hgbook
view en/mq.tex @ 19:187702df428b
Piles of new content for MQ chapter - cookbook stuff.
| author | Bryan O'Sullivan <bos@serpentine.com> |
|---|---|
| date | Fri Jul 07 19:56:53 2006 -0700 (2006-07-07) |
| parents | e6f4088ebe52 |
| children | 9d5b6d303ef5 |
line source
1 \chapter{Managing change with Mercurial Queues}
2 \label{chap:mq}
4 \section{The patch management problem}
5 \label{sec:mq:patch-mgmt}
7 Here is a common scenario: you need to install a software package from
8 source, but you find a bug that you must fix in the source before you
9 can start using the package. You make your changes, forget about the
10 package for a while, and a few months later you need to upgrade to a
11 newer version of the package. If the newer version of the package
12 still has the bug, you must extract your fix from the older source
13 tree and apply it against the newer version. This is a tedious task,
14 and it's easy to make mistakes.
16 This is a simple case of the ``patch management'' problem. You have
17 an ``upstream'' source tree that you can't change; you need to make
18 some local changes on top of the upstream tree; and you'd like to be
19 able to keep those changes separate, so that you can apply them to
20 newer versions of the upstream source.
22 The patch management problem arises in many situations. Probably the
23 most visible is that a user of an open source software project will
24 contribute a bug fix or new feature to the project's maintainers in the
25 form of a patch.
27 Distributors of operating systems that include open source software
28 often need to make changes to the packages they distribute so that
29 they will build properly in their environments.
31 When you have few changes to maintain, it is easy to manage a single
32 patch using the standard \texttt{diff} and \texttt{patch} programs
33 (see section~\ref{sec:mq:patch} for a discussion of these tools).
34 Once the number of changes grows, it starts to makes sense to maintain
35 patches as discrete ``chunks of work,'' so that for example a single
36 patch will contain only one bug fix (the patch might modify several
37 files, but it's doing ``only one thing''), and you may have a number
38 of such patches for different bugs you need fixed and local changes
39 you require. In this situation, if you submit a bug fix patch to the
40 upstream maintainers of a package and they include your fix in a
41 subsequent release, you can simply drop that single patch when you're
42 updating to the newer release.
44 Maintaining a single patch against an upstream tree is a little
45 tedious and error-prone, but not difficult. However, the complexity
46 of the problem grows rapidly as the number of patches you have to
47 maintain increases. With more than a tiny number of patches in hand,
48 understanding which ones you have applied and maintaining them moves
49 from messy to overwhelming.
51 Fortunately, Mercurial includes a powerful extension, Mercurial Queues
52 (or simply ``MQ''), that massively simplifies the patch management
53 problem.
55 \section{The prehistory of Mercurial Queues}
56 \label{sec:mq:history}
58 During the late 1990s, several Linux kernel developers started to
59 maintain ``patch series'' that modified the behaviour of the Linux
60 kernel. Some of these series were focused on stability, some on
61 feature coverage, and others were more speculative.
63 The sizes of these patch series grew rapidly. In 2002, Andrew Morton
64 published some shell scripts he had been using to automate the task of
65 managing his patch queues. Andrew was successfully using these
66 scripts to manage hundreds (sometimes thousands) of patches on top of
67 the Linux kernel.
69 \subsection{A patchwork quilt}
70 \label{sec:mq:quilt}
73 In early 2003, Andreas Gruenbacher and Martin Quinson borrowed the
74 approach of Andrew's scripts and published a tool called ``patchwork
75 quilt''~\cite{web:quilt}, or simply ``quilt''
76 (see~\cite{gruenbacher:2005} for a paper describing it). Because
77 quilt substantially automated patch management, it rapidly gained a
78 large following among open source software developers.
80 Quilt manages a \emph{stack of patches} on top of a directory tree.
81 To begin, you tell quilt to manage a directory tree; it stores away
82 the names and contents of all files in the tree. To fix a bug, you
83 create a new patch (using a single command), edit the files you need
84 to fix, then ``refresh'' the patch.
86 The refresh step causes quilt to scan the directory tree; it updates
87 the patch with all of the changes you have made. You can create
88 another patch on top of the first, which will track the changes
89 required to modify the tree from ``tree with one patch applied'' to
90 ``tree with two patches applied''.
92 You can \emph{change} which patches are applied to the tree. If you
93 ``pop'' a patch, the changes made by that patch will vanish from the
94 directory tree. Quilt remembers which patches you have popped,
95 though, so you can ``push'' a popped patch again, and the directory
96 tree will be restored to contain the modifications in the patch. Most
97 importantly, you can run the ``refresh'' command at any time, and the
98 topmost applied patch will be updated. This means that you can, at
99 any time, change both which patches are applied and what
100 modifications those patches make.
102 Quilt knows nothing about revision control tools, so it works equally
103 well on top of an unpacked tarball or a Subversion repository.
105 \subsection{From patchwork quilt to Mercurial Queues}
106 \label{sec:mq:quilt-mq}
108 In mid-2005, Chris Mason took the features of quilt and wrote an
109 extension that he called Mercurial Queues, which added quilt-like
110 behaviour to Mercurial.
112 The key difference between quilt and MQ is that quilt knows nothing
113 about revision control systems, while MQ is \emph{integrated} into
114 Mercurial. Each patch that you push is represented as a Mercurial
115 changeset. Pop a patch, and the changeset goes away.
117 This integration makes understanding patches and debugging their
118 effects \emph{enormously} easier. Since every applied patch has an
119 associated changeset, you can use \hgcmdargs{log}{\emph{filename}} to
120 see which changesets and patches affected a file. You can use the
121 \hgext{bisect} extension to binary-search through all changesets and
122 applied patches to see where a bug got introduced or fixed. You can
123 use the \hgcmd{annotate} command to see which changeset or patch
124 modified a particular line of a source file. And so on.
126 Because quilt does not care about revision control tools, it is still
127 a tremendously useful piece of software to know about for situations
128 where you cannot use Mercurial and MQ.
130 \section{Understanding patches}
132 Because MQ doesn't hide its patch-oriented nature, it is helpful to
133 understand what patches are, and a little about the tools that work
134 with them.
136 The traditional Unix \command{diff} command compares two files, and
137 prints a list of differences between them. The \command{patch} command
138 understands these differences as \emph{modifications} to make to a
139 file. Take a look at figure~\ref{ex:mq:diff} for a simple example of
140 these commands in action.
142 \begin{figure}[ht]
143 \interaction{mq.diff.diff}
144 \caption{Simple uses of the \command{diff} and \command{patch} commands}
145 \label{ex:mq:diff}
146 \end{figure}
148 The type of file that \command{diff} generates (and \command{patch}
149 takes as input) is called a ``patch'' or a ``diff''; there is no
150 difference between a patch and a diff. (We'll use the term ``patch'',
151 since it's more commonly used.)
153 A patch file can start with arbitrary text; the \command{patch}
154 command ignores this text, but MQ uses it as the commit message when
155 creating changesets. To find the beginning of the patch content,
156 \command{patch} searches for the first line that starts with the
157 string ``\texttt{diff~-}''.
159 MQ works with \emph{unified} diffs (\command{patch} can accept several
160 other diff formats, but MQ doesn't). A unified diff contains two
161 kinds of header. The \emph{file header} describes the file being
162 modified; it contains the name of the file to modify. When
163 \command{patch} sees a new file header, it looks for a file with that
164 name to start modifying.
166 After the file header comes a series of \emph{hunks}. Each hunk
167 starts with a header; this identifies the range of line numbers within
168 the file that the hunk should modify. Following the header, a hunk
169 starts and ends with a few (usually three) lines of text from the
170 unmodified file; these are called the \emph{context} for the hunk. If
171 there's only a small amount of context between successive hunks,
172 \command{diff} doesn't print a new hunk header; it just runs the hunks
173 together, with a few lines of context between modifications.
175 Each line of context begins with a space character. Within the hunk,
176 a line that begins with ``\texttt{-}'' means ``remove this line,''
177 while a line that begins with ``\texttt{+}'' means ``insert this
178 line.'' For example, a line that is modified is represented by one
179 deletion and one insertion.
181 We will return to ome of the more subtle aspects of patches later (in
182 section~\ref{ex:mq:adv-patch}), but you should have enough information
183 now to use MQ.
185 \section{Getting started with Mercurial Queues}
186 \label{sec:mq:start}
188 Because MQ is implemented as an extension, you must explicitly enable
189 before you can use it. (You don't need to download anything; MQ ships
190 with the standard Mercurial distribution.) To enable MQ, edit your
191 \tildefile{.hgrc} file, and add the lines in figure~\ref{ex:mq:config}.
193 \begin{figure}[ht]
194 \begin{codesample4}
195 [extensions]
196 hgext.mq =
197 \end{codesample4}
198 \label{ex:mq:config}
199 \caption{Contents to add to \tildefile{.hgrc} to enable the MQ extension}
200 \end{figure}
202 Once the extension is enabled, it will make a number of new commands
203 available. To verify that the extension is working, you can use
204 \hgcmd{help} to see if the \hgcmd{qinit} command is now available; see
205 the example in figure~\ref{ex:mq:enabled}.
207 \begin{figure}[ht]
208 \interaction{mq.qinit-help.help}
209 \caption{How to verify that MQ is enabled}
210 \label{ex:mq:enabled}
211 \end{figure}
213 You can use MQ with \emph{any} Mercurial repository, and its commands
214 only operate within that repository. To get started, simply prepare
215 the repository using the \hgcmd{qinit} command (see
216 figure~\ref{ex:mq:qinit}). This command creates an empty directory
217 called \sdirname{.hg/patches}, where MQ will keep its metadata. As
218 with many Mercurial commands, the \hgcmd{qinit} command prints nothing
219 if it succeeds.
221 \begin{figure}[ht]
222 \interaction{mq.tutorial.qinit}
223 \caption{Preparing a repository for use with MQ}
224 \label{ex:mq:qinit}
225 \end{figure}
227 \begin{figure}[ht]
228 \interaction{mq.tutorial.qnew}
229 \caption{Creating a new patch}
230 \label{ex:mq:qnew}
231 \end{figure}
233 \subsection{Creating a new patch}
235 To begin work on a new patch, use the \hgcmd{qnew} command. This
236 command takes one argument, the name of the patch to create. MQ will
237 use this as the name of an actual file in the \sdirname{.hg/patches}
238 directory, as you can see in figure~\ref{ex:mq:qnew}.
240 Also newly present in the \sdirname{.hg/patches} directory are two
241 other files, \sfilename{series} and \sfilename{status}. The
242 \sfilename{series} file lists all of the patches that MQ knows about
243 for this repository, with one patch per line. Mercurial uses the
244 \sfilename{status} file for internal book-keeping; it tracks all of the
245 patches that MQ has \emph{applied} in this repository.
247 \begin{note}
248 You may sometimes want to edit the \sfilename{series} file by hand;
249 for example, to change the sequence in which some patches are
250 applied. However, manually editing the \sfilename{status} file is
251 almost always a bad idea, as it's easy to corrupt MQ's idea of what
252 is happening.
253 \end{note}
255 Once you have created your new patch, you can edit files in the
256 working directory as you usually would. All of the normal Mercurial
257 commands, such as \hgcmd{diff} and \hgcmd{annotate}, work exactly as
258 they did before.
260 \subsection{Refreshing a patch}
262 When you reach a point where you want to save your work, use the
263 \hgcmd{qrefresh} command (figure~\ref{ex:mq:qnew}) to update the patch
264 you are working on. This command folds the changes you have made in
265 the working directory into your patch, and updates its corresponding
266 changeset to contain those changes.
268 \begin{figure}[ht]
269 \interaction{mq.tutorial.qrefresh}
270 \caption{Refreshing a patch}
271 \label{ex:mq:qrefresh}
272 \end{figure}
274 You can run \hgcmd{qrefresh} as often as you like, so it's a good way
275 to ``checkpoint'' your work. Refresh your patch at an opportune
276 time; try an experiment; and if the experiment doesn't work out,
277 \hgcmd{revert} your modifications back to the last time you refreshed.
279 \begin{figure}[ht]
280 \interaction{mq.tutorial.qrefresh2}
281 \caption{Refresh a patch many times to accumulate changes}
282 \label{ex:mq:qrefresh2}
283 \end{figure}
285 \subsection{Stacking and tracking patches}
287 Once you have finished working on a patch, or need to work on another,
288 you can use the \hgcmd{qnew} command again to create a new patch.
289 Mercurial will apply this patch on top of your existing patch. See
290 figure~\ref{ex:mq:qnew2} for an example. Notice that the patch
291 contains the changes in our prior patch as part of its context (you
292 can see this more clearly in the output of \hgcmd{annotate}).
294 \begin{figure}[ht]
295 \interaction{mq.tutorial.qnew2}
296 \caption{Stacking a second patch on top of the first}
297 \label{ex:mq:qnew2}
298 \end{figure}
300 So far, with the exception of \hgcmd{qnew} and \hgcmd{qrefresh}, we've
301 been careful to only use regular Mercurial commands. However, there
302 are more ``natural'' commands you can use when thinking about patches
303 with MQ, as illustrated in figure~\ref{ex:mq:qseries}:
305 \begin{itemize}
306 \item The \hgcmd{qseries} command lists every patch that MQ knows
307 about in this repository, from oldest to newest (most recently
308 \emph{created}).
309 \item The \hgcmd{qapplied} command lists every patch that MQ has
310 \emph{applied} in this repository, again from oldest to newest (most
311 recently applied).
312 \end{itemize}
314 \begin{figure}[ht]
315 \interaction{mq.tutorial.qseries}
316 \caption{Understanding the patch stack with \hgcmd{qseries} and
317 \hgcmd{qapplied}}
318 \label{ex:mq:qseries}
319 \end{figure}
321 \subsection{Manipulating the patch stack}
323 The previous discussion implied that there must be a difference
324 between ``known'' and ``applied'' patches, and there is. MQ can
325 manage a patch without it being applied in the repository.
327 An \emph{applied} patch has a corresponding changeset in the
328 repository, and the effects of the patch and changeset are visible in
329 the working directory. You can undo the application of a patch using
330 the \hgcmd{qpop} command. MQ still \emph{knows about}, or manages, a
331 popped patch, but the patch no longer has a corresponding changeset in
332 the repository, and the working directory does not contain the changes
333 made by the patch. Figure~\ref{fig:mq:stack} illustrates the
334 difference between applied and tracked patches.
336 \begin{figure}[ht]
337 \centering
338 \grafix{mq-stack}
339 \caption{Applied and unapplied patches in the MQ patch stack}
340 \label{fig:mq:stack}
341 \end{figure}
343 You can reapply an unapplied, or popped, patch using the \hgcmd{qpush}
344 command. This creates a new changeset to correspond to the patch, and
345 the patch's changes once again become present in the working
346 directory. See figure~\ref{ex:mq:qpop} for examples of \hgcmd{qpop}
347 and \hgcmd{qpush} in action. Notice that once we have popped a patch
348 or two patches, the output of \hgcmd{qseries} remains the same, while
349 that of \hgcmd{qapplied} has changed.
351 \begin{figure}[ht]
352 \interaction{mq.tutorial.qpop}
353 \caption{Modifying the stack of applied patches}
354 \label{ex:mq:qpop}
355 \end{figure}
357 MQ does not limit you to pushing or popping one patch. You can have
358 no patches, all of them, or any number in between applied at some
359 point in time.
361 \subsection{Working on several patches at once}
363 The \hgcmd{qrefresh} command always refreshes the \emph{topmost}
364 applied patch. This means that you can suspend work on one patch (by
365 refreshing it), pop or push to make a different patch the top, and
366 work on \emph{that} patch for a while.
368 Here's an example that illustrates how you can use this ability.
369 Let's say you're developing a new feature as two patches. The first
370 is a change to the core of your software, and the second---layered on
371 top of the first---changes the user interface to use the code you just
372 added to the core. If you notice a bug in the core while you're
373 working on the UI patch, it's easy to fix the core. Simply
374 \hgcmd{qrefresh} the UI patch to save your in-progress changes, and
375 \hgcmd{qpop} down to the core patch. Fix the core bug,
376 \hgcmd{qrefresh} the core patch, and \hgcmd{qpush} back to the UI
377 patch to continue where you left off.
379 \section{More about patches}
380 \label{sec:mq:adv-patch}
382 MQ uses the GNU \command{patch} command to apply patches, so it's
383 helpful to know about a few more detailed aspects of how
384 \command{patch} works.
386 When \command{patch} applies a hunk, it tries a handful of
387 successively less accurate strategies to try to make the hunk apply.
388 This falling-back technique often makes it possible to take a patch
389 that was generated against an old version of a file, and apply it
390 against a newer version of that file.
392 First, \command{patch} tries an exact match, where the line numbers,
393 the context, and the text to be modified must apply exactly. If it
394 cannot make an exact match, it tries to find an exact match for the
395 context, without honouring the line numbering information. If this
396 succeeds, it prints a line of output saying that the hunk was applied,
397 but at some \emph{offset} from the original line number.
399 If a context-only match fails, \command{patch} removes the first and
400 last lines of the context, and tries a \emph{reduced} context-only
401 match. If the hunk with reduced context succeeds, it prints a message
402 saying that it applied the hunk with a \emph{fuzz factor} (the number
403 after the fuzz factor indicates how many lines of context
404 \command{patch} had to trim before the patch applied).
406 When neither of these techniques works, \command{patch} prints a
407 message saying that the hunk in question was rejected. It saves
408 rejected hunks (also simply called ``rejects'') to a file with the
409 same name, and an added \sfilename{.rej} extension. It also saves an
410 unmodified copy of the file with a \sfilename{.orig} extension; the
411 copy of the file without any extensions will contain any changes made
412 by hunks that \emph{did} apply cleanly. If you have a patch that
413 modifies \filename{foo} with six hunks, and one of them fails to
414 apply, you will have: an unmodified \filename{foo.orig}, a
415 \filename{foo.rej} containing one hunk, and \filename{foo}, containing
416 the changes made by the five successful five hunks.
418 \subsection{Beware the fuzz}
420 While applying a hunk at an offset, or with a fuzz factor, will often
421 be completely successful, these inexact techniques naturally leave
422 open the possibility of corrupting the patched file. The most common
423 cases typically involve applying a patch twice, or at an incorrect
424 location in the file. If \command{patch} or \hgcmd{qpush} ever
425 mentions an offset or fuzz factor, you should make sure that the
426 modified files are correct afterwards.
428 It's often a good idea to refresh a patch that has applied with an
429 offset or fuzz factor; refreshing the patch generates new context
430 information that will make it apply cleanly. I say ``often,'' not
431 ``always,'' because sometimes refreshing a patch will make it fail to
432 apply against a different revision of the underlying files. In some
433 cases, such as when you're maintaining a patch that must sit on top of
434 multiple versions of a source tree, it's acceptable to have a patch
435 apply with some fuzz, provided you've verified the results of the
436 patching process in such cases.
438 \subsection{Handling rejection}
440 If \hgcmd{qpush} fails to apply a patch, it will print an error
441 message and exit. If it has left \sfilename{.rej} files behind, it is
442 usually best to fix up the rejected hunks before you push more patches
443 or do any further work.
445 If your patch \emph{used to} apply cleanly, and no longer does because
446 you've changed the underlying code that your patches are based on,
447 Mercurial Queues can help; see section~\ref{sec:mq:merge} for details.
449 Unfortunately, there aren't any great techniques for dealing with
450 rejected hunks. Most often, you'll need to view the \sfilename{.rej}
451 file and edit the target file, applying the rejected hunks by hand.
453 If you're feeling adventurous, Neil Brown, a Linux kernel hacker,
454 wrote a tool called \command{wiggle}~\cite{web:wiggle}, which is more
455 vigorous than \command{patch} in its attempts to make a patch apply.
457 Another Linux kernel hacker, Chris Mason (the author of Mercurial
458 Queues), wrote a similar tool called \command{rej}~\cite{web:rej},
459 which takes a simple approach to automating the application of hunks
460 rejected by \command{patch}. \command{rej} can help with four common
461 reasons that a hunk may be rejected:
463 \begin{itemize}
464 \item The context in the middle of a hunk has changed.
465 \item A hunk is missing some context at the beginning or end.
466 \item A large hunk might apply better---either entirely or in
467 part---if it was broken up into smaller hunks.
468 \item A hunk removes lines with slightly different content than those
469 currently present in the file.
470 \end{itemize}
472 If you use \command{wiggle} or \command{rej}, you should be doubly
473 careful to check your results when you're done.
475 \section{Getting the best performance out of MQ}
477 MQ is very efficient at handling a large number of patches. I ran
478 some performance experiments in mid-2006 for a talk that I gave at the
479 2006 EuroPython conference~\cite{web:europython}. I used as my data
480 set the Linux 2.6.17-mm1 patch series, which consists of 1,738
481 patches. I applied thes on top of a Linux kernel repository
482 containing all 27,472 revisions between Linux 2.6.12-rc2 and Linux
483 2.6.17.
485 On my old, slow laptop, I was able to
486 \hgcmdargs{qpush}{\hgopt{qpush}{-a}} all 1,738 patches in 3.5 minutes,
487 and \hgcmdargs{qpop}{\hgopt{qpop}{-a}} them all in 30 seconds. I
488 could \hgcmd{qrefresh} one of the biggest patches (which made 22,779
489 lines of changes to 287 files) in 6.6 seconds.
491 Clearly, MQ is well suited to working in large trees, but there are a
492 few tricks you can use to get the best performance of it.
494 First of all, try to ``batch'' operations together. Every time you
495 run \hgcmd{qpush} or \hgcmd{qpop}, these commands scan the working
496 directory once to make sure you haven't made some changes and then
497 forgotten to run \hgcmd{qrefresh}. On a small tree, the time that
498 this scan takes is unnoticeable. However, on a medium-sized tree
499 (containing tens of thousands of files), it can take a second or more.
501 The \hgcmd{qpush} and \hgcmd{qpop} commands allow you to push and pop
502 multiple patches at a time. You can identify the ``destination
503 patch'' that you want to end up at. When you \hgcmd{qpush} with a
504 destination specified, it will push patches until that patch is at the
505 top of the applied stack. When you \hgcmd{qpop} to a destination, MQ
506 will pop patches until the destination patch \emph{is no longer}
507 applied.
509 You can identify a destination patch using either the name of the
510 patch, or by number. If you use numeric addressing, patches are
511 counted from zero; this means that the first patch is zero, the second
512 is one, and so on.
514 \section{Updating your patches when the underlying code changes}
515 \label{sec:mq:merge}
517 It's common to have a stack of patches on top of an underlying
518 repository that you don't modify directly. If you're working on
519 changes to third-party code, or on a feature that is taking longer to
520 develop than the rate of change of the code beneath, you will often
521 need to sync up with the underlying code, and fix up any hunks in your
522 patches that no longer apply. This is called \emph{rebasing} your
523 patch series.
525 The simplest way to do this is to \hgcmdargs{qpop}{\hgopt{qpop}{-a}}
526 your patches, then \hgcmd{pull} changes into the underlying
527 repository, and finally \hgcmdargs{qpush}{\hgopt{qpop}{-a}} your
528 patches again. MQ will stop pushing any time it runs across a patch
529 that fails to apply during conflicts, allowing you to fix your
530 conflicts, \hgcmd{qrefresh} the affected patch, and continue pushing
531 until you have fixed your entire stack.
533 This approach is easy to use and works well if you don't expect
534 changes to the underlying code to affect how well your patches apply.
535 If your patch stack touches code that is modified frequently or
536 invasively in the underlying repository, however, fixing up rejected
537 hunks by hand quickly becomes tiresome.
539 It's possible to partially automate the rebasing process. If your
540 patches apply cleanly against some revision of the underlying repo, MQ
541 can use this information to help you to resolve conflicts between your
542 patches and a different revision.
544 The process is a little involved.
545 \begin{enumerate}
546 \item To begin, \hgcmdargs{qpush}{-a} all of your patches on top of
547 the revision where you know that they apply cleanly.
548 \item Save a backup copy of your patch directory using
549 \hgcmdargs{qsave}{\hgopt{qsave}{-e} \hgopt{qsave}{-c}}. This prints
550 the name of the directory that it has saved the patches in. It will
551 save the patches to a directory called
552 \sdirname{.hg/patches.\emph{N}}, where \texttt{\emph{N}} is a small
553 integer. It also commits a ``save changeset'' on top of your
554 applied patches; this is for internal book-keeping, and records the
555 states of the \sfilename{series} and \sfilename{status} files.
556 \item Use \hgcmd{pull} to bring new changes into the underlying
557 repository. (Don't run \hgcmdargs{pull}{-u}; see below for why.)
558 \item Update to the new tip revision, using
559 \hgcmdargs{update}{\hgopt{update}{-C}} to override the patches you
560 have pushed.
561 \item Merge all patches using \hgcmdargs{qpush}{\hgopt{qpush}{-m}
562 \hgopt{qpush}{-a}}. The \hgopt{qpush}{-m} option to \hgcmd{qpush}
563 tells MQ to perform a three-way merge if the patch fails to apply.
564 \end{enumerate}
566 During the \hgcmdargs{qpush}{\hgopt{qpush}{-m}}, each patch in the
567 \sfilename{series} file is applied normally. If a patch applies with
568 fuzz or rejects, MQ looks at the queue you \hgcmd{qsave}d, and
569 performs a three-way merge with the corresponding changeset. This
570 merge uses Mercurial's normal merge machinery, so it may pop up a GUI
571 merge tool to help you to resolve problems.
573 When you finish resolving the effects of a patch, MQ refreshes your
574 patch based on the result of the merge.
576 At the end of this process, your repository will have one extra head
577 from the old patch queue, and a copy of the old patch queue will be in
578 \sdirname{.hg/patches.\emph{N}}. You can remove the extra head using
579 \hgcmdargs{qpop}{\hgopt{qpop}{-a} \hgopt{qpop}{-n} patches.\emph{N}}
580 or \hgcmd{strip}. You can delete \sdirname{.hg/patches.\emph{N}} once
581 you are sure that you no longer need it as a backup.
583 \section{Managing patches in a repository}
585 Because MQ's \sdirname{.hg/patches} directory resides outside a
586 Mercurial repository's working directory, the ``underlying'' Mercurial
587 repository knows nothing about the management or presence of patches.
589 This presents the interesting possibility of managing the contents of
590 the patch directory as a Mercurial repository in its own right. This
591 can be a useful way to work. For example, you can work on a patch for
592 a while, \hgcmd{qrefresh} it, then \hgcmd{commit} the current state of
593 the patch. This lets you ``roll back'' to that version of the patch
594 later on.
596 In addition, you can then share different versions of the same patch
597 stack among multiple underlying repositories. I use this when I am
598 developing a Linux kernel feature. I have a pristine copy of my
599 kernel sources for each of several CPU architectures, and a cloned
600 repository under each that contains the patches I am working on. When
601 I want to test a change on a different architecture, I push my current
602 patches to the patch repository associated with that kernel tree, pop
603 and push all of my patches, and build and test that kernel.
605 Managing patches in a repository makes it possible for multiple
606 developers to work on the same patch series without colliding with
607 each other, all on top of an underlying source base that they may or
608 may not control.
610 \subsection{MQ support for patch repositories}
612 MQ helps you to work with the \sdirname{.hg/patches} directory as a
613 repository; when you prepare a repository for working with patches
614 using \hgcmd{qinit}, you can pass the \hgopt{qinit}{-c} option to
615 create the \sdirname{.hg/patches} directory as a Mercurial repository.
617 \begin{note}
618 If you forget to use the \hgopt{qinit}{-c} option, you can simply go
619 into the \sdirname{.hg/patches} directory at any time and run
620 \hgcmd{init}. Don't forget to add an entry for the
621 \sfilename{status} file to the \sfilename{.hgignore} file, though
622 (\hgcmdargs{qinit}{\hgopt{qinit}{-c}} does this for you
623 automatically); you \emph{really} don't want to manage the
624 \sfilename{status} file.
625 \end{note}
627 As a convenience, if MQ notices that the \dirname{.hg/patches}
628 directory is a repository, it will automatically \hgcmd{add} every
629 patch that you create and import.
631 Finally, MQ provides a shortcut command, \hgcmd{qcommit}, that runs
632 \hgcmd{commit} in the \sdirname{.hg/patches} directory. This saves
633 some cumbersome typing.
635 \subsection{A few things to watch out for}
637 MQ's support for working with a repository full of patches is limited
638 in a few small respects.
640 MQ cannot automatically detect changes that you make to the patch
641 directory. If you \hgcmd{pull}, manually edit, or \hgcmd{update}
642 changes to patches or the \sfilename{series} file, you will have to
643 \hgcmdargs{qpop}{\hgopt{qpop}{-a}} and then
644 \hgcmdargs{qpush}{\hgopt{qpush}{-a}} in the underlying repository to
645 see those changes show up there. If you forget to do this, you can
646 confuse MQ's idea of which patches are applied.
648 \section{Commands for working with patches}
649 \label{sec:mq:tools}
651 Once you've been working with patches for a while, you'll find
652 yourself hungry for tools that will help you to understand and
653 manipulate the patches you're dealing with.
655 The \command{diffstat} command~\cite{web:diffstat} generates a
656 histogram of the modifications made to each file in a patch. It
657 provides a good way to ``get a sense of'' a patch---which files it
658 affects, and how much change it introduces to each file and as a
659 whole. (I find that it's a good idea to use \command{diffstat}'s
660 \texttt{-p} option as a matter of course, as otherwise it will try to
661 do clever things with prefixes of file names that inevitably confuse
662 at least me.)
664 \begin{figure}[ht]
665 \interaction{mq.tools.tools}
666 \caption{The \command{diffstat}, \command{filterdiff}, and \command{lsdiff} commands}
667 \label{ex:mq:tools}
668 \end{figure}
670 The \package{patchutils} package~\cite{web:patchutils} is invaluable.
671 It provides a set of small utilities that follow the ``Unix
672 philosophy;'' each does one useful thing with a patch. The
673 \package{patchutils} command I use most is \command{filterdiff}, which
674 extracts subsets from a patch file. For example, given a patch that
675 modifies hundreds of files across dozens of directories, a single
676 invocation of \command{filterdiff} can generate a smaller patch that
677 only touches files whose names match a particular glob pattern.
679 \section{Good ways to work with patches}
681 Whether you are working on a patch series to submit to a free software
682 or open source project, or a series that you intend to treat as a
683 sequence of regular changesets when you're done, you can use some
684 simple techniques to keep your work well organised.
686 Give your patches descriptive names. A good name for a patch might be
687 \filename{rework-device-alloc.patch}, because it will immediately give
688 you a hint what the purpose of the patch is. Long names shouldn't be
689 a problem; you won't be typing the names often, but you \emph{will} be
690 running commands like \hgcmd{qapplied} and \hgcmd{qtop} over and over.
691 Good naming becomes especially important when you have a number of
692 patches to work with, or if you are juggling a number of different
693 tasks and your patches only get a fraction of your attention.
695 Be aware of what patch you're working on. Use the \hgcmd{qtop}
696 command and skim over the text of your patches frequently---for
697 example, using \hgcmdargs{tip}{\hgopt{tip}{-p}})---to be sure of where
698 you stand. I have several times worked on and \hgcmd{qrefresh}ed a
699 patch other than the one I intended, and it's often tricky to migrate
700 changes into the right patch after making them in the wrong one.
702 For this reason, it is very much worth investing a little time to
703 learn how to use some of the third-party tools I described in
704 section~\ref{sec:mq:tools}, particularly \command{diffstat} and
705 \command{filterdiff}. The former will give you a quick idea of what
706 changes your patch is making, while the latter makes it easy to splice
707 hunks selectively out of one patch and into another.
709 \section{MQ cookbook}
711 \subsection{Manage ``trivial'' patches}
713 Because the overhead of dropping files into a new Mercurial repository
714 is so low, it makes a lot of sense to manage patches this way even if
715 you simply want to make a few changes to a source tarball that you
716 downloaded.
718 Begin by downloading and unpacking the source tarball,
719 and turning it into a Mercurial repository.
720 \interaction{mq.tarball.download}
722 Continue by creating a patch stack and making your changes.
723 \interaction{mq.tarball.qinit}
725 Let's say a few weeks or months pass, and your package author releases
726 a new version. First, bring their changes into the repository.
727 \interaction{mq.tarball.newsource}
728 The pipeline starting with \hgcmd{locate} above deletes all files in
729 the working directory, so that \hgcmd{commit}'s
730 \hgopt{commit}{--addremove} option can actually tell which files have
731 really been removed in the newer version of the source.
733 Finally, you can apply your patches on top of the new tree.
734 \interaction{mq.tarball.repush}
736 \subsection{Combining entire patches}
737 \label{sec:mq:combine}
739 It's easy to combine entire patches.
741 \begin{enumerate}
742 \item \hgcmd{qpop} your applied patches until neither patch is
743 applied.
744 \item Concatenate the patches that you want to combine together:
745 \begin{codesample4}
746 cat patch-to-drop.patch >> patch-to-augment.patch
747 \end{codesample4}
748 The description from the first patch (if you have one) will be used
749 as the commit comment when you \hgcmd{qpush} the combined patch.
750 Edit the patch description if you need to.
751 \item Use the \hgcmd{qdel} command to delete the patch you're dropping
752 from the \sfilename{series} file.
753 \item \hgcmd{qpush} the combined patch. Fix up any rejects.
754 \item \hgcmd{qrefresh} the combined patch to tidy it up.
755 \end{enumerate}
757 \subsection{Merging part of one patch into another}
759 Merging \emph{part} of one patch into another is more difficult than
760 combining entire patches.
762 If you want to move changes to entire files, you can use
763 \command{filterdiff}'s \cmdopt{filterdiff}{-i} and
764 \cmdopt{filterdiff}{-x} options to choose the modifications to snip
765 out of one patch, concatenating its output onto the end of the patch
766 you want to merge into. You usually won't need to modify the patch
767 you've merged the changes from. Instead, MQ will report some rejected
768 hunks when you \hgcmd{qpush} it (from the hunks you moved into the
769 other patch), and you can simply \hgcmd{qrefresh} the patch to drop
770 the duplicate hunks.
772 If you have a patch that has multiple hunks modifying a file, and you
773 only want to move a few of those hunks, the job becomes more messy,
774 but you can still partly automate it. Use \cmdargs{lsdiff}{-nvv} to
775 print some metadata about the patch.
776 \interaction{mq.tools.lsdiff}
778 This command prints three different kinds of number:
779 \begin{itemize}
780 \item a \emph{file number} to identify each file modified in the patch;
781 \item the line number within a modified file that a hunk starts at; and
782 \item a \emph{hunk number} to identify that hunk.
783 \end{itemize}
785 You'll have to use some visual inspection, and reading of the patch,
786 to identify the file and hunk numbers you'll want, but you can then
787 pass them to to \command{filterdiff}'s \cmdopt{filterdiff}{--files}
788 and \cmdopt{filterdiff}{--hunks} options, to select exactly the file
789 and hunk you want to extract.
791 Once you have this hunk, you can concatenate it onto the end of your
792 destination patch and continue with the remainder of
793 section~\ref{sec:mq:combine}.
795 %%% Local Variables:
796 %%% mode: latex
797 %%% TeX-master: "00book"
798 %%% End: