Mercurial > hgbook
comparison en/ch01-intro.xml @ 831:acf9dc5f088d
Add a skeletal preface.
author | Bryan O'Sullivan <bos@serpentine.com> |
---|---|
date | Thu, 07 May 2009 21:07:35 -0700 |
parents | en/ch00-preface.xml@b338f5490029 |
children |
comparison
equal
deleted
inserted
replaced
830:cbdff5945f9d | 831:acf9dc5f088d |
---|---|
1 <!-- vim: set filetype=docbkxml shiftwidth=2 autoindent expandtab tw=77 : --> | |
2 | |
3 <chapter id="chap:intro"> | |
4 <?dbhtml filename="how-did-we-get-here.html"?> | |
5 <title>How did we get here?</title> | |
6 | |
7 <sect1> | |
8 <title>Why revision control? Why Mercurial?</title> | |
9 | |
10 <para id="x_6d">Revision control is the process of managing multiple | |
11 versions of a piece of information. In its simplest form, this | |
12 is something that many people do by hand: every time you modify | |
13 a file, save it under a new name that contains a number, each | |
14 one higher than the number of the preceding version.</para> | |
15 | |
16 <para id="x_6e">Manually managing multiple versions of even a single file is | |
17 an error-prone task, though, so software tools to help automate | |
18 this process have long been available. The earliest automated | |
19 revision control tools were intended to help a single user to | |
20 manage revisions of a single file. Over the past few decades, | |
21 the scope of revision control tools has expanded greatly; they | |
22 now manage multiple files, and help multiple people to work | |
23 together. The best modern revision control tools have no | |
24 problem coping with thousands of people working together on | |
25 projects that consist of hundreds of thousands of files.</para> | |
26 | |
27 <para id="x_6f">The arrival of distributed revision control is relatively | |
28 recent, and so far this new field has grown due to people's | |
29 willingness to explore ill-charted territory.</para> | |
30 | |
31 <para id="x_70">I am writing a book about distributed revision control | |
32 because I believe that it is an important subject that deserves | |
33 a field guide. I chose to write about Mercurial because it is | |
34 the easiest tool to learn the terrain with, and yet it scales to | |
35 the demands of real, challenging environments where many other | |
36 revision control tools buckle.</para> | |
37 | |
38 <sect2> | |
39 <title>Why use revision control?</title> | |
40 | |
41 <para id="x_71">There are a number of reasons why you or your team might | |
42 want to use an automated revision control tool for a | |
43 project.</para> | |
44 | |
45 <itemizedlist> | |
46 <listitem><para id="x_72">It will track the history and evolution of | |
47 your project, so you don't have to. For every change, | |
48 you'll have a log of <emphasis>who</emphasis> made it; | |
49 <emphasis>why</emphasis> they made it; | |
50 <emphasis>when</emphasis> they made it; and | |
51 <emphasis>what</emphasis> the change | |
52 was.</para></listitem> | |
53 <listitem><para id="x_73">When you're working with other people, | |
54 revision control software makes it easier for you to | |
55 collaborate. For example, when people more or less | |
56 simultaneously make potentially incompatible changes, the | |
57 software will help you to identify and resolve those | |
58 conflicts.</para></listitem> | |
59 <listitem><para id="x_74">It can help you to recover from mistakes. If | |
60 you make a change that later turns out to be in error, you | |
61 can revert to an earlier version of one or more files. In | |
62 fact, a <emphasis>really</emphasis> good revision control | |
63 tool will even help you to efficiently figure out exactly | |
64 when a problem was introduced (see <xref | |
65 linkend="sec:undo:bisect"/> for details).</para></listitem> | |
66 <listitem><para id="x_75">It will help you to work simultaneously on, | |
67 and manage the drift between, multiple versions of your | |
68 project.</para></listitem> | |
69 </itemizedlist> | |
70 | |
71 <para id="x_76">Most of these reasons are equally | |
72 valid&emdash;at least in theory&emdash;whether you're working | |
73 on a project by yourself, or with a hundred other | |
74 people.</para> | |
75 | |
76 <para id="x_77">A key question about the practicality of revision control | |
77 at these two different scales (<quote>lone hacker</quote> and | |
78 <quote>huge team</quote>) is how its | |
79 <emphasis>benefits</emphasis> compare to its | |
80 <emphasis>costs</emphasis>. A revision control tool that's | |
81 difficult to understand or use is going to impose a high | |
82 cost.</para> | |
83 | |
84 <para id="x_78">A five-hundred-person project is likely to collapse under | |
85 its own weight almost immediately without a revision control | |
86 tool and process. In this case, the cost of using revision | |
87 control might hardly seem worth considering, since | |
88 <emphasis>without</emphasis> it, failure is almost | |
89 guaranteed.</para> | |
90 | |
91 <para id="x_79">On the other hand, a one-person <quote>quick hack</quote> | |
92 might seem like a poor place to use a revision control tool, | |
93 because surely the cost of using one must be close to the | |
94 overall cost of the project. Right?</para> | |
95 | |
96 <para id="x_7a">Mercurial uniquely supports <emphasis>both</emphasis> of | |
97 these scales of development. You can learn the basics in just | |
98 a few minutes, and due to its low overhead, you can apply | |
99 revision control to the smallest of projects with ease. Its | |
100 simplicity means you won't have a lot of abstruse concepts or | |
101 command sequences competing for mental space with whatever | |
102 you're <emphasis>really</emphasis> trying to do. At the same | |
103 time, Mercurial's high performance and peer-to-peer nature let | |
104 you scale painlessly to handle large projects.</para> | |
105 | |
106 <para id="x_7b">No revision control tool can rescue a poorly run project, | |
107 but a good choice of tools can make a huge difference to the | |
108 fluidity with which you can work on a project.</para> | |
109 | |
110 </sect2> | |
111 | |
112 <sect2> | |
113 <title>The many names of revision control</title> | |
114 | |
115 <para id="x_7c">Revision control is a diverse field, so much so that it is | |
116 referred to by many names and acronyms. Here are a few of the | |
117 more common variations you'll encounter:</para> | |
118 <itemizedlist> | |
119 <listitem><para id="x_7d">Revision control (RCS)</para></listitem> | |
120 <listitem><para id="x_7e">Software configuration management (SCM), or | |
121 configuration management</para></listitem> | |
122 <listitem><para id="x_7f">Source code management</para></listitem> | |
123 <listitem><para id="x_80">Source code control, or source | |
124 control</para></listitem> | |
125 <listitem><para id="x_81">Version control | |
126 (VCS)</para></listitem></itemizedlist> | |
127 <para id="x_82">Some people claim that these terms actually have different | |
128 meanings, but in practice they overlap so much that there's no | |
129 agreed or even useful way to tease them apart.</para> | |
130 | |
131 </sect2> | |
132 </sect1> | |
133 | |
134 <sect1> | |
135 <title>About the examples in this book</title> | |
136 | |
137 <para id="x_84">This book takes an unusual approach to code samples. Every | |
138 example is <quote>live</quote>&emdash;each one is actually the result | |
139 of a shell script that executes the Mercurial commands you see. | |
140 Every time an image of the book is built from its sources, all | |
141 the example scripts are automatically run, and their current | |
142 results compared against their expected results.</para> | |
143 | |
144 <para id="x_85">The advantage of this approach is that the examples are | |
145 always accurate; they describe <emphasis>exactly</emphasis> the | |
146 behavior of the version of Mercurial that's mentioned at the | |
147 front of the book. If I update the version of Mercurial that | |
148 I'm documenting, and the output of some command changes, the | |
149 build fails.</para> | |
150 | |
151 <para id="x_86">There is a small disadvantage to this approach, which is | |
152 that the dates and times you'll see in examples tend to be | |
153 <quote>squashed</quote> together in a way that they wouldn't be | |
154 if the same commands were being typed by a human. Where a human | |
155 can issue no more than one command every few seconds, with any | |
156 resulting timestamps correspondingly spread out, my automated | |
157 example scripts run many commands in one second.</para> | |
158 | |
159 <para id="x_87">As an instance of this, several consecutive commits in an | |
160 example can show up as having occurred during the same second. | |
161 You can see this occur in the <literal | |
162 role="hg-ext">bisect</literal> example in <xref | |
163 linkend="sec:undo:bisect"/>, for instance.</para> | |
164 | |
165 <para id="x_88">So when you're reading examples, don't place too much weight | |
166 on the dates or times you see in the output of commands. But | |
167 <emphasis>do</emphasis> be confident that the behavior you're | |
168 seeing is consistent and reproducible.</para> | |
169 | |
170 </sect1> | |
171 | |
172 <sect1> | |
173 <title>Trends in the field</title> | |
174 | |
175 <para id="x_89">There has been an unmistakable trend in the development and | |
176 use of revision control tools over the past four decades, as | |
177 people have become familiar with the capabilities of their tools | |
178 and constrained by their limitations.</para> | |
179 | |
180 <para id="x_8a">The first generation began by managing single files on | |
181 individual computers. Although these tools represented a huge | |
182 advance over ad-hoc manual revision control, their locking model | |
183 and reliance on a single computer limited them to small, | |
184 tightly-knit teams.</para> | |
185 | |
186 <para id="x_8b">The second generation loosened these constraints by moving | |
187 to network-centered architectures, and managing entire projects | |
188 at a time. As projects grew larger, they ran into new problems. | |
189 With clients needing to talk to servers very frequently, server | |
190 scaling became an issue for large projects. An unreliable | |
191 network connection could prevent remote users from being able to | |
192 talk to the server at all. As open source projects started | |
193 making read-only access available anonymously to anyone, people | |
194 without commit privileges found that they could not use the | |
195 tools to interact with a project in a natural way, as they could | |
196 not record their changes.</para> | |
197 | |
198 <para id="x_8c">The current generation of revision control tools is | |
199 peer-to-peer in nature. All of these systems have dropped the | |
200 dependency on a single central server, and allow people to | |
201 distribute their revision control data to where it's actually | |
202 needed. Collaboration over the Internet has moved from | |
203 constrained by technology to a matter of choice and consensus. | |
204 Modern tools can operate offline indefinitely and autonomously, | |
205 with a network connection only needed when syncing changes with | |
206 another repository.</para> | |
207 | |
208 </sect1> | |
209 <sect1> | |
210 <title>A few of the advantages of distributed revision | |
211 control</title> | |
212 | |
213 <para id="x_8d">Even though distributed revision control tools have for | |
214 several years been as robust and usable as their | |
215 previous-generation counterparts, people using older tools have | |
216 not yet necessarily woken up to their advantages. There are a | |
217 number of ways in which distributed tools shine relative to | |
218 centralised ones.</para> | |
219 | |
220 <para id="x_8e">For an individual developer, distributed tools are almost | |
221 always much faster than centralised tools. This is for a simple | |
222 reason: a centralised tool needs to talk over the network for | |
223 many common operations, because most metadata is stored in a | |
224 single copy on the central server. A distributed tool stores | |
225 all of its metadata locally. All else being equal, talking over | |
226 the network adds overhead to a centralised tool. Don't | |
227 underestimate the value of a snappy, responsive tool: you're | |
228 going to spend a lot of time interacting with your revision | |
229 control software.</para> | |
230 | |
231 <para id="x_8f">Distributed tools are indifferent to the vagaries of your | |
232 server infrastructure, again because they replicate metadata to | |
233 so many locations. If you use a centralised system and your | |
234 server catches fire, you'd better hope that your backup media | |
235 are reliable, and that your last backup was recent and actually | |
236 worked. With a distributed tool, you have many backups | |
237 available on every contributor's computer.</para> | |
238 | |
239 <para id="x_90">The reliability of your network will affect distributed | |
240 tools far less than it will centralised tools. You can't even | |
241 use a centralised tool without a network connection, except for | |
242 a few highly constrained commands. With a distributed tool, if | |
243 your network connection goes down while you're working, you may | |
244 not even notice. The only thing you won't be able to do is talk | |
245 to repositories on other computers, something that is relatively | |
246 rare compared with local operations. If you have a far-flung | |
247 team of collaborators, this may be significant.</para> | |
248 | |
249 <sect2> | |
250 <title>Advantages for open source projects</title> | |
251 | |
252 <para id="x_91">If you take a shine to an open source project and decide | |
253 that you would like to start hacking on it, and that project | |
254 uses a distributed revision control tool, you are at once a | |
255 peer with the people who consider themselves the | |
256 <quote>core</quote> of that project. If they publish their | |
257 repositories, you can immediately copy their project history, | |
258 start making changes, and record your work, using the same | |
259 tools in the same ways as insiders. By contrast, with a | |
260 centralised tool, you must use the software in a <quote>read | |
261 only</quote> mode unless someone grants you permission to | |
262 commit changes to their central server. Until then, you won't | |
263 be able to record changes, and your local modifications will | |
264 be at risk of corruption any time you try to update your | |
265 client's view of the repository.</para> | |
266 | |
267 <sect3> | |
268 <title>The forking non-problem</title> | |
269 | |
270 <para id="x_92">It has been suggested that distributed revision control | |
271 tools pose some sort of risk to open source projects because | |
272 they make it easy to <quote>fork</quote> the development of | |
273 a project. A fork happens when there are differences in | |
274 opinion or attitude between groups of developers that cause | |
275 them to decide that they can't work together any longer. | |
276 Each side takes a more or less complete copy of the | |
277 project's source code, and goes off in its own | |
278 direction.</para> | |
279 | |
280 <para id="x_93">Sometimes the camps in a fork decide to reconcile their | |
281 differences. With a centralised revision control system, the | |
282 <emphasis>technical</emphasis> process of reconciliation is | |
283 painful, and has to be performed largely by hand. You have | |
284 to decide whose revision history is going to | |
285 <quote>win</quote>, and graft the other team's changes into | |
286 the tree somehow. This usually loses some or all of one | |
287 side's revision history.</para> | |
288 | |
289 <para id="x_94">What distributed tools do with respect to forking is | |
290 they make forking the <emphasis>only</emphasis> way to | |
291 develop a project. Every single change that you make is | |
292 potentially a fork point. The great strength of this | |
293 approach is that a distributed revision control tool has to | |
294 be really good at <emphasis>merging</emphasis> forks, | |
295 because forks are absolutely fundamental: they happen all | |
296 the time.</para> | |
297 | |
298 <para id="x_95">If every piece of work that everybody does, all the | |
299 time, is framed in terms of forking and merging, then what | |
300 the open source world refers to as a <quote>fork</quote> | |
301 becomes <emphasis>purely</emphasis> a social issue. If | |
302 anything, distributed tools <emphasis>lower</emphasis> the | |
303 likelihood of a fork:</para> | |
304 <itemizedlist> | |
305 <listitem><para id="x_96">They eliminate the social distinction that | |
306 centralised tools impose: that between insiders (people | |
307 with commit access) and outsiders (people | |
308 without).</para></listitem> | |
309 <listitem><para id="x_97">They make it easier to reconcile after a | |
310 social fork, because all that's involved from the | |
311 perspective of the revision control software is just | |
312 another merge.</para></listitem></itemizedlist> | |
313 | |
314 <para id="x_98">Some people resist distributed tools because they want | |
315 to retain tight control over their projects, and they | |
316 believe that centralised tools give them this control. | |
317 However, if you're of this belief, and you publish your CVS | |
318 or Subversion repositories publicly, there are plenty of | |
319 tools available that can pull out your entire project's | |
320 history (albeit slowly) and recreate it somewhere that you | |
321 don't control. So while your control in this case is | |
322 illusory, you are forgoing the ability to fluidly | |
323 collaborate with whatever people feel compelled to mirror | |
324 and fork your history.</para> | |
325 | |
326 </sect3> | |
327 </sect2> | |
328 <sect2> | |
329 <title>Advantages for commercial projects</title> | |
330 | |
331 <para id="x_99">Many commercial projects are undertaken by teams that are | |
332 scattered across the globe. Contributors who are far from a | |
333 central server will see slower command execution and perhaps | |
334 less reliability. Commercial revision control systems attempt | |
335 to ameliorate these problems with remote-site replication | |
336 add-ons that are typically expensive to buy and cantankerous | |
337 to administer. A distributed system doesn't suffer from these | |
338 problems in the first place. Better yet, you can easily set | |
339 up multiple authoritative servers, say one per site, so that | |
340 there's no redundant communication between repositories over | |
341 expensive long-haul network links.</para> | |
342 | |
343 <para id="x_9a">Centralised revision control systems tend to have | |
344 relatively low scalability. It's not unusual for an expensive | |
345 centralised system to fall over under the combined load of | |
346 just a few dozen concurrent users. Once again, the typical | |
347 response tends to be an expensive and clunky replication | |
348 facility. Since the load on a central server&emdash;if you have | |
349 one at all&emdash;is many times lower with a distributed tool | |
350 (because all of the data is replicated everywhere), a single | |
351 cheap server can handle the needs of a much larger team, and | |
352 replication to balance load becomes a simple matter of | |
353 scripting.</para> | |
354 | |
355 <para id="x_9b">If you have an employee in the field, troubleshooting a | |
356 problem at a customer's site, they'll benefit from distributed | |
357 revision control. The tool will let them generate custom | |
358 builds, try different fixes in isolation from each other, and | |
359 search efficiently through history for the sources of bugs and | |
360 regressions in the customer's environment, all without needing | |
361 to connect to your company's network.</para> | |
362 | |
363 </sect2> | |
364 </sect1> | |
365 <sect1> | |
366 <title>Why choose Mercurial?</title> | |
367 | |
368 <para id="x_9c">Mercurial has a unique set of properties that make it a | |
369 particularly good choice as a revision control system.</para> | |
370 <itemizedlist> | |
371 <listitem><para id="x_9d">It is easy to learn and use.</para></listitem> | |
372 <listitem><para id="x_9e">It is lightweight.</para></listitem> | |
373 <listitem><para id="x_9f">It scales excellently.</para></listitem> | |
374 <listitem><para id="x_a0">It is easy to | |
375 customise.</para></listitem></itemizedlist> | |
376 | |
377 <para id="x_a1">If you are at all familiar with revision control systems, | |
378 you should be able to get up and running with Mercurial in less | |
379 than five minutes. Even if not, it will take no more than a few | |
380 minutes longer. Mercurial's command and feature sets are | |
381 generally uniform and consistent, so you can keep track of a few | |
382 general rules instead of a host of exceptions.</para> | |
383 | |
384 <para id="x_a2">On a small project, you can start working with Mercurial in | |
385 moments. Creating new changes and branches; transferring changes | |
386 around (whether locally or over a network); and history and | |
387 status operations are all fast. Mercurial attempts to stay | |
388 nimble and largely out of your way by combining low cognitive | |
389 overhead with blazingly fast operations.</para> | |
390 | |
391 <para id="x_a3">The usefulness of Mercurial is not limited to small | |
392 projects: it is used by projects with hundreds to thousands of | |
393 contributors, each containing tens of thousands of files and | |
394 hundreds of megabytes of source code.</para> | |
395 | |
396 <para id="x_a4">If the core functionality of Mercurial is not enough for | |
397 you, it's easy to build on. Mercurial is well suited to | |
398 scripting tasks, and its clean internals and implementation in | |
399 Python make it easy to add features in the form of extensions. | |
400 There are a number of popular and useful extensions already | |
401 available, ranging from helping to identify bugs to improving | |
402 performance.</para> | |
403 | |
404 </sect1> | |
405 <sect1> | |
406 <title>Mercurial compared with other tools</title> | |
407 | |
408 <para id="x_a5">Before you read on, please understand that this section | |
409 necessarily reflects my own experiences, interests, and (dare I | |
410 say it) biases. I have used every one of the revision control | |
411 tools listed below, in most cases for several years at a | |
412 time.</para> | |
413 | |
414 | |
415 <sect2> | |
416 <title>Subversion</title> | |
417 | |
418 <para id="x_a6">Subversion is a popular revision control tool, developed | |
419 to replace CVS. It has a centralised client/server | |
420 architecture.</para> | |
421 | |
422 <para id="x_a7">Subversion and Mercurial have similarly named commands for | |
423 performing the same operations, so if you're familiar with | |
424 one, it is easy to learn to use the other. Both tools are | |
425 portable to all popular operating systems.</para> | |
426 | |
427 <para id="x_a8">Prior to version 1.5, Subversion had no useful support for | |
428 merges. At the time of writing, its merge tracking capability | |
429 is new, and known to be <ulink | |
430 url="http://svnbook.red-bean.com/nightly/en/svn.branchmerge.advanced.html#svn.branchmerge.advanced.finalword">complicated | |
431 and buggy</ulink>.</para> | |
432 | |
433 <para id="x_a9">Mercurial has a substantial performance advantage over | |
434 Subversion on every revision control operation I have | |
435 benchmarked. I have measured its advantage as ranging from a | |
436 factor of two to a factor of six when compared with Subversion | |
437 1.4.3's <emphasis>ra_local</emphasis> file store, which is the | |
438 fastest access method available. In more realistic | |
439 deployments involving a network-based store, Subversion will | |
440 be at a substantially larger disadvantage. Because many | |
441 Subversion commands must talk to the server and Subversion | |
442 does not have useful replication facilities, server capacity | |
443 and network bandwidth become bottlenecks for modestly large | |
444 projects.</para> | |
445 | |
446 <para id="x_aa">Additionally, Subversion incurs substantial storage | |
447 overhead to avoid network transactions for a few common | |
448 operations, such as finding modified files | |
449 (<literal>status</literal>) and displaying modifications | |
450 against the current revision (<literal>diff</literal>). As a | |
451 result, a Subversion working copy is often the same size as, | |
452 or larger than, a Mercurial repository and working directory, | |
453 even though the Mercurial repository contains a complete | |
454 history of the project.</para> | |
455 | |
456 <para id="x_ab">Subversion is widely supported by third party tools. | |
457 Mercurial currently lags considerably in this area. This gap | |
458 is closing, however, and indeed some of Mercurial's GUI tools | |
459 now outshine their Subversion equivalents. Like Mercurial, | |
460 Subversion has an excellent user manual.</para> | |
461 | |
462 <para id="x_ac">Because Subversion doesn't store revision history on the | |
463 client, it is well suited to managing projects that deal with | |
464 lots of large, opaque binary files. If you check in fifty | |
465 revisions to an incompressible 10MB file, Subversion's | |
466 client-side space usage stays constant The space used by any | |
467 distributed SCM will grow rapidly in proportion to the number | |
468 of revisions, because the differences between each revision | |
469 are large.</para> | |
470 | |
471 <para id="x_ad">In addition, it's often difficult or, more usually, | |
472 impossible to merge different versions of a binary file. | |
473 Subversion's ability to let a user lock a file, so that they | |
474 temporarily have the exclusive right to commit changes to it, | |
475 can be a significant advantage to a project where binary files | |
476 are widely used.</para> | |
477 | |
478 <para id="x_ae">Mercurial can import revision history from a Subversion | |
479 repository. It can also export revision history to a | |
480 Subversion repository. This makes it easy to <quote>test the | |
481 waters</quote> and use Mercurial and Subversion in parallel | |
482 before deciding to switch. History conversion is incremental, | |
483 so you can perform an initial conversion, then small | |
484 additional conversions afterwards to bring in new | |
485 changes.</para> | |
486 | |
487 | |
488 </sect2> | |
489 <sect2> | |
490 <title>Git</title> | |
491 | |
492 <para id="x_af">Git is a distributed revision control tool that was | |
493 developed for managing the Linux kernel source tree. Like | |
494 Mercurial, its early design was somewhat influenced by | |
495 Monotone.</para> | |
496 | |
497 <para id="x_b0">Git has a very large command set, with version 1.5.0 | |
498 providing 139 individual commands. It has something of a | |
499 reputation for being difficult to learn. Compared to Git, | |
500 Mercurial has a strong focus on simplicity.</para> | |
501 | |
502 <para id="x_b1">In terms of performance, Git is extremely fast. In | |
503 several cases, it is faster than Mercurial, at least on Linux, | |
504 while Mercurial performs better on other operations. However, | |
505 on Windows, the performance and general level of support that | |
506 Git provides is, at the time of writing, far behind that of | |
507 Mercurial.</para> | |
508 | |
509 <para id="x_b2">While a Mercurial repository needs no maintenance, a Git | |
510 repository requires frequent manual <quote>repacks</quote> of | |
511 its metadata. Without these, performance degrades, while | |
512 space usage grows rapidly. A server that contains many Git | |
513 repositories that are not rigorously and frequently repacked | |
514 will become heavily disk-bound during backups, and there have | |
515 been instances of daily backups taking far longer than 24 | |
516 hours as a result. A freshly packed Git repository is | |
517 slightly smaller than a Mercurial repository, but an unpacked | |
518 repository is several orders of magnitude larger.</para> | |
519 | |
520 <para id="x_b3">The core of Git is written in C. Many Git commands are | |
521 implemented as shell or Perl scripts, and the quality of these | |
522 scripts varies widely. I have encountered several instances | |
523 where scripts charged along blindly in the presence of errors | |
524 that should have been fatal.</para> | |
525 | |
526 <para id="x_b4">Mercurial can import revision history from a Git | |
527 repository.</para> | |
528 | |
529 | |
530 </sect2> | |
531 <sect2> | |
532 <title>CVS</title> | |
533 | |
534 <para id="x_b5">CVS is probably the most widely used revision control tool | |
535 in the world. Due to its age and internal untidiness, it has | |
536 been only lightly maintained for many years.</para> | |
537 | |
538 <para id="x_b6">It has a centralised client/server architecture. It does | |
539 not group related file changes into atomic commits, making it | |
540 easy for people to <quote>break the build</quote>: one person | |
541 can successfully commit part of a change and then be blocked | |
542 by the need for a merge, causing other people to see only a | |
543 portion of the work they intended to do. This also affects | |
544 how you work with project history. If you want to see all of | |
545 the modifications someone made as part of a task, you will | |
546 need to manually inspect the descriptions and timestamps of | |
547 the changes made to each file involved (if you even know what | |
548 those files were).</para> | |
549 | |
550 <para id="x_b7">CVS has a muddled notion of tags and branches that I will | |
551 not attempt to even describe. It does not support renaming of | |
552 files or directories well, making it easy to corrupt a | |
553 repository. It has almost no internal consistency checking | |
554 capabilities, so it is usually not even possible to tell | |
555 whether or how a repository is corrupt. I would not recommend | |
556 CVS for any project, existing or new.</para> | |
557 | |
558 <para id="x_b8">Mercurial can import CVS revision history. However, there | |
559 are a few caveats that apply; these are true of every other | |
560 revision control tool's CVS importer, too. Due to CVS's lack | |
561 of atomic changes and unversioned filesystem hierarchy, it is | |
562 not possible to reconstruct CVS history completely accurately; | |
563 some guesswork is involved, and renames will usually not show | |
564 up. Because a lot of advanced CVS administration has to be | |
565 done by hand and is hence error-prone, it's common for CVS | |
566 importers to run into multiple problems with corrupted | |
567 repositories (completely bogus revision timestamps and files | |
568 that have remained locked for over a decade are just two of | |
569 the less interesting problems I can recall from personal | |
570 experience).</para> | |
571 | |
572 <para id="x_b9">Mercurial can import revision history from a CVS | |
573 repository.</para> | |
574 | |
575 | |
576 </sect2> | |
577 <sect2> | |
578 <title>Commercial tools</title> | |
579 | |
580 <para id="x_ba">Perforce has a centralised client/server architecture, | |
581 with no client-side caching of any data. Unlike modern | |
582 revision control tools, Perforce requires that a user run a | |
583 command to inform the server about every file they intend to | |
584 edit.</para> | |
585 | |
586 <para id="x_bb">The performance of Perforce is quite good for small teams, | |
587 but it falls off rapidly as the number of users grows beyond a | |
588 few dozen. Modestly large Perforce installations require the | |
589 deployment of proxies to cope with the load their users | |
590 generate.</para> | |
591 | |
592 | |
593 </sect2> | |
594 <sect2> | |
595 <title>Choosing a revision control tool</title> | |
596 | |
597 <para id="x_bc">With the exception of CVS, all of the tools listed above | |
598 have unique strengths that suit them to particular styles of | |
599 work. There is no single revision control tool that is best | |
600 in all situations.</para> | |
601 | |
602 <para id="x_bd">As an example, Subversion is a good choice for working | |
603 with frequently edited binary files, due to its centralised | |
604 nature and support for file locking.</para> | |
605 | |
606 <para id="x_be">I personally find Mercurial's properties of simplicity, | |
607 performance, and good merge support to be a compelling | |
608 combination that has served me well for several years.</para> | |
609 | |
610 | |
611 </sect2> | |
612 </sect1> | |
613 <sect1> | |
614 <title>Switching from another tool to Mercurial</title> | |
615 | |
616 <para id="x_bf">Mercurial is bundled with an extension named <literal | |
617 role="hg-ext">convert</literal>, which can incrementally | |
618 import revision history from several other revision control | |
619 tools. By <quote>incremental</quote>, I mean that you can | |
620 convert all of a project's history to date in one go, then rerun | |
621 the conversion later to obtain new changes that happened after | |
622 the initial conversion.</para> | |
623 | |
624 <para id="x_c0">The revision control tools supported by <literal | |
625 role="hg-ext">convert</literal> are as follows:</para> | |
626 <itemizedlist> | |
627 <listitem><para id="x_c1">Subversion</para></listitem> | |
628 <listitem><para id="x_c2">CVS</para></listitem> | |
629 <listitem><para id="x_c3">Git</para></listitem> | |
630 <listitem><para id="x_c4">Darcs</para></listitem></itemizedlist> | |
631 | |
632 <para id="x_c5">In addition, <literal role="hg-ext">convert</literal> can | |
633 export changes from Mercurial to Subversion. This makes it | |
634 possible to try Subversion and Mercurial in parallel before | |
635 committing to a switchover, without risking the loss of any | |
636 work.</para> | |
637 | |
638 <para id="x_c6">The <command role="hg-ext-convert">convert</command> command | |
639 is easy to use. Simply point it at the path or URL of the | |
640 source repository, optionally give it the name of the | |
641 destination repository, and it will start working. After the | |
642 initial conversion, just run the same command again to import | |
643 new changes.</para> | |
644 </sect1> | |
645 | |
646 <sect1> | |
647 <title>A short history of revision control</title> | |
648 | |
649 <para id="x_c7">The best known of the old-time revision control tools is | |
650 SCCS (Source Code Control System), which Marc Rochkind wrote at | |
651 Bell Labs, in the early 1970s. SCCS operated on individual | |
652 files, and required every person working on a project to have | |
653 access to a shared workspace on a single system. Only one | |
654 person could modify a file at any time; arbitration for access | |
655 to files was via locks. It was common for people to lock files, | |
656 and later forget to unlock them, preventing anyone else from | |
657 modifying those files without the help of an | |
658 administrator.</para> | |
659 | |
660 <para id="x_c8">Walter Tichy developed a free alternative to SCCS in the | |
661 early 1980s; he called his program RCS (Revision Control System). | |
662 Like SCCS, RCS required developers to work in a single shared | |
663 workspace, and to lock files to prevent multiple people from | |
664 modifying them simultaneously.</para> | |
665 | |
666 <para id="x_c9">Later in the 1980s, Dick Grune used RCS as a building block | |
667 for a set of shell scripts he initially called cmt, but then | |
668 renamed to CVS (Concurrent Versions System). The big innovation | |
669 of CVS was that it let developers work simultaneously and | |
670 somewhat independently in their own personal workspaces. The | |
671 personal workspaces prevented developers from stepping on each | |
672 other's toes all the time, as was common with SCCS and RCS. Each | |
673 developer had a copy of every project file, and could modify | |
674 their copies independently. They had to merge their edits prior | |
675 to committing changes to the central repository.</para> | |
676 | |
677 <para id="x_ca">Brian Berliner took Grune's original scripts and rewrote | |
678 them in C, releasing in 1989 the code that has since developed | |
679 into the modern version of CVS. CVS subsequently acquired the | |
680 ability to operate over a network connection, giving it a | |
681 client/server architecture. CVS's architecture is centralised; | |
682 only the server has a copy of the history of the project. Client | |
683 workspaces just contain copies of recent versions of the | |
684 project's files, and a little metadata to tell them where the | |
685 server is. CVS has been enormously successful; it is probably | |
686 the world's most widely used revision control system.</para> | |
687 | |
688 <para id="x_cb">In the early 1990s, Sun Microsystems developed an early | |
689 distributed revision control system, called TeamWare. A | |
690 TeamWare workspace contains a complete copy of the project's | |
691 history. TeamWare has no notion of a central repository. (CVS | |
692 relied upon RCS for its history storage; TeamWare used | |
693 SCCS.)</para> | |
694 | |
695 <para id="x_cc">As the 1990s progressed, awareness grew of a number of | |
696 problems with CVS. It records simultaneous changes to multiple | |
697 files individually, instead of grouping them together as a | |
698 single logically atomic operation. It does not manage its file | |
699 hierarchy well; it is easy to make a mess of a repository by | |
700 renaming files and directories. Worse, its source code is | |
701 difficult to read and maintain, which made the <quote>pain | |
702 level</quote> of fixing these architectural problems | |
703 prohibitive.</para> | |
704 | |
705 <para id="x_cd">In 2001, Jim Blandy and Karl Fogel, two developers who had | |
706 worked on CVS, started a project to replace it with a tool that | |
707 would have a better architecture and cleaner code. The result, | |
708 Subversion, does not stray from CVS's centralised client/server | |
709 model, but it adds multi-file atomic commits, better namespace | |
710 management, and a number of other features that make it a | |
711 generally better tool than CVS. Since its initial release, it | |
712 has rapidly grown in popularity.</para> | |
713 | |
714 <para id="x_ce">More or less simultaneously, Graydon Hoare began working on | |
715 an ambitious distributed revision control system that he named | |
716 Monotone. While Monotone addresses many of CVS's design flaws | |
717 and has a peer-to-peer architecture, it goes beyond earlier (and | |
718 subsequent) revision control tools in a number of innovative | |
719 ways. It uses cryptographic hashes as identifiers, and has an | |
720 integral notion of <quote>trust</quote> for code from different | |
721 sources.</para> | |
722 | |
723 <para id="x_cf">Mercurial began life in 2005. While a few aspects of its | |
724 design are influenced by Monotone, Mercurial focuses on ease of | |
725 use, high performance, and scalability to very large | |
726 projects.</para> | |
727 </sect1> | |
728 </chapter> | |
729 | |
730 <!-- | |
731 local variables: | |
732 sgml-parent-document: ("00book.xml" "book" "chapter") | |
733 end: | |
734 --> |