hgbook

annotate en/hook.tex @ 54:e94202d88199

find changesets by author, revision, files, or words in the commit message
Tix fypos.
author Bryan O'Sullivan <bos@serpentine.com>
date Wed Aug 02 13:08:56 2006 -0700 (2006-08-02)
parents 497aa3c9d4ce
children 8806b2875f10
rev   line source
bos@34 1 \chapter{Handling repository events with hooks}
bos@34 2 \label{chap:hook}
bos@34 3
bos@34 4 Mercurial offers a powerful mechanism to let you perform automated
bos@34 5 actions in response to events that occur in a repository. In some
bos@34 6 cases, you can even control Mercurial's response to those events.
bos@34 7
bos@34 8 The name Mercurial uses for one of these actions is a \emph{hook}.
bos@34 9 Hooks are called ``triggers'' in some revision control systems, but
bos@34 10 the two names refer to the same idea.
bos@34 11
bos@38 12 \section{An overview of hooks in Mercurial}
bos@38 13
bos@41 14 Here is a brief list of the hooks that Mercurial supports. We will
bos@41 15 revisit each of these hooks in more detail later, in
bos@41 16 section~\ref{sec:hook:ref}.
bos@41 17
bos@38 18 \begin{itemize}
bos@38 19 \item[\small\hook{changegroup}] This is run after a group of
bos@41 20 changesets has been brought into the repository from elsewhere.
bos@38 21 \item[\small\hook{commit}] This is run after a new changeset has been
bos@41 22 created in the local repository.
bos@38 23 \item[\small\hook{incoming}] This is run once for each new changeset
bos@38 24 that is brought into the repository from elsewhere. Notice the
bos@38 25 difference from \hook{changegroup}, which is run once per
bos@41 26 \emph{group} of changesets brought in.
bos@38 27 \item[\small\hook{outgoing}] This is run after a group of changesets
bos@41 28 has been transmitted from this repository.
bos@38 29 \item[\small\hook{prechangegroup}] This is run before starting to
bos@41 30 bring a group of changesets into the repository.
bos@41 31 \item[\small\hook{precommit}] Controlling. This is run before starting
bos@41 32 a commit.
bos@41 33 \item[\small\hook{preoutgoing}] Controlling. This is run before
bos@41 34 starting to transmit a group of changesets from this repository.
bos@41 35 \item[\small\hook{pretag}] Controlling. This is run before creating a tag.
bos@41 36 \item[\small\hook{pretxnchangegroup}] Controlling. This is run after a
bos@41 37 group of changesets has been brought into the local repository from
bos@41 38 another, but before the transaction completes that will make the
bos@41 39 changes permanent in the repository.
bos@41 40 \item[\small\hook{pretxncommit}] Controlling. This is run after a new
bos@41 41 changeset has been created in the local repository, but before the
bos@41 42 transaction completes that will make it permanent.
bos@41 43 \item[\small\hook{preupdate}] Controlling. This is run before starting
bos@41 44 an update or merge of the working directory.
bos@38 45 \item[\small\hook{tag}] This is run after a tag is created.
bos@38 46 \item[\small\hook{update}] This is run after an update or merge of the
bos@38 47 working directory has finished.
bos@38 48 \end{itemize}
bos@41 49 Each of the hooks whose description begins with the word
bos@41 50 ``Controlling'' has the ability to determine whether an activity can
bos@41 51 proceed. If the hook succeeds, the activity may proceed; if it fails,
bos@41 52 the activity is either not permitted or undone, depending on the hook.
bos@38 53
bos@38 54 \section{Hooks and security}
bos@38 55
bos@38 56 \subsection{Hooks are run with your privileges}
bos@38 57
bos@38 58 When you run a Mercurial command in a repository, and the command
bos@41 59 causes a hook to run, that hook runs on \emph{your} system, under
bos@41 60 \emph{your} user account, with \emph{your} privilege level. Since
bos@41 61 hooks are arbitrary pieces of executable code, you should treat them
bos@41 62 with an appropriate level of suspicion. Do not install a hook unless
bos@41 63 you are confident that you know who created it and what it does.
bos@38 64
bos@38 65 In some cases, you may be exposed to hooks that you did not install
bos@38 66 yourself. If you work with Mercurial on an unfamiliar system,
bos@38 67 Mercurial will run hooks defined in that system's global \hgrc\ file.
bos@38 68
bos@38 69 If you are working with a repository owned by another user, Mercurial
bos@41 70 can run hooks defined in that user's repository, but it will still run
bos@41 71 them as ``you''. For example, if you \hgcmd{pull} from that
bos@41 72 repository, and its \sfilename{.hg/hgrc} defines a local
bos@41 73 \hook{outgoing} hook, that hook will run under your user account, even
bos@41 74 though you don't own that repository.
bos@38 75
bos@38 76 \begin{note}
bos@38 77 This only applies if you are pulling from a repository on a local or
bos@38 78 network filesystem. If you're pulling over http or ssh, any
bos@41 79 \hook{outgoing} hook will run under whatever account is executing
bos@41 80 the server process, on the server.
bos@38 81 \end{note}
bos@38 82
bos@38 83 XXX To see what hooks are defined in a repository, use the
bos@38 84 \hgcmdargs{config}{hooks} command. If you are working in one
bos@38 85 repository, but talking to another that you do not own (e.g.~using
bos@38 86 \hgcmd{pull} or \hgcmd{incoming}), remember that it is the other
bos@38 87 repository's hooks you should be checking, not your own.
bos@38 88
bos@38 89 \subsection{Hooks do not propagate}
bos@38 90
bos@38 91 In Mercurial, hooks are not revision controlled, and do not propagate
bos@38 92 when you clone, or pull from, a repository. The reason for this is
bos@38 93 simple: a hook is a completely arbitrary piece of executable code. It
bos@38 94 runs under your user identity, with your privilege level, on your
bos@38 95 machine.
bos@38 96
bos@38 97 It would be extremely reckless for any distributed revision control
bos@38 98 system to implement revision-controlled hooks, as this would offer an
bos@38 99 easily exploitable way to subvert the accounts of users of the
bos@38 100 revision control system.
bos@38 101
bos@38 102 Since Mercurial does not propagate hooks, if you are collaborating
bos@38 103 with other people on a common project, you should not assume that they
bos@38 104 are using the same Mercurial hooks as you are, or that theirs are
bos@38 105 correctly configured. You should document the hooks you expect people
bos@38 106 to use.
bos@38 107
bos@38 108 In a corporate intranet, this is somewhat easier to control, as you
bos@38 109 can for example provide a ``standard'' installation of Mercurial on an
bos@38 110 NFS filesystem, and use a site-wide \hgrc\ file to define hooks that
bos@38 111 all users will see. However, this too has its limits; see below.
bos@38 112
bos@38 113 \subsection{Hooks can be overridden}
bos@38 114
bos@38 115 Mercurial allows you to override a hook definition by redefining the
bos@38 116 hook. You can disable it by setting its value to the empty string, or
bos@38 117 change its behaviour as you wish.
bos@38 118
bos@38 119 If you deploy a system-~or site-wide \hgrc\ file that defines some
bos@38 120 hooks, you should thus understand that your users can disable or
bos@38 121 override those hooks.
bos@38 122
bos@38 123 \subsection{Ensuring that critical hooks are run}
bos@38 124
bos@38 125 Sometimes you may want to enforce a policy that you do not want others
bos@38 126 to be able to work around. For example, you may have a requirement
bos@38 127 that every changeset must pass a rigorous set of tests. Defining this
bos@38 128 requirement via a hook in a site-wide \hgrc\ won't work for remote
bos@38 129 users on laptops, and of course local users can subvert it at will by
bos@38 130 overriding the hook.
bos@38 131
bos@38 132 Instead, you can set up your policies for use of Mercurial so that
bos@38 133 people are expected to propagate changes through a well-known
bos@38 134 ``canonical'' server that you have locked down and configured
bos@38 135 appropriately.
bos@38 136
bos@38 137 One way to do this is via a combination of social engineering and
bos@38 138 technology. Set up a restricted-access account; users can push
bos@38 139 changes over the network to repositories managed by this account, but
bos@38 140 they cannot log into the account and run normal shell commands. In
bos@38 141 this scenario, a user can commit a changeset that contains any old
bos@38 142 garbage they want.
bos@38 143
bos@38 144 When someone pushes a changeset to the server that everyone pulls
bos@38 145 from, the server will test the changeset before it accepts it as
bos@38 146 permanent, and reject it if it fails to pass the test suite. If
bos@38 147 people only pull changes from this filtering server, it will serve to
bos@38 148 ensure that all changes that people pull have been automatically
bos@38 149 vetted.
bos@38 150
bos@41 151 \section{Using hooks with shared access to a repository}
bos@41 152
bos@41 153 If you want to use hooks to so some automated work in a repository
bos@54 154 that a number of people have shared access to, you need to be careful
bos@41 155 in how you do this.
bos@41 156
bos@41 157 Mercurial only locks a repository when it is writing to the
bos@41 158 repository, and only the parts of Mercurial that write to the
bos@41 159 repository pay attention to locks. Write locks are necessary to
bos@41 160 prevent multiple simultaneous writers from scribbling on each other's
bos@41 161 work, corrupting the repository.
bos@41 162
bos@41 163 Because Mercurial is careful with the order in which it reads and
bos@41 164 writes data, it does not need to acquire a lock when it wants to read
bos@41 165 data from the repository. The parts of Mercurial that read from the
bos@41 166 repository never pay attention to locks. This lockless reading scheme
bos@41 167 greatly increases performance and concurrency.
bos@41 168
bos@41 169 With great performance comes a trade-off, though, one which has the
bos@41 170 potential to cause you trouble unless you're aware of it. To describe
bos@41 171 this requires a little detail about how Mercurial adds changesets to a
bos@41 172 repository and reads those changes.
bos@41 173
bos@41 174 When Mercurial \emph{writes} metadata, it writes it straight into the
bos@41 175 destination file. It writes file data first, then manifest data
bos@41 176 (which contains pointers to the new file data), then changelog data
bos@41 177 (which contains pointers to the new manifest data). Before the first
bos@41 178 write to each file, it stores a record of where the end of the file
bos@41 179 was in its transaction log. If the transaction must be rolled back,
bos@54 180 Mercurial simply truncates each file back to the size it was before the
bos@41 181 transaction began.
bos@41 182
bos@41 183 When Mercurial \emph{reads} metadata, it reads the changelog first,
bos@41 184 then everything else. Since a reader will only access parts of the
bos@41 185 manifest or file metadata that it can see in the changelog, it can
bos@41 186 never see partially written data.
bos@41 187
bos@41 188 Some controlling hooks (\hook{pretxncommit} and
bos@41 189 \hook{pretxnchangegroup}) run when a transaction is almost complete.
bos@41 190 All of the metadata has been written, but Mercurial can still roll the
bos@41 191 transaction back and cause the newly-written data to disappear.
bos@41 192
bos@41 193 If one of these hooks runs for long, it opens a window in which a
bos@41 194 reader can see the metadata for changesets that are, strictly
bos@41 195 speaking, not yet permanent. The longer the hook runs, the bigger the
bos@41 196 window.
bos@41 197
bos@41 198 A good use for the \hook{pretxnchangegroup} hook would be to
bos@41 199 automatically build and test incoming changes before they are accepted
bos@41 200 into the repository, so that you can guarantee that nobody can push
bos@41 201 changes to this repository that ``break the build''. But if a client
bos@41 202 can pull changes while they're being tested, the usefulness of the
bos@41 203 test is zero; someone can pull untested changes.
bos@41 204
bos@41 205 The safest answer to this challenge is to set up such a ``gatekeeper''
bos@41 206 repository as \emph{unidirectional}. It can take changes pushed in
bos@41 207 from the outside, but nobody can pull changes from it. Use the
bos@41 208 \hook{preoutgoing} hook to lock it down. Configure a
bos@41 209 \hook{changegroup} hook so that if a build or test succeeds, the hook
bos@41 210 will push the new changes out to another repository that people
bos@41 211 \emph{can} pull from.
bos@41 212
bos@34 213 \section{A short tutorial on using hooks}
bos@34 214 \label{sec:hook:simple}
bos@34 215
bos@34 216 It is easy to write a Mercurial hook. Let's start with a hook that
bos@34 217 runs when you finish a \hgcmd{commit}, and simply prints the hash of
bos@34 218 the changeset you just created. The hook is called \hook{commit}.
bos@34 219
bos@34 220 \begin{figure}[ht]
bos@34 221 \interaction{hook.simple.init}
bos@34 222 \caption{A simple hook that runs when a changeset is committed}
bos@34 223 \label{ex:hook:init}
bos@34 224 \end{figure}
bos@34 225
bos@34 226 All hooks follow the pattern in example~\ref{ex:hook:init}. You add
bos@34 227 an entry to the \rcsection{hooks} section of your \hgrc\. On the left
bos@34 228 is the name of the event to trigger on; on the right is the action to
bos@34 229 take. As you can see, you can run an arbitrary shell command in a
bos@34 230 hook. Mercurial passes extra information to the hook using
bos@34 231 environment variables (look for \envar{HG\_NODE} in the example).
bos@34 232
bos@34 233 \subsection{Performing multiple actions per event}
bos@34 234
bos@34 235 Quite often, you will want to define more than one hook for a
bos@34 236 particular kind of event, as shown in example~\ref{ex:hook:ext}.
bos@34 237 Mercurial lets you do this by adding an \emph{extension} to the end of
bos@34 238 a hook's name. You extend a hook's name by giving the name of the
bos@34 239 hook, followed by a full stop (the ``\texttt{.}'' character), followed
bos@34 240 by some more text of your choosing. For example, Mercurial will run
bos@34 241 both \texttt{commit.foo} and \texttt{commit.bar} when the
bos@34 242 \texttt{commit} event occurs.
bos@34 243
bos@34 244 \begin{figure}[ht]
bos@34 245 \interaction{hook.simple.ext}
bos@34 246 \caption{Defining a second \hook{commit} hook}
bos@34 247 \label{ex:hook:ext}
bos@34 248 \end{figure}
bos@34 249
bos@34 250 To give a well-defined order of execution when there are multiple
bos@34 251 hooks defined for an event, Mercurial sorts hooks by extension, and
bos@34 252 executes the hook commands in this sorted order. In the above
bos@34 253 example, it will execute \texttt{commit.bar} before
bos@34 254 \texttt{commit.foo}, and \texttt{commit} before both.
bos@34 255
bos@34 256 It is a good idea to use a somewhat descriptive extension when you
bos@34 257 define a new hook. This will help you to remember what the hook was
bos@34 258 for. If the hook fails, you'll get an error message that contains the
bos@34 259 hook name and extension, so using a descriptive extension could give
bos@34 260 you an immediate hint as to why the hook failed (see
bos@34 261 section~\ref{sec:hook:perm} for an example).
bos@34 262
bos@34 263 \subsection{Controlling whether an activity can proceed}
bos@34 264 \label{sec:hook:perm}
bos@34 265
bos@34 266 In our earlier examples, we used the \hook{commit} hook, which is
bos@34 267 run after a commit has completed. This is one of several Mercurial
bos@34 268 hooks that run after an activity finishes. Such hooks have no way of
bos@34 269 influencing the activity itself.
bos@34 270
bos@34 271 Mercurial defines a number of events that occur before an activity
bos@34 272 starts; or after it starts, but before it finishes. Hooks that
bos@34 273 trigger on these events have the added ability to choose whether the
bos@34 274 activity can continue, or will abort.
bos@34 275
bos@34 276 The \hook{pretxncommit} hook runs after a commit has all but
bos@34 277 completed. In other words, the metadata representing the changeset
bos@34 278 has been written out to disk, but the transaction has not yet been
bos@34 279 allowed to complete. The \hook{pretxncommit} hook has the ability to
bos@34 280 decide whether the transaction can complete, or must be rolled back.
bos@34 281
bos@34 282 If the \hook{pretxncommit} hook exits with a status code of zero, the
bos@34 283 transaction is allowed to complete; the commit finishes; and the
bos@34 284 \hook{commit} hook is run. If the \hook{pretxncommit} hook exits with
bos@34 285 a non-zero status code, the transaction is rolled back; the metadata
bos@34 286 representing the changeset is erased; and the \hook{commit} hook is
bos@34 287 not run.
bos@34 288
bos@34 289 \begin{figure}[ht]
bos@34 290 \interaction{hook.simple.pretxncommit}
bos@34 291 \caption{Using the \hook{pretxncommit} hook to control commits}
bos@34 292 \label{ex:hook:pretxncommit}
bos@34 293 \end{figure}
bos@34 294
bos@34 295 The hook in example~\ref{ex:hook:pretxncommit} checks that a commit
bos@34 296 comment contains a bug ID. If it does, the commit can complete. If
bos@34 297 not, the commit is rolled back.
bos@34 298
bos@37 299 \section{Writing your own hooks}
bos@37 300
bos@37 301 When you are writing a hook, you might find it useful to run Mercurial
bos@37 302 either with the \hggopt{-v} option, or the \rcitem{ui}{verbose} config
bos@37 303 item set to ``true''. When you do so, Mercurial will print a message
bos@37 304 before it calls each hook.
bos@37 305
bos@37 306 \subsection{Choosing how your hook should run}
bos@37 307 \label{sec:hook:lang}
bos@34 308
bos@34 309 You can write a hook either as a normal program---typically a shell
bos@37 310 script---or as a Python function that is executed within the Mercurial
bos@34 311 process.
bos@34 312
bos@34 313 Writing a hook as an external program has the advantage that it
bos@34 314 requires no knowledge of Mercurial's internals. You can call normal
bos@34 315 Mercurial commands to get any added information you need. The
bos@34 316 trade-off is that external hooks are slower than in-process hooks.
bos@34 317
bos@34 318 An in-process Python hook has complete access to the Mercurial API,
bos@34 319 and does not ``shell out'' to another process, so it is inherently
bos@34 320 faster than an external hook. It is also easier to obtain much of the
bos@34 321 information that a hook requires by using the Mercurial API than by
bos@34 322 running Mercurial commands.
bos@34 323
bos@34 324 If you are comfortable with Python, or require high performance,
bos@34 325 writing your hooks in Python may be a good choice. However, when you
bos@34 326 have a straightforward hook to write and you don't need to care about
bos@34 327 performance (probably the majority of hooks), a shell script is
bos@34 328 perfectly fine.
bos@34 329
bos@37 330 \subsection{Hook parameters}
bos@34 331 \label{sec:hook:param}
bos@34 332
bos@34 333 Mercurial calls each hook with a set of well-defined parameters. In
bos@34 334 Python, a parameter is passed as a keyword argument to your hook
bos@34 335 function. For an external program, a parameter is passed as an
bos@34 336 environment variable.
bos@34 337
bos@34 338 Whether your hook is written in Python or as a shell script, the
bos@37 339 hook-specific parameter names and values will be the same. A boolean
bos@37 340 parameter will be represented as a boolean value in Python, but as the
bos@37 341 number 1 (for ``true'') or 0 (for ``false'') as an environment
bos@37 342 variable for an external hook. If a hook parameter is named
bos@37 343 \texttt{foo}, the keyword argument for a Python hook will also be
bos@51 344 named \texttt{foo}, while the environment variable for an external
bos@51 345 hook will be named \texttt{HG\_FOO}.
bos@37 346
bos@37 347 \subsection{Hook return values and activity control}
bos@37 348
bos@37 349 A hook that executes successfully must exit with a status of zero if
bos@37 350 external, or return boolean ``false'' if in-process. Failure is
bos@37 351 indicated with a non-zero exit status from an external hook, or an
bos@37 352 in-process hook returning boolean ``true''. If an in-process hook
bos@37 353 raises an exception, the hook is considered to have failed.
bos@37 354
bos@37 355 For a hook that controls whether an activity can proceed, zero/false
bos@37 356 means ``allow'', while non-zero/true/exception means ``deny''.
bos@37 357
bos@37 358 \subsection{Writing an external hook}
bos@37 359
bos@37 360 When you define an external hook in your \hgrc\ and the hook is run,
bos@37 361 its value is passed to your shell, which interprets it. This means
bos@37 362 that you can use normal shell constructs in the body of the hook.
bos@37 363
bos@37 364 An executable hook is always run with its current directory set to a
bos@37 365 repository's root directory.
bos@37 366
bos@37 367 Each hook parameter is passed in as an environment variable; the name
bos@37 368 is upper-cased, and prefixed with the string ``\texttt{HG\_}''.
bos@37 369
bos@37 370 With the exception of hook parameters, Mercurial does not set or
bos@37 371 modify any environment variables when running a hook. This is useful
bos@37 372 to remember if you are writing a site-wide hook that may be run by a
bos@37 373 number of different users with differing environment variables set.
bos@37 374 In multi-user situations, you should not rely on environment variables
bos@37 375 being set to the values you have in your environment when testing the
bos@37 376 hook.
bos@37 377
bos@37 378 \subsection{Telling Mercurial to use an in-process hook}
bos@37 379
bos@37 380 The \hgrc\ syntax for defining an in-process hook is slightly
bos@37 381 different than for an executable hook. The value of the hook must
bos@37 382 start with the text ``\texttt{python:}'', and continue with the
bos@37 383 fully-qualified name of a callable object to use as the hook's value.
bos@37 384
bos@37 385 The module in which a hook lives is automatically imported when a hook
bos@37 386 is run. So long as you have the module name and \envar{PYTHONPATH}
bos@37 387 right, it should ``just work''.
bos@37 388
bos@37 389 The following \hgrc\ example snippet illustrates the syntax and
bos@37 390 meaning of the notions we just described.
bos@37 391 \begin{codesample2}
bos@37 392 [hooks]
bos@37 393 commit.example = python:mymodule.submodule.myhook
bos@37 394 \end{codesample2}
bos@37 395 When Mercurial runs the \texttt{commit.example} hook, it imports
bos@37 396 \texttt{mymodule.submodule}, looks for the callable object named
bos@37 397 \texttt{myhook}, and calls it.
bos@37 398
bos@37 399 \subsection{Writing an in-process hook}
bos@37 400
bos@37 401 The simplest in-process hook does nothing, but illustrates the basic
bos@37 402 shape of the hook API:
bos@37 403 \begin{codesample2}
bos@37 404 def myhook(ui, repo, **kwargs):
bos@37 405 pass
bos@37 406 \end{codesample2}
bos@37 407 The first argument to a Python hook is always a
bos@37 408 \pymodclass{mercurial.ui}{ui} object. The second is a repository object;
bos@37 409 at the moment, it is always an instance of
bos@37 410 \pymodclass{mercurial.localrepo}{localrepository}. Following these two
bos@37 411 arguments are other keyword arguments. Which ones are passed in
bos@37 412 depends on the hook being called, but a hook can ignore arguments it
bos@37 413 doesn't care about by dropping them into a keyword argument dict, as
bos@37 414 with \texttt{**kwargs} above.
bos@34 415
bos@44 416 \section{Some hook examples}
bos@44 417
bos@49 418 \subsection{Writing meaningful commit messages}
bos@49 419
bos@49 420 It's hard to imagine a useful commit message being very short. The
bos@49 421 simple \hook{pretxncommit} hook of figure~\ref{ex:hook:msglen.run}
bos@49 422 will prevent you from committing a changeset with a message that is
bos@49 423 less than ten bytes long.
bos@49 424
bos@49 425 \begin{figure}[ht]
bos@49 426 \interaction{hook.msglen.run}
bos@49 427 \caption{A hook that forbids overly short commit messages}
bos@49 428 \label{ex:hook:msglen.run}
bos@49 429 \end{figure}
bos@49 430
bos@49 431 \subsection{Checking for trailing whitespace}
bos@44 432
bos@44 433 An interesting use of a commit-related hook is to help you to write
bos@44 434 cleaner code. A simple example of ``cleaner code'' is the dictum that
bos@44 435 a change should not add any new lines of text that contain ``trailing
bos@44 436 whitespace''. Trailing whitespace is a series of space and tab
bos@44 437 characters at the end of a line of text. In most cases, trailing
bos@44 438 whitespace is unnecessary, invisible noise, but it is occasionally
bos@49 439 problematic, and people often prefer to get rid of it.
bos@44 440
bos@44 441 You can use either the \hook{precommit} or \hook{pretxncommit} hook to
bos@44 442 tell whether you have a trailing whitespace problem. If you use the
bos@44 443 \hook{precommit} hook, the hook will not know which files you are
bos@44 444 committing, so it will have to check every modified file in the
bos@44 445 repository for trailing white space. If you want to commit a change
bos@44 446 to just the file \filename{foo}, but the file \filename{bar} contains
bos@44 447 trailing whitespace, doing a check in the \hook{precommit} hook will
bos@44 448 prevent you from committing \filename{foo} due to the problem with
bos@44 449 \filename{bar}. This doesn't seem right.
bos@44 450
bos@44 451 Should you choose the \hook{pretxncommit} hook, the check won't occur
bos@44 452 until just before the transaction for the commit completes. This will
bos@44 453 allow you to check for problems only the exact files that are being
bos@44 454 committed. However, if you entered the commit message interactively
bos@44 455 and the hook fails, the transaction will roll back; you'll have to
bos@44 456 re-enter the commit message after you fix the trailing whitespace and
bos@44 457 run \hgcmd{commit} again.
bos@44 458
bos@44 459 \begin{figure}[ht]
bos@44 460 \interaction{hook.ws.simple}
bos@44 461 \caption{A simple hook that checks for trailing whitespace}
bos@44 462 \label{ex:hook:ws.simple}
bos@44 463 \end{figure}
bos@44 464
bos@44 465 Figure~\ref{ex:hook:ws.simple} introduces a simple \hook{pretxncommit}
bos@44 466 hook that checks for trailing whitespace. This hook is short, but not
bos@44 467 very helpful. It exits with an error status if a change adds a line
bos@44 468 with trailing whitespace to any file, but does not print any
bos@49 469 information that might help us to identify the offending file or
bos@49 470 line. It also has the nice property of not paying attention to
bos@49 471 unmodified lines; only lines that introduce new trailing whitespace
bos@49 472 cause problems.
bos@49 473
bos@49 474 \begin{figure}[ht]
bos@49 475 \interaction{hook.ws.better}
bos@49 476 \caption{A better trailing whitespace hook}
bos@49 477 \label{ex:hook:ws.better}
bos@49 478 \end{figure}
bos@49 479
bos@49 480 The example of figure~\ref{ex:hook:ws.better} is much more complex,
bos@49 481 but also more useful. It parses a unified diff to see if any lines
bos@49 482 add trailing whitespace, and prints the name of the file and the line
bos@49 483 number of each such occurrence. Even better, if the change adds
bos@49 484 trailing whitespace, this hook saves the commit comment and prints the
bos@49 485 name of the save file before exiting and telling Mercurial to roll the
bos@49 486 transaction back, so you can use
bos@49 487 \hgcmdargs{commit}{\hgopt{commit}{-l}~\emph{filename}} to reuse the
bos@49 488 saved commit message once you've corrected the problem.
bos@49 489
bos@49 490 As a final aside, note in figure~\ref{ex:hook:ws.better} the use of
bos@49 491 \command{perl}'s in-place editing feature to get rid of trailing
bos@49 492 whitespace from a file. This is concise and useful enough that I will
bos@49 493 reproduce it here.
bos@49 494 \begin{codesample2}
bos@54 495 perl -pi -e 's,\\s+\$,,' filename
bos@49 496 \end{codesample2}
bos@49 497
bos@49 498 \section{Bundled hooks}
bos@49 499
bos@49 500 Mercurial ships with several bundled hooks. You can find them in the
bos@49 501 \dirname{hgext} directory of a Mercurial source tree. If you are
bos@49 502 using a Mercurial binary package, the hooks will be located in the
bos@49 503 \dirname{hgext} directory of wherever your package installer put
bos@49 504 Mercurial.
bos@49 505
bos@49 506 \subsection{\hgext{acl}---access control for parts of a repository}
bos@49 507
bos@49 508 The \hgext{acl} extension lets you control which remote users are
bos@49 509 allowed to push changesets to a networked server. You can protect any
bos@49 510 portion of a repository (including the entire repo), so that a
bos@49 511 specific remote user can push changes that do not affect the protected
bos@49 512 portion.
bos@49 513
bos@49 514 This extension implements access control based on the identity of the
bos@49 515 user performing a push, \emph{not} on who committed the changesets
bos@49 516 they're pushing. (If access control based on committer was to work
bos@49 517 properly, it would require commits to be cryptographically signed,
bos@49 518 which is an onerous and hence unusual policy to enforce.)
bos@49 519
bos@49 520 XXX More help.
bos@44 521
bos@39 522 \section{Hook reference}
bos@41 523 \label{sec:hook:ref}
bos@39 524
bos@39 525 \subsection{In-process hook execution}
bos@39 526
bos@39 527 An in-process hook is called with arguments of the following form:
bos@39 528 \begin{codesample2}
bos@39 529 def myhook(ui, repo, **kwargs):
bos@39 530 pass
bos@39 531 \end{codesample2}
bos@39 532 The \texttt{ui} parameter is a \pymodclass{mercurial.ui}{ui} object.
bos@39 533 The \texttt{repo} parameter is a
bos@39 534 \pymodclass{mercurial.localrepo}{localrepository} object. The
bos@39 535 names and values of the \texttt{**kwargs} parameters depend on the
bos@39 536 hook being invoked, with the following common features:
bos@39 537 \begin{itemize}
bos@39 538 \item If a parameter is named \texttt{node} or
bos@39 539 \texttt{parent\emph{N}}, it will contain a hexadecimal changeset ID.
bos@39 540 The empty string is used to represent ``null changeset ID'' instead
bos@39 541 of a string of zeroes.
bos@39 542 \item Boolean-valued parameters are represented as Python
bos@39 543 \texttt{bool} objects.
bos@39 544 \end{itemize}
bos@39 545
bos@39 546 An in-process hook is called without a change to the process's working
bos@39 547 directory (unlike external hooks, which are run in the root of the
bos@39 548 repository). It must not change the process's working directory. If
bos@39 549 it were to do so, it would probably cause calls to the Mercurial API,
bos@39 550 or operations after the hook finishes, to fail.
bos@39 551
bos@39 552 If a hook returns a boolean ``false'' value, it is considered to
bos@39 553 have succeeded. If it returns a boolean ``true'' value or raises an
bos@39 554 exception, it is considered to have failed.
bos@39 555
bos@39 556 \subsection{External hook execution}
bos@39 557
bos@39 558 An external hook is passed to the user's shell for execution, so
bos@39 559 features of that shell, such as variable substitution and command
bos@39 560 redirection, are available. The hook is run in the root directory of
bos@39 561 the repository.
bos@39 562
bos@39 563 Hook parameters are passed to the hook as environment variables. Each
bos@39 564 environment variable's name is converted in upper case and prefixed
bos@39 565 with the string ``\texttt{HG\_}''. For example, if the name of a
bos@39 566 parameter is ``\texttt{node}'', the name of the environment variable
bos@39 567 representing that parameter will be ``\texttt{HG\_NODE}''.
bos@39 568
bos@39 569 A boolean parameter is represented as the string ``\texttt{1}'' for
bos@39 570 ``true'', ``\texttt{0}'' for ``false''. If an environment variable is
bos@39 571 named \envar{HG\_NODE}, \envar{HG\_PARENT1} or \envar{HG\_PARENT2}, it
bos@39 572 contains a changeset ID represented as a hexadecimal string. The
bos@39 573 empty string is used to represent ``null changeset ID'' instead of a
bos@39 574 string of zeroes.
bos@39 575
bos@39 576 If a hook exits with a status of zero, it is considered to have
bos@39 577 succeeded. If it exits with a non-zero status, it is considered to
bos@39 578 have failed.
bos@39 579
bos@39 580 \subsection{The \hook{changegroup} hook}
bos@39 581 \label{sec:hook:changegroup}
bos@39 582
bos@40 583 This hook is run after a group of pre-existing changesets has been
bos@40 584 added to the repository, for example via a \hgcmd{pull} or
bos@40 585 \hgcmd{unbundle}. This hook is run once per operation that added one
bos@41 586 or more changesets. This is in contrast to the \hook{incoming} hook,
bos@41 587 which is run once per changeset, regardless of whether the changesets
bos@41 588 arrive in a group.
bos@41 589
bos@41 590 Some possible uses for this hook include kicking off an automated
bos@41 591 build or test of the added changesets, updating a bug database, or
bos@41 592 notifying subscribers that a repository contains new changes.
bos@40 593
bos@40 594 Parameters to this hook:
bos@40 595 \begin{itemize}
bos@40 596 \item[\texttt{node}] A changeset ID. The changeset ID of the first
bos@40 597 changeset in the group that was added. All changesets between this
bos@40 598 and \index{tags!\texttt{tip}}\texttt{tip}, inclusive, were added by
bos@40 599 a single \hgcmd{pull}, \hgcmd{push} or \hgcmd{unbundle}.
bos@40 600 \end{itemize}
bos@40 601
bos@40 602 See also: \hook{incoming} (section~\ref{sec:hook:incoming}),
bos@40 603 \hook{prechangegroup} (section~\ref{sec:hook:prechangegroup}),
bos@40 604 \hook{pretxnchangegroup} (section~\ref{sec:hook:pretxnchangegroup})
bos@39 605
bos@39 606 \subsection{The \hook{commit} hook}
bos@39 607 \label{sec:hook:commit}
bos@39 608
bos@40 609 This hook is run after a new changeset has been created.
bos@40 610
bos@40 611 Parameters to this hook:
bos@40 612 \begin{itemize}
bos@40 613 \item[\texttt{node}] A changeset ID. The changeset ID of the newly
bos@40 614 committed changeset.
bos@40 615 \item[\texttt{parent1}] A changeset ID. The changeset ID of the first
bos@40 616 parent of the newly committed changeset.
bos@40 617 \item[\texttt{parent2}] A changeset ID. The changeset ID of the second
bos@40 618 parent of the newly committed changeset.
bos@40 619 \end{itemize}
bos@40 620
bos@40 621 See also: \hook{precommit} (section~\ref{sec:hook:precommit}),
bos@40 622 \hook{pretxncommit} (section~\ref{sec:hook:pretxncommit})
bos@40 623
bos@40 624 \subsection{The \hook{incoming} hook}
bos@40 625 \label{sec:hook:incoming}
bos@40 626
bos@40 627 This hook is run after a pre-existing changeset has been added to the
bos@40 628 repository, for example via a \hgcmd{push}. If a group of changesets
bos@40 629 was added in a single operation, this hook is called once for each
bos@40 630 added changeset.
bos@40 631
bos@41 632 You can use this hook for the same purposes as the \hook{changegroup}
bos@41 633 hook (section~\ref{sec:hook:changegroup}); it's simply more convenient
bos@54 634 sometimes to run a hook once per group of changesets, while other
bos@41 635 times it's handier once per changeset.
bos@41 636
bos@40 637 Parameters to this hook:
bos@40 638 \begin{itemize}
bos@40 639 \item[\texttt{node}] A changeset ID. The ID of the newly added
bos@39 640 changeset.
bos@40 641 \end{itemize}
bos@40 642
bos@40 643 See also: \hook{changegroup} (section~\ref{sec:hook:changegroup}) \hook{prechangegroup} (section~\ref{sec:hook:prechangegroup}), \hook{pretxnchangegroup} (section~\ref{sec:hook:pretxnchangegroup})
bos@40 644
bos@40 645 \subsection{The \hook{outgoing} hook}
bos@40 646 \label{sec:hook:outgoing}
bos@40 647
bos@40 648 This hook is run after a group of changesets has been propagated out
bos@40 649 of this repository, for example by a \hgcmd{push} or \hgcmd{bundle}
bos@40 650 command.
bos@40 651
bos@41 652 One possible use for this hook is to notify administrators that
bos@41 653 changes have been pulled.
bos@41 654
bos@40 655 Parameters to this hook:
bos@40 656 \begin{itemize}
bos@40 657 \item[\texttt{node}] A changeset ID. The changeset ID of the first
bos@40 658 changeset of the group that was sent.
bos@40 659 \item[\texttt{source}] A string. The source of the of the operation.
bos@40 660 If a remote client pulled changes from this repository,
bos@40 661 \texttt{source} will be \texttt{serve}. If the client that obtained
bos@40 662 changes from this repository was local, \texttt{source} will be
bos@40 663 \texttt{bundle}, \texttt{pull}, or \texttt{push}, depending on the
bos@40 664 operation the client performed.
bos@40 665 \end{itemize}
bos@40 666
bos@40 667 See also: \hook{preoutgoing} (section~\ref{sec:hook:preoutgoing})
bos@40 668
bos@40 669 \subsection{The \hook{prechangegroup} hook}
bos@40 670 \label{sec:hook:prechangegroup}
bos@40 671
bos@41 672 This controlling hook is run before Mercurial begins to add a group of
bos@41 673 changesets from another repository.
bos@41 674
bos@41 675 This hook does not have any information about the changesets to be
bos@41 676 added, because it is run before transmission of those changesets is
bos@41 677 allowed to begin. If this hook fails, the changesets will not be
bos@41 678 transmitted.
bos@41 679
bos@41 680 One use for this hook is to prevent external changes from being added
bos@41 681 to a repository, for example to ``freeze'' a server-hosted branch
bos@41 682 temporarily or permanently.
bos@41 683
bos@40 684 This hook is not passed any parameters.
bos@40 685
bos@40 686 See also: \hook{changegroup} (section~\ref{sec:hook:changegroup}),
bos@40 687 \hook{incoming} (section~\ref{sec:hook:incoming}), ,
bos@40 688 \hook{pretxnchangegroup} (section~\ref{sec:hook:pretxnchangegroup})
bos@40 689
bos@40 690 \subsection{The \hook{precommit} hook}
bos@40 691 \label{sec:hook:precommit}
bos@40 692
bos@41 693 This hook is run before Mercurial begins to commit a new changeset.
bos@41 694 It is run before Mercurial has any of the metadata for the commit,
bos@41 695 such as the files to be committed, the commit message, or the commit
bos@41 696 date.
bos@41 697
bos@41 698 One use for this hook is to disable the ability to commit new
bos@41 699 changesets, while still allowing incoming changesets. Another is to
bos@41 700 run a build or test, and only allow the commit to begin if the build
bos@41 701 or test succeeds.
bos@40 702
bos@40 703 Parameters to this hook:
bos@40 704 \begin{itemize}
bos@40 705 \item[\texttt{parent1}] A changeset ID. The changeset ID of the first
bos@40 706 parent of the working directory.
bos@40 707 \item[\texttt{parent2}] A changeset ID. The changeset ID of the second
bos@40 708 parent of the working directory.
bos@40 709 \end{itemize}
bos@40 710 If the commit proceeds, the parents of the working directory will
bos@40 711 become the parents of the new changeset.
bos@40 712
bos@40 713 See also: \hook{commit} (section~\ref{sec:hook:commit}),
bos@40 714 \hook{pretxncommit} (section~\ref{sec:hook:pretxncommit})
bos@40 715
bos@40 716 \subsection{The \hook{preoutgoing} hook}
bos@40 717 \label{sec:hook:preoutgoing}
bos@40 718
bos@40 719 This hook is invoked before Mercurial knows the identities of the
bos@40 720 changesets to be transmitted.
bos@40 721
bos@41 722 One use for this hook is to prevent changes from being transmitted to
bos@41 723 another repository.
bos@41 724
bos@40 725 Parameters to this hook:
bos@40 726 \begin{itemize}
bos@40 727 \item[\texttt{source}] A string. The source of the operation that is
bos@40 728 attempting to obtain changes from this repository. See the
bos@40 729 documentation for the \texttt{source} parameter to the
bos@40 730 \hook{outgoing} hook, in section~\ref{sec:hook:outgoing}, for
bos@40 731 possible values of this parameter..
bos@40 732 \end{itemize}
bos@40 733
bos@40 734 See also: \hook{outgoing} (section~\ref{sec:hook:outgoing})
bos@40 735
bos@40 736 \subsection{The \hook{pretag} hook}
bos@40 737 \label{sec:hook:pretag}
bos@40 738
bos@41 739 This controlling hook is run before a tag is created. If the hook
bos@41 740 succeeds, creation of the tag proceeds. If the hook fails, the tag is
bos@41 741 not created.
bos@41 742
bos@40 743 Parameters to this hook:
bos@40 744 \begin{itemize}
bos@40 745 \item[\texttt{local}] A boolean. Whether the tag is local to this
bos@40 746 repository instance (i.e.~stored in \sfilename{.hg/tags}) or managed
bos@40 747 by Mercurial (stored in \sfilename{.hgtags}).
bos@40 748 \item[\texttt{node}] A changeset ID. The ID of the changeset to be tagged.
bos@40 749 \item[\texttt{tag}] A string. The name of the tag to be created.
bos@40 750 \end{itemize}
bos@40 751
bos@40 752 If the tag to be created is revision-controlled, the \hook{precommit}
bos@40 753 and \hook{pretxncommit} hooks (sections~\ref{sec:hook:commit}
bos@40 754 and~\ref{sec:hook:pretxncommit}) will also be run.
bos@40 755
bos@40 756 See also: \hook{tag} (section~\ref{sec:hook:tag})
bos@40 757
bos@40 758 \subsection{The \hook{pretxnchangegroup} hook}
bos@40 759 \label{sec:hook:pretxnchangegroup}
bos@40 760
bos@41 761 This controlling hook is run before a transaction---that manages the
bos@41 762 addition of a group of new changesets from outside the
bos@41 763 repository---completes. If the hook succeeds, the transaction
bos@41 764 completes, and all of the changesets become permanent within this
bos@41 765 repository. If the hook fails, the transaction is rolled back, and
bos@41 766 the data for the changesets is erased.
bos@41 767
bos@41 768 This hook can access the metadata associated with the almost-added
bos@41 769 changesets, but it should not do anything permanent with this data.
bos@41 770 It must also not modify the working directory.
bos@41 771
bos@41 772 While this hook is running, if other Mercurial processes access this
bos@41 773 repository, they will be able to see the almost-added changesets as if
bos@41 774 they are permanent. This may lead to race conditions if you do not
bos@41 775 take steps to avoid them.
bos@41 776
bos@41 777 This hook can be used to automatically vet a group of changesets. If
bos@41 778 the hook fails, all of the changesets are ``rejected'' when the
bos@41 779 transaction rolls back.
bos@41 780
bos@40 781 Parameters to this hook are the same as for the \hook{changegroup}
bos@40 782 hook; see section~\ref{sec:hook:changegroup} for details.
bos@40 783
bos@40 784 See also: \hook{changegroup} (section~\ref{sec:hook:changegroup}),
bos@40 785 \hook{incoming} (section~\ref{sec:hook:incoming}),
bos@40 786 \hook{prechangegroup} (section~\ref{sec:hook:prechangegroup})
bos@40 787
bos@40 788 \subsection{The \hook{pretxncommit} hook}
bos@40 789 \label{sec:hook:pretxncommit}
bos@40 790
bos@41 791 This controlling hook is run before a transaction---that manages a new
bos@41 792 commit---completes. If the hook succeeds, the transaction completes
bos@41 793 and the changeset becomes permanent within this repository. If the
bos@41 794 hook fails, the transaction is rolled back, and the commit data is
bos@41 795 erased.
bos@41 796
bos@41 797 This hook can access the metadata associated with the almost-new
bos@41 798 changeset, but it should not do anything permanent with this data. It
bos@41 799 must also not modify the working directory.
bos@41 800
bos@41 801 While this hook is running, if other Mercurial processes access this
bos@41 802 repository, they will be able to see the almost-new changeset as if it
bos@41 803 is permanent. This may lead to race conditions if you do not take
bos@41 804 steps to avoid them.
bos@41 805
bos@40 806 Parameters to this hook are the same as for the \hook{commit} hook;
bos@40 807 see section~\ref{sec:hook:commit} for details.
bos@40 808
bos@40 809 See also: \hook{precommit} (section~\ref{sec:hook:precommit})
bos@40 810
bos@40 811 \subsection{The \hook{preupdate} hook}
bos@40 812 \label{sec:hook:preupdate}
bos@40 813
bos@41 814 This controlling hook is run before an update or merge of the working
bos@41 815 directory begins. It is run only if Mercurial's normal pre-update
bos@41 816 checks determine that the update or merge can proceed. If the hook
bos@41 817 succeeds, the update or merge may proceed; if it fails, the update or
bos@41 818 merge does not start.
bos@41 819
bos@40 820 Parameters to this hook:
bos@40 821 \begin{itemize}
bos@40 822 \item[\texttt{parent1}] A changeset ID. The ID of the parent that the
bos@40 823 working directory is to be updated to. If the working directory is
bos@40 824 being merged, it will not change this parent.
bos@40 825 \item[\texttt{parent2}] A changeset ID. Only set if the working
bos@40 826 directory is being merged. The ID of the revision that the working
bos@40 827 directory is being merged with.
bos@40 828 \end{itemize}
bos@40 829
bos@40 830 See also: \hook{update} (section~\ref{sec:hook:update})
bos@40 831
bos@40 832 \subsection{The \hook{tag} hook}
bos@40 833 \label{sec:hook:tag}
bos@40 834
bos@41 835 This hook is run after a tag has been created.
bos@41 836
bos@40 837 Parameters to this hook are the same as for the \hook{pretag} hook;
bos@40 838 see section~\ref{sec:hook:pretag} for details.
bos@40 839
bos@40 840 If the created tag is revision-controlled, the \hook{commit} hook
bos@41 841 (section~\ref{sec:hook:commit}) is run before this hook.
bos@40 842
bos@40 843 See also: \hook{pretag} (section~\ref{sec:hook:pretag})
bos@40 844
bos@40 845 \subsection{The \hook{update} hook}
bos@40 846 \label{sec:hook:update}
bos@40 847
bos@41 848 This hook is run after an update or merge of the working directory
bos@41 849 completes. Since a merge can fail (if the external \command{hgmerge}
bos@41 850 command fails to resolve conflicts in a file), this hook communicates
bos@41 851 whether the update or merge completed cleanly.
bos@41 852
bos@40 853 \begin{itemize}
bos@40 854 \item[\texttt{error}] A boolean. Indicates whether the update or
bos@40 855 merge completed successfully.
bos@40 856 \item[\texttt{parent1}] A changeset ID. The ID of the parent that the
bos@40 857 working directory was updated to. If the working directory was
bos@40 858 merged, it will not have changed this parent.
bos@40 859 \item[\texttt{parent2}] A changeset ID. Only set if the working
bos@40 860 directory was merged. The ID of the revision that the working
bos@40 861 directory was merged with.
bos@40 862 \end{itemize}
bos@40 863
bos@40 864 See also: \hook{preupdate} (section~\ref{sec:hook:preupdate})
bos@34 865
bos@34 866 %%% Local Variables:
bos@34 867 %%% mode: latex
bos@34 868 %%% TeX-master: "00book"
bos@34 869 %%% End: