| rev |
line source |
|
bos@559
|
1 <!-- vim: set filetype=docbkxml shiftwidth=2 autoindent expandtab tw=77 : -->
|
|
bos@559
|
2
|
|
bos@559
|
3 <chapter id="chap:mq-collab">
|
|
bos@559
|
4 <title>Advanced uses of Mercurial Queues</title>
|
|
bos@559
|
5
|
|
bos@559
|
6 <para>While it's easy to pick up straightforward uses of Mercurial
|
|
bos@559
|
7 Queues, use of a little discipline and some of MQ's less
|
|
bos@559
|
8 frequently used capabilities makes it possible to work in
|
|
bos@559
|
9 complicated development environments.</para>
|
|
bos@559
|
10
|
|
bos@559
|
11 <para>In this chapter, I will use as an example a technique I have
|
|
bos@559
|
12 used to manage the development of an Infiniband device driver for
|
|
bos@559
|
13 the Linux kernel. The driver in question is large (at least as
|
|
bos@559
|
14 drivers go), with 25,000 lines of code spread across 35 source
|
|
bos@559
|
15 files. It is maintained by a small team of developers.</para>
|
|
bos@559
|
16
|
|
bos@559
|
17 <para>While much of the material in this chapter is specific to
|
|
bos@559
|
18 Linux, the same principles apply to any code base for which you're
|
|
bos@559
|
19 not the primary owner, and upon which you need to do a lot of
|
|
bos@559
|
20 development.</para>
|
|
bos@559
|
21
|
|
bos@559
|
22 <sect1>
|
|
bos@559
|
23 <title>The problem of many targets</title>
|
|
bos@559
|
24
|
|
bos@559
|
25 <para>The Linux kernel changes rapidly, and has never been
|
|
bos@559
|
26 internally stable; developers frequently make drastic changes
|
|
bos@559
|
27 between releases. This means that a version of the driver that
|
|
bos@559
|
28 works well with a particular released version of the kernel will
|
|
bos@559
|
29 not even <emphasis>compile</emphasis> correctly against,
|
|
bos@559
|
30 typically, any other version.</para>
|
|
bos@559
|
31
|
|
bos@559
|
32 <para>To maintain a driver, we have to keep a number of distinct
|
|
bos@559
|
33 versions of Linux in mind.</para>
|
|
bos@559
|
34 <itemizedlist>
|
|
bos@559
|
35 <listitem><para>One target is the main Linux kernel development
|
|
bos@559
|
36 tree. Maintenance of the code is in this case partly shared
|
|
bos@559
|
37 by other developers in the kernel community, who make
|
|
bos@559
|
38 <quote>drive-by</quote> modifications to the driver as they
|
|
bos@559
|
39 develop and refine kernel subsystems.</para>
|
|
bos@559
|
40 </listitem>
|
|
bos@559
|
41 <listitem><para>We also maintain a number of
|
|
bos@559
|
42 <quote>backports</quote> to older versions of the Linux
|
|
bos@559
|
43 kernel, to support the needs of customers who are running
|
|
bos@559
|
44 older Linux distributions that do not incorporate our
|
|
bos@559
|
45 drivers. (To <emphasis>backport</emphasis> a piece of code
|
|
bos@559
|
46 is to modify it to work in an older version of its target
|
|
bos@559
|
47 environment than the version it was developed for.)</para>
|
|
bos@559
|
48 </listitem>
|
|
bos@559
|
49 <listitem><para>Finally, we make software releases on a schedule
|
|
bos@559
|
50 that is necessarily not aligned with those used by Linux
|
|
bos@559
|
51 distributors and kernel developers, so that we can deliver
|
|
bos@559
|
52 new features to customers without forcing them to upgrade
|
|
bos@559
|
53 their entire kernels or distributions.</para>
|
|
bos@559
|
54 </listitem></itemizedlist>
|
|
bos@559
|
55
|
|
bos@559
|
56 <sect2>
|
|
bos@559
|
57 <title>Tempting approaches that don't work well</title>
|
|
bos@559
|
58
|
|
bos@559
|
59 <para>There are two <quote>standard</quote> ways to maintain a
|
|
bos@559
|
60 piece of software that has to target many different
|
|
bos@559
|
61 environments.</para>
|
|
bos@559
|
62
|
|
bos@559
|
63 <para>The first is to maintain a number of branches, each
|
|
bos@559
|
64 intended for a single target. The trouble with this approach
|
|
bos@559
|
65 is that you must maintain iron discipline in the flow of
|
|
bos@559
|
66 changes between repositories. A new feature or bug fix must
|
|
bos@559
|
67 start life in a <quote>pristine</quote> repository, then
|
|
bos@559
|
68 percolate out to every backport repository. Backport changes
|
|
bos@559
|
69 are more limited in the branches they should propagate to; a
|
|
bos@559
|
70 backport change that is applied to a branch where it doesn't
|
|
bos@559
|
71 belong will probably stop the driver from compiling.</para>
|
|
bos@559
|
72
|
|
bos@559
|
73 <para>The second is to maintain a single source tree filled with
|
|
bos@559
|
74 conditional statements that turn chunks of code on or off
|
|
bos@559
|
75 depending on the intended target. Because these
|
|
bos@559
|
76 <quote>ifdefs</quote> are not allowed in the Linux kernel
|
|
bos@559
|
77 tree, a manual or automatic process must be followed to strip
|
|
bos@559
|
78 them out and yield a clean tree. A code base maintained in
|
|
bos@559
|
79 this fashion rapidly becomes a rat's nest of conditional
|
|
bos@559
|
80 blocks that are difficult to understand and maintain.</para>
|
|
bos@559
|
81
|
|
bos@559
|
82 <para>Neither of these approaches is well suited to a situation
|
|
bos@559
|
83 where you don't <quote>own</quote> the canonical copy of a
|
|
bos@559
|
84 source tree. In the case of a Linux driver that is
|
|
bos@559
|
85 distributed with the standard kernel, Linus's tree contains
|
|
bos@559
|
86 the copy of the code that will be treated by the world as
|
|
bos@559
|
87 canonical. The upstream version of <quote>my</quote> driver
|
|
bos@559
|
88 can be modified by people I don't know, without me even
|
|
bos@559
|
89 finding out about it until after the changes show up in
|
|
bos@559
|
90 Linus's tree.</para>
|
|
bos@559
|
91
|
|
bos@559
|
92 <para>These approaches have the added weakness of making it
|
|
bos@559
|
93 difficult to generate well-formed patches to submit
|
|
bos@559
|
94 upstream.</para>
|
|
bos@559
|
95
|
|
bos@559
|
96 <para>In principle, Mercurial Queues seems like a good candidate
|
|
bos@559
|
97 to manage a development scenario such as the above. While
|
|
bos@559
|
98 this is indeed the case, MQ contains a few added features that
|
|
bos@559
|
99 make the job more pleasant.</para>
|
|
bos@559
|
100
|
|
bos@559
|
101 </sect2>
|
|
bos@559
|
102 </sect1>
|
|
bos@559
|
103 <sect1>
|
|
bos@559
|
104 <title>Conditionally applying patches with guards</title>
|
|
bos@559
|
105
|
|
bos@559
|
106 <para>Perhaps the best way to maintain sanity with so many targets
|
|
bos@559
|
107 is to be able to choose specific patches to apply for a given
|
|
bos@559
|
108 situation. MQ provides a feature called <quote>guards</quote>
|
|
bos@559
|
109 (which originates with quilt's <literal>guards</literal>
|
|
bos@559
|
110 command) that does just this. To start off, let's create a
|
|
bos@567
|
111 simple repository for experimenting in.</para>
|
|
bos@567
|
112
|
|
bos@567
|
113 &interaction.mq.guards.init;
|
|
bos@567
|
114
|
|
bos@567
|
115 <para>This gives us a tiny repository that contains two patches
|
|
bos@567
|
116 that don't have any dependencies on each other, because they
|
|
bos@567
|
117 touch different files.</para>
|
|
bos@559
|
118
|
|
bos@559
|
119 <para>The idea behind conditional application is that you can
|
|
bos@559
|
120 <quote>tag</quote> a patch with a <emphasis>guard</emphasis>,
|
|
bos@559
|
121 which is simply a text string of your choosing, then tell MQ to
|
|
bos@559
|
122 select specific guards to use when applying patches. MQ will
|
|
bos@559
|
123 then either apply, or skip over, a guarded patch, depending on
|
|
bos@559
|
124 the guards that you have selected.</para>
|
|
bos@559
|
125
|
|
bos@559
|
126 <para>A patch can have an arbitrary number of guards; each one is
|
|
bos@559
|
127 <emphasis>positive</emphasis> (<quote>apply this patch if this
|
|
bos@559
|
128 guard is selected</quote>) or <emphasis>negative</emphasis>
|
|
bos@559
|
129 (<quote>skip this patch if this guard is selected</quote>). A
|
|
bos@559
|
130 patch with no guards is always applied.</para>
|
|
bos@559
|
131
|
|
bos@559
|
132 </sect1>
|
|
bos@559
|
133 <sect1>
|
|
bos@559
|
134 <title>Controlling the guards on a patch</title>
|
|
bos@559
|
135
|
|
bos@559
|
136 <para>The <command role="hg-ext-mq">qguard</command> command lets
|
|
bos@559
|
137 you determine which guards should apply to a patch, or display
|
|
bos@559
|
138 the guards that are already in effect. Without any arguments, it
|
|
bos@567
|
139 displays the guards on the current topmost patch.</para>
|
|
bos@567
|
140
|
|
bos@567
|
141 &interaction.mq.guards.qguard;
|
|
bos@567
|
142
|
|
bos@567
|
143 <para>To set a positive guard on a patch, prefix the name of the
|
|
bos@567
|
144 guard with a <quote><literal>+</literal></quote>.</para>
|
|
bos@567
|
145
|
|
bos@567
|
146 &interaction.mq.guards.qguard.pos;
|
|
bos@567
|
147
|
|
bos@567
|
148 <para>To set a negative guard
|
|
bos@559
|
149 on a patch, prefix the name of the guard with a
|
|
bos@567
|
150 <quote><literal>-</literal></quote>.</para>
|
|
bos@567
|
151
|
|
bos@567
|
152 &interaction.mq.guards.qguard.neg;
|
|
bos@559
|
153
|
|
bos@559
|
154 <note>
|
|
bos@559
|
155 <para> The <command role="hg-ext-mq">qguard</command> command
|
|
bos@559
|
156 <emphasis>sets</emphasis> the guards on a patch; it doesn't
|
|
bos@559
|
157 <emphasis>modify</emphasis> them. What this means is that if
|
|
bos@559
|
158 you run <command role="hg-cmd">hg qguard +a +b</command> on a
|
|
bos@559
|
159 patch, then <command role="hg-cmd">hg qguard +c</command> on
|
|
bos@559
|
160 the same patch, the <emphasis>only</emphasis> guard that will
|
|
bos@559
|
161 be set on it afterwards is <literal>+c</literal>.</para>
|
|
bos@559
|
162 </note>
|
|
bos@559
|
163
|
|
bos@559
|
164 <para>Mercurial stores guards in the <filename
|
|
bos@559
|
165 role="special">series</filename> file; the form in which they
|
|
bos@559
|
166 are stored is easy both to understand and to edit by hand. (In
|
|
bos@559
|
167 other words, you don't have to use the <command
|
|
bos@559
|
168 role="hg-ext-mq">qguard</command> command if you don't want
|
|
bos@559
|
169 to; it's okay to simply edit the <filename
|
|
bos@567
|
170 role="special">series</filename> file.)</para>
|
|
bos@567
|
171
|
|
bos@567
|
172 &interaction.mq.guards.series;
|
|
bos@559
|
173
|
|
bos@559
|
174 </sect1>
|
|
bos@559
|
175 <sect1>
|
|
bos@559
|
176 <title>Selecting the guards to use</title>
|
|
bos@559
|
177
|
|
bos@559
|
178 <para>The <command role="hg-ext-mq">qselect</command> command
|
|
bos@559
|
179 determines which guards are active at a given time. The effect
|
|
bos@559
|
180 of this is to determine which patches MQ will apply the next
|
|
bos@559
|
181 time you run <command role="hg-ext-mq">qpush</command>. It has
|
|
bos@559
|
182 no other effect; in particular, it doesn't do anything to
|
|
bos@559
|
183 patches that are already applied.</para>
|
|
bos@559
|
184
|
|
bos@559
|
185 <para>With no arguments, the <command
|
|
bos@559
|
186 role="hg-ext-mq">qselect</command> command lists the guards
|
|
bos@559
|
187 currently in effect, one per line of output. Each argument is
|
|
bos@567
|
188 treated as the name of a guard to apply.</para>
|
|
bos@567
|
189
|
|
bos@567
|
190 &interaction.mq.guards.qselect.foo;
|
|
bos@567
|
191
|
|
bos@567
|
192 <para>In case you're interested, the currently selected guards are
|
|
bos@567
|
193 stored in the <filename role="special">guards</filename> file.</para>
|
|
bos@567
|
194
|
|
bos@567
|
195 &interaction.mq.guards.qselect.cat;
|
|
bos@567
|
196
|
|
bos@567
|
197 <para>We can see the effect the selected guards have when we run
|
|
bos@567
|
198 <command role="hg-ext-mq">qpush</command>.</para>
|
|
bos@567
|
199
|
|
bos@567
|
200 &interaction.mq.guards.qselect.qpush;
|
|
bos@559
|
201
|
|
bos@559
|
202 <para>A guard cannot start with a
|
|
bos@559
|
203 <quote><literal>+</literal></quote> or
|
|
bos@559
|
204 <quote><literal>-</literal></quote> character. The name of a
|
|
bos@559
|
205 guard must not contain white space, but most other characters
|
|
bos@559
|
206 are acceptable. If you try to use a guard with an invalid name,
|
|
bos@567
|
207 MQ will complain:</para>
|
|
bos@567
|
208
|
|
bos@567
|
209 &interaction.mq.guards.qselect.error;
|
|
bos@567
|
210
|
|
bos@567
|
211 <para>Changing the selected guards changes the patches that are
|
|
bos@567
|
212 applied.</para>
|
|
bos@567
|
213
|
|
bos@567
|
214 &interaction.mq.guards.qselect.quux;
|
|
bos@567
|
215
|
|
bos@567
|
216 <para>You can see in the example below that negative guards take
|
|
bos@567
|
217 precedence over positive guards.</para>
|
|
bos@567
|
218
|
|
bos@567
|
219 &interaction.mq.guards.qselect.foobar;
|
|
bos@559
|
220
|
|
bos@559
|
221 </sect1>
|
|
bos@559
|
222 <sect1>
|
|
bos@559
|
223 <title>MQ's rules for applying patches</title>
|
|
bos@559
|
224
|
|
bos@559
|
225 <para>The rules that MQ uses when deciding whether to apply a
|
|
bos@559
|
226 patch are as follows.</para>
|
|
bos@559
|
227 <itemizedlist>
|
|
bos@559
|
228 <listitem><para>A patch that has no guards is always
|
|
bos@559
|
229 applied.</para>
|
|
bos@559
|
230 </listitem>
|
|
bos@559
|
231 <listitem><para>If the patch has any negative guard that matches
|
|
bos@559
|
232 any currently selected guard, the patch is skipped.</para>
|
|
bos@559
|
233 </listitem>
|
|
bos@559
|
234 <listitem><para>If the patch has any positive guard that matches
|
|
bos@559
|
235 any currently selected guard, the patch is applied.</para>
|
|
bos@559
|
236 </listitem>
|
|
bos@559
|
237 <listitem><para>If the patch has positive or negative guards,
|
|
bos@559
|
238 but none matches any currently selected guard, the patch is
|
|
bos@559
|
239 skipped.</para>
|
|
bos@559
|
240 </listitem></itemizedlist>
|
|
bos@559
|
241
|
|
bos@559
|
242 </sect1>
|
|
bos@559
|
243 <sect1>
|
|
bos@559
|
244 <title>Trimming the work environment</title>
|
|
bos@559
|
245
|
|
bos@559
|
246 <para>In working on the device driver I mentioned earlier, I don't
|
|
bos@559
|
247 apply the patches to a normal Linux kernel tree. Instead, I use
|
|
bos@559
|
248 a repository that contains only a snapshot of the source files
|
|
bos@559
|
249 and headers that are relevant to Infiniband development. This
|
|
bos@559
|
250 repository is 1% the size of a kernel repository, so it's easier
|
|
bos@559
|
251 to work with.</para>
|
|
bos@559
|
252
|
|
bos@559
|
253 <para>I then choose a <quote>base</quote> version on top of which
|
|
bos@559
|
254 the patches are applied. This is a snapshot of the Linux kernel
|
|
bos@559
|
255 tree as of a revision of my choosing. When I take the snapshot,
|
|
bos@559
|
256 I record the changeset ID from the kernel repository in the
|
|
bos@559
|
257 commit message. Since the snapshot preserves the
|
|
bos@559
|
258 <quote>shape</quote> and content of the relevant parts of the
|
|
bos@559
|
259 kernel tree, I can apply my patches on top of either my tiny
|
|
bos@559
|
260 repository or a normal kernel tree.</para>
|
|
bos@559
|
261
|
|
bos@559
|
262 <para>Normally, the base tree atop which the patches apply should
|
|
bos@559
|
263 be a snapshot of a very recent upstream tree. This best
|
|
bos@559
|
264 facilitates the development of patches that can easily be
|
|
bos@559
|
265 submitted upstream with few or no modifications.</para>
|
|
bos@559
|
266
|
|
bos@559
|
267 </sect1>
|
|
bos@559
|
268 <sect1>
|
|
bos@559
|
269 <title>Dividing up the <filename role="special">series</filename>
|
|
bos@559
|
270 file</title>
|
|
bos@559
|
271
|
|
bos@559
|
272 <para>I categorise the patches in the <filename
|
|
bos@559
|
273 role="special">series</filename> file into a number of logical
|
|
bos@559
|
274 groups. Each section of like patches begins with a block of
|
|
bos@559
|
275 comments that describes the purpose of the patches that
|
|
bos@559
|
276 follow.</para>
|
|
bos@559
|
277
|
|
bos@559
|
278 <para>The sequence of patch groups that I maintain follows. The
|
|
bos@559
|
279 ordering of these groups is important; I'll describe why after I
|
|
bos@559
|
280 introduce the groups.</para>
|
|
bos@559
|
281 <itemizedlist>
|
|
bos@559
|
282 <listitem><para>The <quote>accepted</quote> group. Patches that
|
|
bos@559
|
283 the development team has submitted to the maintainer of the
|
|
bos@559
|
284 Infiniband subsystem, and which he has accepted, but which
|
|
bos@559
|
285 are not present in the snapshot that the tiny repository is
|
|
bos@559
|
286 based on. These are <quote>read only</quote> patches,
|
|
bos@559
|
287 present only to transform the tree into a similar state as
|
|
bos@559
|
288 it is in the upstream maintainer's repository.</para>
|
|
bos@559
|
289 </listitem>
|
|
bos@559
|
290 <listitem><para>The <quote>rework</quote> group. Patches that I
|
|
bos@559
|
291 have submitted, but that the upstream maintainer has
|
|
bos@559
|
292 requested modifications to before he will accept
|
|
bos@559
|
293 them.</para>
|
|
bos@559
|
294 </listitem>
|
|
bos@559
|
295 <listitem><para>The <quote>pending</quote> group. Patches that
|
|
bos@559
|
296 I have not yet submitted to the upstream maintainer, but
|
|
bos@559
|
297 which we have finished working on. These will be <quote>read
|
|
bos@559
|
298 only</quote> for a while. If the upstream maintainer
|
|
bos@559
|
299 accepts them upon submission, I'll move them to the end of
|
|
bos@559
|
300 the <quote>accepted</quote> group. If he requests that I
|
|
bos@559
|
301 modify any, I'll move them to the beginning of the
|
|
bos@559
|
302 <quote>rework</quote> group.</para>
|
|
bos@559
|
303 </listitem>
|
|
bos@559
|
304 <listitem><para>The <quote>in progress</quote> group. Patches
|
|
bos@559
|
305 that are actively being developed, and should not be
|
|
bos@559
|
306 submitted anywhere yet.</para>
|
|
bos@559
|
307 </listitem>
|
|
bos@559
|
308 <listitem><para>The <quote>backport</quote> group. Patches that
|
|
bos@559
|
309 adapt the source tree to older versions of the kernel
|
|
bos@559
|
310 tree.</para>
|
|
bos@559
|
311 </listitem>
|
|
bos@559
|
312 <listitem><para>The <quote>do not ship</quote> group. Patches
|
|
bos@559
|
313 that for some reason should never be submitted upstream.
|
|
bos@559
|
314 For example, one such patch might change embedded driver
|
|
bos@559
|
315 identification strings to make it easier to distinguish, in
|
|
bos@559
|
316 the field, between an out-of-tree version of the driver and
|
|
bos@559
|
317 a version shipped by a distribution vendor.</para>
|
|
bos@559
|
318 </listitem></itemizedlist>
|
|
bos@559
|
319
|
|
bos@559
|
320 <para>Now to return to the reasons for ordering groups of patches
|
|
bos@559
|
321 in this way. We would like the lowest patches in the stack to
|
|
bos@559
|
322 be as stable as possible, so that we will not need to rework
|
|
bos@559
|
323 higher patches due to changes in context. Putting patches that
|
|
bos@559
|
324 will never be changed first in the <filename
|
|
bos@559
|
325 role="special">series</filename> file serves this
|
|
bos@559
|
326 purpose.</para>
|
|
bos@559
|
327
|
|
bos@559
|
328 <para>We would also like the patches that we know we'll need to
|
|
bos@559
|
329 modify to be applied on top of a source tree that resembles the
|
|
bos@559
|
330 upstream tree as closely as possible. This is why we keep
|
|
bos@559
|
331 accepted patches around for a while.</para>
|
|
bos@559
|
332
|
|
bos@559
|
333 <para>The <quote>backport</quote> and <quote>do not ship</quote>
|
|
bos@559
|
334 patches float at the end of the <filename
|
|
bos@559
|
335 role="special">series</filename> file. The backport patches
|
|
bos@559
|
336 must be applied on top of all other patches, and the <quote>do
|
|
bos@559
|
337 not ship</quote> patches might as well stay out of harm's
|
|
bos@559
|
338 way.</para>
|
|
bos@559
|
339
|
|
bos@559
|
340 </sect1>
|
|
bos@559
|
341 <sect1>
|
|
bos@559
|
342 <title>Maintaining the patch series</title>
|
|
bos@559
|
343
|
|
bos@559
|
344 <para>In my work, I use a number of guards to control which
|
|
bos@559
|
345 patches are to be applied.</para>
|
|
bos@559
|
346
|
|
bos@559
|
347 <itemizedlist>
|
|
bos@559
|
348 <listitem><para><quote>Accepted</quote> patches are guarded with
|
|
bos@559
|
349 <literal>accepted</literal>. I enable this guard most of
|
|
bos@559
|
350 the time. When I'm applying the patches on top of a tree
|
|
bos@559
|
351 where the patches are already present, I can turn this patch
|
|
bos@559
|
352 off, and the patches that follow it will apply
|
|
bos@559
|
353 cleanly.</para>
|
|
bos@559
|
354 </listitem>
|
|
bos@559
|
355 <listitem><para>Patches that are <quote>finished</quote>, but
|
|
bos@559
|
356 not yet submitted, have no guards. If I'm applying the
|
|
bos@559
|
357 patch stack to a copy of the upstream tree, I don't need to
|
|
bos@559
|
358 enable any guards in order to get a reasonably safe source
|
|
bos@559
|
359 tree.</para>
|
|
bos@559
|
360 </listitem>
|
|
bos@559
|
361 <listitem><para>Those patches that need reworking before being
|
|
bos@559
|
362 resubmitted are guarded with
|
|
bos@559
|
363 <literal>rework</literal>.</para>
|
|
bos@559
|
364 </listitem>
|
|
bos@559
|
365 <listitem><para>For those patches that are still under
|
|
bos@559
|
366 development, I use <literal>devel</literal>.</para>
|
|
bos@559
|
367 </listitem>
|
|
bos@559
|
368 <listitem><para>A backport patch may have several guards, one
|
|
bos@559
|
369 for each version of the kernel to which it applies. For
|
|
bos@559
|
370 example, a patch that backports a piece of code to 2.6.9
|
|
bos@559
|
371 will have a <literal>2.6.9</literal> guard.</para>
|
|
bos@559
|
372 </listitem></itemizedlist>
|
|
bos@559
|
373 <para>This variety of guards gives me considerable flexibility in
|
|
bos@559
|
374 determining what kind of source tree I want to end up with. For
|
|
bos@559
|
375 most situations, the selection of appropriate guards is
|
|
bos@559
|
376 automated during the build process, but I can manually tune the
|
|
bos@559
|
377 guards to use for less common circumstances.</para>
|
|
bos@559
|
378
|
|
bos@559
|
379 <sect2>
|
|
bos@559
|
380 <title>The art of writing backport patches</title>
|
|
bos@559
|
381
|
|
bos@559
|
382 <para>Using MQ, writing a backport patch is a simple process.
|
|
bos@559
|
383 All such a patch has to do is modify a piece of code that uses
|
|
bos@559
|
384 a kernel feature not present in the older version of the
|
|
bos@559
|
385 kernel, so that the driver continues to work correctly under
|
|
bos@559
|
386 that older version.</para>
|
|
bos@559
|
387
|
|
bos@559
|
388 <para>A useful goal when writing a good backport patch is to
|
|
bos@559
|
389 make your code look as if it was written for the older version
|
|
bos@559
|
390 of the kernel you're targeting. The less obtrusive the patch,
|
|
bos@559
|
391 the easier it will be to understand and maintain. If you're
|
|
bos@559
|
392 writing a collection of backport patches to avoid the
|
|
bos@559
|
393 <quote>rat's nest</quote> effect of lots of
|
|
bos@559
|
394 <literal>#ifdef</literal>s (hunks of source code that are only
|
|
bos@559
|
395 used conditionally) in your code, don't introduce
|
|
bos@559
|
396 version-dependent <literal>#ifdef</literal>s into the patches.
|
|
bos@559
|
397 Instead, write several patches, each of which makes
|
|
bos@559
|
398 unconditional changes, and control their application using
|
|
bos@559
|
399 guards.</para>
|
|
bos@559
|
400
|
|
bos@559
|
401 <para>There are two reasons to divide backport patches into a
|
|
bos@559
|
402 distinct group, away from the <quote>regular</quote> patches
|
|
bos@559
|
403 whose effects they modify. The first is that intermingling the
|
|
bos@559
|
404 two makes it more difficult to use a tool like the <literal
|
|
bos@559
|
405 role="hg-ext">patchbomb</literal> extension to automate the
|
|
bos@559
|
406 process of submitting the patches to an upstream maintainer.
|
|
bos@559
|
407 The second is that a backport patch could perturb the context
|
|
bos@559
|
408 in which a subsequent regular patch is applied, making it
|
|
bos@559
|
409 impossible to apply the regular patch cleanly
|
|
bos@559
|
410 <emphasis>without</emphasis> the earlier backport patch
|
|
bos@559
|
411 already being applied.</para>
|
|
bos@559
|
412
|
|
bos@559
|
413 </sect2>
|
|
bos@559
|
414 </sect1>
|
|
bos@559
|
415 <sect1>
|
|
bos@559
|
416 <title>Useful tips for developing with MQ</title>
|
|
bos@559
|
417
|
|
bos@559
|
418 <sect2>
|
|
bos@559
|
419 <title>Organising patches in directories</title>
|
|
bos@559
|
420
|
|
bos@559
|
421 <para>If you're working on a substantial project with MQ, it's
|
|
bos@559
|
422 not difficult to accumulate a large number of patches. For
|
|
bos@559
|
423 example, I have one patch repository that contains over 250
|
|
bos@559
|
424 patches.</para>
|
|
bos@559
|
425
|
|
bos@559
|
426 <para>If you can group these patches into separate logical
|
|
bos@559
|
427 categories, you can if you like store them in different
|
|
bos@559
|
428 directories; MQ has no problems with patch names that contain
|
|
bos@559
|
429 path separators.</para>
|
|
bos@559
|
430
|
|
bos@559
|
431 </sect2>
|
|
bos@559
|
432 <sect2 id="mq-collab:tips:interdiff">
|
|
bos@559
|
433 <title>Viewing the history of a patch</title>
|
|
bos@559
|
434
|
|
bos@559
|
435 <para>If you're developing a set of patches over a long time,
|
|
bos@559
|
436 it's a good idea to maintain them in a repository, as
|
|
bos@559
|
437 discussed in section <xref linkend="sec:mq:repo"/>. If you do
|
|
bos@559
|
438 so, you'll quickly
|
|
bos@559
|
439 discover that using the <command role="hg-cmd">hg
|
|
bos@559
|
440 diff</command> command to look at the history of changes to
|
|
bos@559
|
441 a patch is unworkable. This is in part because you're looking
|
|
bos@559
|
442 at the second derivative of the real code (a diff of a diff),
|
|
bos@559
|
443 but also because MQ adds noise to the process by modifying
|
|
bos@559
|
444 time stamps and directory names when it updates a
|
|
bos@559
|
445 patch.</para>
|
|
bos@559
|
446
|
|
bos@559
|
447 <para>However, you can use the <literal
|
|
bos@559
|
448 role="hg-ext">extdiff</literal> extension, which is bundled
|
|
bos@559
|
449 with Mercurial, to turn a diff of two versions of a patch into
|
|
bos@559
|
450 something readable. To do this, you will need a third-party
|
|
bos@559
|
451 package called <literal role="package">patchutils</literal>
|
|
bos@559
|
452 <citation>web:patchutils</citation>. This provides a command
|
|
bos@559
|
453 named <command>interdiff</command>, which shows the
|
|
bos@559
|
454 differences between two diffs as a diff. Used on two versions
|
|
bos@559
|
455 of the same diff, it generates a diff that represents the diff
|
|
bos@559
|
456 from the first to the second version.</para>
|
|
bos@559
|
457
|
|
bos@559
|
458 <para>You can enable the <literal
|
|
bos@559
|
459 role="hg-ext">extdiff</literal> extension in the usual way,
|
|
bos@559
|
460 by adding a line to the <literal
|
|
bos@559
|
461 role="rc-extensions">extensions</literal> section of your
|
|
bos@559
|
462 <filename role="special"> /.hgrc</filename>.</para>
|
|
bos@559
|
463 <programlisting>[extensions] extdiff =</programlisting>
|
|
bos@559
|
464 <para>The <command>interdiff</command> command expects to be
|
|
bos@559
|
465 passed the names of two files, but the <literal
|
|
bos@559
|
466 role="hg-ext">extdiff</literal> extension passes the program
|
|
bos@559
|
467 it runs a pair of directories, each of which can contain an
|
|
bos@559
|
468 arbitrary number of files. We thus need a small program that
|
|
bos@559
|
469 will run <command>interdiff</command> on each pair of files in
|
|
bos@559
|
470 these two directories. This program is available as <filename
|
|
bos@559
|
471 role="special">hg-interdiff</filename> in the <filename
|
|
bos@559
|
472 class="directory">examples</filename> directory of the
|
|
bos@559
|
473 source code repository that accompanies this book. <!--
|
|
bos@559
|
474 &example.hg-interdiff; --></para>
|
|
bos@559
|
475
|
|
bos@559
|
476 <para>With the <filename role="special">hg-interdiff</filename>
|
|
bos@559
|
477 program in your shell's search path, you can run it as
|
|
bos@559
|
478 follows, from inside an MQ patch directory:</para>
|
|
bos@559
|
479 <programlisting>hg extdiff -p hg-interdiff -r A:B
|
|
bos@559
|
480 my-change.patch</programlisting>
|
|
bos@559
|
481 <para>Since you'll probably want to use this long-winded command
|
|
bos@559
|
482 a lot, you can get <literal role="hg-ext">hgext</literal> to
|
|
bos@559
|
483 make it available as a normal Mercurial command, again by
|
|
bos@559
|
484 editing your <filename role="special">
|
|
bos@559
|
485 /.hgrc</filename>.</para>
|
|
bos@559
|
486 <programlisting>[extdiff] cmd.interdiff =
|
|
bos@559
|
487 hg-interdiff</programlisting>
|
|
bos@559
|
488 <para>This directs <literal role="hg-ext">hgext</literal> to
|
|
bos@559
|
489 make an <literal>interdiff</literal> command available, so you
|
|
bos@559
|
490 can now shorten the previous invocation of <command
|
|
bos@559
|
491 role="hg-ext-extdiff">extdiff</command> to something a
|
|
bos@559
|
492 little more wieldy.</para>
|
|
bos@559
|
493 <programlisting>hg interdiff -r A:B
|
|
bos@559
|
494 my-change.patch</programlisting>
|
|
bos@559
|
495
|
|
bos@559
|
496 <note>
|
|
bos@559
|
497 <para> The <command>interdiff</command> command works well
|
|
bos@559
|
498 only if the underlying files against which versions of a
|
|
bos@559
|
499 patch are generated remain the same. If you create a patch,
|
|
bos@559
|
500 modify the underlying files, and then regenerate the patch,
|
|
bos@559
|
501 <command>interdiff</command> may not produce useful
|
|
bos@559
|
502 output.</para>
|
|
bos@559
|
503 </note>
|
|
bos@559
|
504
|
|
bos@559
|
505 <para>The <literal role="hg-ext">extdiff</literal> extension is
|
|
bos@559
|
506 useful for more than merely improving the presentation of MQ
|
|
bos@559
|
507 patches. To read more about it, go to section <xref
|
|
bos@559
|
508 linkend="sec:hgext:extdiff"/>.</para>
|
|
bos@559
|
509
|
|
bos@559
|
510 </sect2>
|
|
bos@559
|
511 </sect1>
|
|
bos@559
|
512 </chapter>
|
|
bos@559
|
513
|
|
bos@559
|
514 <!--
|
|
bos@559
|
515 local variables:
|
|
bos@559
|
516 sgml-parent-document: ("00book.xml" "book" "chapter")
|
|
bos@559
|
517 end:
|
|
bos@559
|
518 -->
|
