tag versions in the time line history, darcs lets you tag any configuration of patches at any time, and pass the tags around between branches.
With the option --tag to get you can easily get
a named version in the repository
as a new branch.
This part of darcs becomes a bit complicated, and the description given here is slightly simplified.
Conflicting patches are created when you record changes to the same line in two different repositories. Same line does not mean the same line number and file name, but the same line added by a common depended-upon patch.
If you are using a darcs-2 repository (Section ),
darcs does not consider two patches making the same change to be a
conflict, much in the same fashion as other version control systems.
(The caveat here is two non-identical patches with some identical
changes may conflict. For the most part, darcs should just do what you
expect).
A conflict happens when two conflicting patches meet in the same repository. This is no problem for darcs; it can happily pull together just any patches. But it is a problem for the files in working (and pristine). The conflict can be thought of as two patches telling darcs different things about what a file should look like.
Darcs escapes this problem by ignoring those parts5.7of the patches that conflict. They are ignored in both patches. If patch A changes the line ``FIXME'' to ``FIXED'', and patch B changes the same line to ``DONE'', the two patches together will produce the line ``FIXME''. Darcs doesn't care which one you pulled into the repository first, you still get the same result when the conflicting patches meet. All other changes made by A and B are performed as normal.
Darcs can mark a conflict for you in working.
This is done with mark-conflicts.
Conflicts are marked such that both conflicting changes
are inserted with special delimiter lines around them.
Then you can merge the two changes by hand,
and remove the delimiters.
When you pull patches,
darcs automatically performs a mark-conflicts for you if a conflict happens.
You can remove the markup with revert,
Remember that the result will be the lines from
the previous version common to both conflicting patches.
The conflict marking can be redone again with mark-conflicts.
A special case is when a pulled patch conflicts with unrecorded changes in the repository.
The conflict will be automatically marked as usual,
but since the markup is also an unrecorded change,
it will get mixed in with your unrecorded changes.
There is no guarantee you can revert only the markup after this,
and resolve will not be able to redo this markup later if you remove it.
It is good practice to record important changes before pulling.
mark-conflicts can't mark complicated conflicts.
In that case you'll have to use darcs diff and other commands
to understand what the conflict is all about.
If for example two conflicting patches create the same file,
mark-conflicts will pick just one of them,
and no delimiters are inserted.
So watch out if darcs tells you about a conflict.
mark-conflicts can also be used to check for unresolved conflicts.
If there are none, darcs replies ``No conflicts to resolve''.
While pull reports when a conflict happens,
obliterate and get don't.
A conflict is resolved
(not marked, as with the command mark-conflicts)
as soon as some new patch depends on the conflicting patches.
This will usually be the resolve patch you record after manually putting together the pieces
from the conflict markup produced by mark-conflicts (or pull).
But it can just as well be a tag.
So don't forget to fix conflicts before you accidently ``resolve'' them by recording other patches.
If the conflict is with one of your not-yet-published patches, you may choose to amend that patch rather than creating a resolve patch.
If you want to back out and wait with the conflict,
you can obliterate the conflicting patch you just pulled.
Before you can do that you have to revert the conflict markups
that pull inserted when the conflict happened.
This is how darcs itself is developed. There are many contributors to
darcs, but every contribution is reviewed and manually applied by myself.
For this sort of a situation, darcs send is ideal, since the barrier for
contributions is very low, which helps encourage contributors.
One could simply set the _darcs/prefs/email value to the project
mailing list, but I also use darcs send to send my changes to the main
server, so instead the email address is set to
``Davids Darcs Repo <droundy@abridgegame.org>''. My
.procmailrc
file on the server has the following rule:
:0
* ^TODavids Darcs Repo
|(umask 022; darcs apply --reply darcs-devel@abridgegame.org \
--repodir /path/to/repo --verify /path/to/allowed_keys)
This causes darcs apply to be run on any email sent to ``Davids Darcs
Repo''.
apply actually applies them only if they are signed by an
authorized key. Currently, the only authorized key is mine, but of course
this could be extended easily enough.
The central darcs repository contains the following values in its
_darcs/prefs/defaults:
apply test apply verbose apply happy-forwardingThe first line tells apply to always run the test suite. The test suite is in fact the main reason I use send rather than push, since it allows me to easily continue working (or put my computer to sleep) while the tests are being run on the main server. The second line is just there to improve the email response that I get when a patch has either been applied or failed the tests. The third line makes darcs not complain about unsigned patches, but just to forward them to
darcs-devel.
On my development computer, I have in my .muttrc the following
alias, which allows me to easily apply patches that I get via email
directly to my darcs working directory:
macro pager A "<pipe-entry>(umask 022; darcs apply --no-test -v \
--repodir ~/darcs)"
This section describes the development method used for the density
functional theory code DFT++, which is available at
http://dft.physics.cornell.edu/dft.
We have a number of workstations which all mount the same /home via NFS.
We created a special ``dft'' user, with the central repository living in that
user's home directory. The ssh public keys of authorized persons are added to
the ``dft'' user's .ssh/allowed_keys, and we commit patches to this
repository using
darcs push. As in Section ,
we have the central repository set to run the test suite before the push goes
through.
Note that no one ever runs as the dft user.
A subtlety that we ran into showed up in the running of the test suite.
Since our test suite includes the running of MPI programs, it must be run
in a directory that is mounted across our cluster. To achieve this, we set
the $DARCS_TMPDIR environment variable to ~/tmp.
Note that even though there are only four active developers at the moment,
the distributed nature of darcs still plays a large role. Each developer
works on a feature until it is stable, a process that often takes quite a
few patches, and only once it is stable does he
push to the central repository.
It's easy to have several personal development trees using darcs, even when working on a team or with shared code. The number and method of using each personal tree is limited only by such grand limitations as: your disk space, your imagination, available time, etc.
For example, if the central darcs repository for your development team
is 
, you can create a local working directory for feature
. This local working directory contains a full copy of 
(as of the time of the ``darcs get'' operation) and can be denoted
. In the midst of working on feature , you realize it
requires the development of a separate feature 
. Rather than
intermingling and in the same working tree, you can create
a new working tree for working on , where that working tree
contains the repository 
.
While working on , other developers may have made other changes;
these changes can be retrieved on a per-patch selection basis by
periodic ``darcs pull'' operations.
When your work on is completed, you can publish it for the use
of other developers by a ``darcs push'' (or ``darcs send'') from
to . Independently of the publishing of , you can merge
your work into your working tree by a ``darcs pull ''
in the development tree (or ``darcs push'' from to ).
When your work on is completed, you publish it as well by a
``darcs push'' from to .
Your local feature development efforts for or can each
consist of multiple patches. When pushing or pulling to other trees,
these patches can either all be selected or specific patches can be
selected. Thus, if you introduce a set of debugging calls into the
code, you can commit the debugging code in a distictly separate patch
(or patches) that you will not push to .
As discussed in the section above, a developer may have various changes to their local development repositories that they do not ever wish to publish to a group repository (e.g. personal debugging code), but which they would like to keep in their local repository, and perhaps even share amongst their local repositories.
This is easily done via darcs, since those private changes can be
committed in patches that are separate from other patches; during the
process of pushing patches to the common repository (), the
developer is queried for which patches should be moved to () on a
patch-by-patch basis.
The --complement flag for the ``darcs pull'' operation can
further simplify this effort. If the developer copies (via ``darcs
push'' or ``darcs pull'') all private patches into a special
repository/working tree (
), then those patches are easily
disregarded for pulling by adding --complement to the ``darcs
pull'' line and listing after the primary source repository.
The --complement flag is only available for ``darcs pull'', and
not ``darcs push'' or ``darcs send'', requiring the user to have pull
access to the target repository. While the actual public repository
is often not shared in this manner, it's simple to create a local
version of the public repository to act as the staging area for that
public repository.
The following example extends the two feature addition example in the
previous section using a local staging repository (
) and a
private patch repository:
$ cd working-dir $ darcs get http://server/repos/Rc Rl $ darcs get Rl R1 $ cd R1 ...development of f1 $ darcs record -m'p1: f1 initial work' ... $ darcs record -m'p2: my debugging tracepoints' ... $ cd .. $ darcs get http://server/repos/Rc R2 $ cd R2 $ darcs pull -p p2 ../R1 ... development of f2 $ darcs record -m'p3: f2 finished' $ cd .. $ darcs get Rl Rp $ cd Rp $ darcs pull -p p2 ../R2 $ cd ../Rl $ darcs pull --complement ../R2 ../Rp $ darcs send ... for publishing f2 patches to Rc $ cd ../R1 $ darcs pull ../R2 ... updates R1 with f2 changes from R2 ... more development of f1 $ darcs record -m'p4: f1 feature finished.' $ cd ../Rl $ darcs pull --complement ../R1 ../Rp $ darcs send
The general format of a darcs command is
% darcs COMMAND OPTIONS ARGUMENTS ...Here
COMMAND is a command such as add or record, which of
course may have one or more arguments. Options have the form
--option or -o, while arguments vary from command to
command. There are many options which are common to a number of different
commands, which will be summarized here.
If you wish, you may use any unambiguous beginning of a command name as a
shortcut: for darcs record, you could type darcs recor or
darcs rec, but not darcs re since that could be confused with
darcs replace, darcs revert and darcs remove.
In some cases, COMMAND actually consists of two words, a
super-command and a subcommand. For example, the ``display the
manifest'' command has the form darcs query manifest.
Not all commands modify the ``patches'' of your repository (that is, the named patches which other users can pull); some commands only affect the copy of the source tree you're working on (your ``working directory''), and some affect both. This table summarizes what you should expect from each one and will hopefully serve as guide when you're having doubts about which command to use.
| affects | patches | working directory |
| record | yes | no |
| unrecord | yes | no |
| rollback | yes | yes |
| revert | no | yes |
| unrevert | no | yes |
| pull | yes | yes |
| obliterate | yes | yes |
| apply | yes | yes |
| push6.1 | no | no |
| send6.2 | no | no |
| put6.3 | no | no |
COMMAND accepts --help as an argument, which tells it to
provide a bit of help. Among other things, this help always provides an
accurate listing of the options available with that command, and is
guaranteed never to be out of sync with the version of darcs you actually
have installed (unlike this manual, which could be for an entirely
different version of darcs).
% darcs COMMAND --help
--disable option, which can be used in
_darcs/prefs/defaults to disable some commands in the repository. This
can be helpful if you want to protect the repository from accidental use of
advanced commands like obliterate, unpull, unrecord or amend-record.
--verbose option, which tells darcs to
provide additional output. The amount of verbosity varies from command to
command. Commands that accept --verbose\verb also accept --quiet\verb,
which surpresses non-error output, and --normal-verbosity\verb which can be
used to restore the default verbosity if --verbose or --quiet is in
the defaults file.
--debug option, which causes darcs to generate
additional output that may be useful for debugging its behavior, but which otherwise
would not be interesting.
--repodir option, which allows you to
specify the directory of the repository in which to perform the command.
This option is used with commands, such as whatsnew, that ordinarily would
be performed within a repository directory, and allows you to use those
commands without actually being in the repository directory when calling the
command. This is useful when running darcs in a pipe, as might be the case
when running apply from a mailer.
Some commands, such as pull require a remote repository to be specified,
either from the command line or as a default. The --remote-repo
provides an alternative way to supply this remote repository path. This flag
can be seen as temporarily ``replacing'' the default repository. Setting it
causes the command to ignore the default repository (it also does not affect,
i.e. overwrite the default repository). On the other hand, if any other
repositories are supplied as command line arguments, this flag will be ignored
(and the default repository may be overwritten).
Many commands operate on a patch or patches that have already been recorded.
There are a number of options that specify which patches are selected for
these operations: --patch, --match, --tag, and variants
on these, which for --patch are --patches,
--from-patch, and --to-patch. The --patch and
--tag forms simply take (POSIX extended, aka egrep) regular
expressions and match them against tag and patch names. --match,
described below, allows more powerful patterns.
The plural forms of these options select all matching patches. The singular forms select the last matching patch. The range (from and to) forms select patches after or up to (both inclusive) the last matching patch.
These options use the current order of patches in the repository. darcs may reorder patches, so this is not necessarily the order of creation or the order in which patches were applied. However, as long as you are just recording patches in your own repository, they will remain in order.
When a patch or a group of patches is selected, all patches they depend on
get silently selected too. For example: darcs pull --patches bugfix
means ``pull all the patches with `bugfix' in their name, along with any
patches they require.'' If you really only want patches with `bugfix' in
their name, you should use the --no-deps option, which makes darcs
exclude any matched patches from the selection which have dependencies that
are themselves not explicitly matched by the selection.
For unrecord, unpull and obliterate, patches that
depend on the selected patches are silently included, or if
--no-deps is used selected patches with dependencies on not selected
patches are excluded from the selection.
Currently --match accepts five primitive match types, although
there are plans to expand it to match more patterns. Also, note that the
syntax is still preliminary and subject to change.
The first match type accepts a literal string which is checked against the patch name. The syntax is
darcs annotate --summary --match 'exact foo+bar'This is useful for situations where a patch name contains characters that could be considered special for regul