1 {% extends "templates/_base.html" %}
3 {% block file_metadata %}
4 <meta name="title" content="Version Control With Subversion" />
5 {% endblock file_metadata %}
9 <li><a href="#s:basics">Basic Use</a></li>
10 <li><a href="#s:merge">Merging Conflicts</a></li>
11 <li><a href="#s:rollback">Recovering Old Versions</a></li>
12 <li><a href="#s:setup">Setting up a Repository</a></li>
13 <li><a href="#s:provenance">Provenance</a></li>
14 <li><a href="#s:summary">Summing Up</a></li>
18 Wolfman and Dracula have been hired by Universal Missions
19 (a space services spinoff from Euphoric State University)
20 to figure out where the company should send its next planetary lander.
21 They want to be able to work on the plans at the same time,
22 but they have run into problems doing this in the past.
24 each one will spend a lot of time waiting for the other to finish.
26 if they work on their own copies and email changes back and forth
27 they know that things will be lost, overwritten, or duplicated.
31 The right solution is to use a
32 <a href="glossary.html#version-control-system">version control system</a>
34 Version control is better than mailing files back and forth because:
40 It's hard (but not impossible) to accidentally overlook or overwrite someone's changes,
41 because the version control system highlights them automatically.
45 It keeps a record of who made what changes when,
46 so that if people have questions later on,
52 Nothing that is committed to version control is ever lost.
53 This means it can be used like the "undo" feature in an editor,
54 and since all old versions of files are saved
55 it's always possible to go back in time to see exactly who wrote what on a particular day,
56 or what version of a program was used to generate a particular set of results.
62 The rest of this chapter will explore how to use
63 a popular open source version control system called Subversion.
64 It does not have all the features of some newer systems,
65 such as <a href="git.html">Git</a>,
66 but it is still widely used,
67 and is simpler to pick up than those more advanced alternatives.
68 No matter which system you use,
69 the most important thing to learn is not the details of their more obscure commands,
70 but the workflow that they encourage.
74 <h2>For Instructors</h2>
77 Version control is the most important practical skill we introduce.
78 As the last paragraph of the introduction above says,
79 the workflow matters more than the ins and outs of any particular tool.
80 By the end of 90 minutes,
81 the instructor should be able to get learners to chant,
82 "Update, edit, merge, commit," in unison,
83 and have them understand what those terms mean
84 and why that's a good way to structure their working day.
88 Provided there aren't network problems,
89 this entire lesson can be covered in <span class="duration">90 minutes</span>.
90 The example at the end
91 showing how to use Subversion keywords to track provenance
92 is the "ah ha!" moment for many learners.
94 skip the material on recovering old versions of files
95 in order to get to this section instead.
96 (The fact that provenance is harder in Git,
97 both mechanically and conceptually,
98 is one reason to keep teaching Subversion.)
102 <h3>Prerequisites</h3>
104 Basic shell concepts and skills
105 (<code>ls</code>, <code>cd</code>, <code>mkdir</code>,
107 basic shell scripting
108 (for the discussion of <a href="#s:provenance">provenance</a>).
113 <h3>Teaching Notes</h3>
116 Make sure the network is working <em>before</em> starting this lesson.
119 Give learners a ten-minute overview of what version control does for them
120 before diving into the watch-and-do practicals.
121 Most of them will have tried to co-author papers by emailing files back and forth,
122 or will have biked into the office
123 only to realize that the USB key with last night's work
124 is still on the kitchen table.
125 Instructors can also make jokes about directories with names like
127 "final version revised",
128 "final version with reviewer three's corrections",
129 "really final version",
131 "come on this really has to be the last version"
132 to motivate version control as a better way to collaborate
133 and as a better way to back work up.
136 Version control is typically taught after the shell,
137 so collect learners' names during that session
138 and create a repository for them to share
139 with their names as both their IDs and their passwords.
140 The easiest way to create the repository is to use
141 a server managed by an ISP such as Dreamhost,
142 or on SourceForge, Google Code, or some other "forge" site,
143 all of which provide web interfaces for repository creation and management.
144 If your learners are advanced enough to be using SSH,
145 you can instead create it on any server they can access,
146 and connect with the <code>svn+ssh</code> protocol instead of HTTPS.
149 Be very clear what files learners are to edit
150 and what user IDs they are to use
151 when giving instructions.
152 It is common for them to edit the instructor's biography,
153 or to use the instructor's user ID and password when committing.
154 Be equally clear <em>when</em> they are to edit things:
155 it's also common for someone to edit the file the instructor is editing
156 and commit changes while the instructor is explaining what's going on,
157 so that a conflict occurs when the instructor comes to commit the file.
160 Learners could do most exercises with repositories on their own machines,
161 but it's hard for them to see how version control helps collaboration
162 unless they're sharing a repository with other learners.
164 showing learners who changed what using <code>svn blame</code>
165 is only compelling if a file has been edited by at least two people.
168 If some learners are using Windows,
169 there will inevitably be issues merging files with different line endings.
170 <code>svn diff -x -w</code> is supposed to suppress differences in whitespace,
171 but we have found that it doesn't always work as advertised.
178 <section id="s:basics">
181 <div class="understand">
182 <h3>Learning Objectives</h3>
184 <li>Draw a diagram showing the places version control stores information.</li>
185 <li>Check out a working copy of a repository.</li>
186 <li>View the history of changes to a project.</li>
187 <li>Explain why working copies of different projects should not overlap.</li>
188 <li>Add files to a project.</li>
189 <li>Commit changes made to a working copy to a repository.</li>
190 <li>Update a working copy to get changes from the repository.</li>
191 <li>Compare the current state of a working copy to the last update from the repository, and to the current state of the repository.</li>
192 <li>Explain what "version 123 of <code>xyz.txt</code>" actually means.</li>
197 A version control system keeps the master copy of a file
198 in a <a href="glossary.html#repository">repository</a>
199 located on a <a href="glossary.html#server">server</a>—a computer
200 that is never used directly by people,
201 but only by their programs
202 (<a href="#f:repository">Figure 1</a>).
203 No-one ever edits the master copy directly.
205 Wolfman and Dracula each have a <a href="glossary.html#working-copy">working copy</a>
206 on their own machines.
207 They can each edit their working copies whenever and however they want.
210 <figure id="f:repository">
211 <img src="svn/repository.png" alt="Repositories and Working Copies" />
212 <figcaption>Figure 1: Repositories and Working Copies</figcaption>
216 When Wolfman is ready to share his changes with Dracula,
217 he <a href="glossary.html#commit">commits</a> his work to the repository
218 (<a href="#f:workflow">Figure 2</a>).
219 Dracula can then <a href="glossary.html#update">update</a> his working copy
220 to get those changes when he's ready for them.
222 when Dracula finishes working on something,
223 he can commit and so that Wolfman can update.
226 <figure id="f:workflow">
227 <img src="svn/workflow.png" alt="Sharing Files Through Version Control" />
228 <figcaption>Figure 2: Sharing Files Through Version Control</figcaption>
232 If this is all there was to version control,
233 it would be no better than FTP or Dropbox.
234 But what if Dracula and Wolfman change their working copies at the same time?
235 If Wolfman commits first,
236 his changes are simply copied to the repository
237 (<a href="#f:merge_first_commit">Figure 3</a>):
240 <figure id="f:merge_first_commit">
241 <img src="svn/merge_first_commit.png" alt="Wolfman Commits First" />
242 <figcaption>Figure 3: Wolfman Commits First</figcaption>
246 If Dracula now tries to commit something that would overwrite Wolfman's changes
247 the version control system detects the <a href="glossary.html#conflict">conflict</a>,
249 and tells Dracula that there's a problem
250 (<a href="#f:merge_second_commit">Figure 4</a>):
253 <figure id="f:merge_second_commit">
254 <img src="svn/merge_second_commit.png" alt="Dracula Has a Conflict" />
255 <figcaption>Figure 4: Dracula Has a Conflict</figcaption>
259 Dracula must <a href="glossary.html#resolve">resolve</a> that conflict
260 before the version control system will allow him to commit his work.
261 He can accept what Wolfman did,
262 replace it with what he has done,
263 or write something new that combines the two—that's up to him
264 (<a href="#f:merge_resolve">Figure 5</a>).
265 Once he has cleaned things up, he can go ahead and try committing again.
266 If all of the conflicts have been resolved,
267 the version control will accept it this time.
270 <figure id="f:merge_resolve">
271 <img src="svn/merge_resolve.png" alt="Resolving the Conflict" />
272 <figcaption>Figure 5: Resolving the Conflict</figcaption>
276 <h3>Forgiveness vs. Permission</h3>
279 Old-fashioned version control systems prevented conflicts from happening
280 by <a href="glossary.html#lock">locking</a> the master copy
281 whenever someone was working on it.
282 This <a href="glossary.html#pessimistic-concurrency">pessimistic</a> strategy
283 guaranteed that a second person (or monster)
284 could never make changes to the same file at the same time,
285 but it also meant that people had to take turns editing files.
289 Most of today's version control systems use
290 an <a href="glossary.html#optimistic-concurrency">optimistic</a> strategy instead:
291 people are always allowed to edit their working copies,
292 and if a conflict occurs,
293 the version control system helps them sort it out after the fact.
298 To see how this actually works,
299 let's assume that the Mummy
300 (Dracula and Wolfman's boss)
301 has already put some notes in a version control repository
302 whose URL is <code>https://universal.software-carpentry.org/explore</code>.
303 Every repository has an address like this that uniquely identifies the location of the master copy.
307 <h3>There's More Than One Way To Do It</h3>
310 We will drive Subversion from the command line in our examples,
311 but if you prefer using a GUI,
312 there are many for you to choose from.
313 Please see the <a href="ref.html#s:svn:gui">reference</a> for links.
319 and Dracula has just joined the project.
320 In order to get a working copy on his computer,
321 Dracula has to <a href="glossary.html#check-out">check out</a> a copy of the repository.
322 He only has to do this once per project:
323 once he has a working copy,
324 he can update it over and over again to get other people's work.
328 While in his home directory,
329 Dracula types the command:
333 $ <span class="in">svn checkout https://universal.software-carpentry.org/explore</span>
337 This creates a new directory called <code>explore</code>
338 and fills it with a copy of the repository's contents
339 (<a href="#f:example_repo">Figure 6</a>).
343 <span class="out">A explore/jupiter
345 A explore/mars/mons-olympus.txt
346 A explore/mars/cydonia.txt
348 A explore/earth/himalayas.txt
349 A explore/earth/antarctica.txt
350 A explore/earth/carlsbad.txt
351 Checked out revision 6.</span>
354 <figure id="f:example_repo">
355 <img src="svn/example_repo.png" alt="Example Repository" />
356 <figcaption>Figure 6: Example Repository</figcaption>
360 Dracula can then go into this directory
361 and use regular shell commands to view the files:
365 $ <span class="in">cd explore</span>
366 $ <span class="in">ls</span>
367 <span class="out">earth jupiter mars</span>
368 $ <span class="in">ls *</span>
369 <span class="out">earth:
370 antarctica.txt carlsbad.txt himalayas.txt
375 cydonia.txt mons-olympus.txt</span>
379 <h3>Don't Let the Working Copies Overlap</h3>
382 It's very important that the working copies of different project do not overlap;
384 we should never try to check out one project inside a working copy of another project.
385 The reason is that Subversion stories information about
386 the current state of a working copy
387 in special sub-directories called <code>.svn</code>:
391 $ <span class="in">pwd</span>
392 <span class="out">/home/dracula/explore</span>
393 $ <span class="in">ls -a</span>
394 <span class="out">. .. .svn earth jupiter mars</span>
395 $ <span class="in">ls -F .svn</span>
396 <span class="out">entries prop-base/ props/ text-base/ tmp/</span>
400 If two working copies overlap,
401 the files in the <code>.svn</code> directories for one repository
402 will be clobbered by the other repository's <code>.svn</code> files,
403 and Subversion will become hopelessly confused.
408 Dracula can find out more about the history of the project
409 using Subversion's <code>log</code> command:
413 $ <span class="in">svn log</span>
414 <span class="out">------------------------------------------------------------------------
415 r6 | mummy | 2010-07-26 09:21:10 -0400 (Mon, 26 Jul 2010) | 1 line
417 Damn the budget---the Jovian moons would be a _perfect_ place to explore.
418 ------------------------------------------------------------------------
419 r5 | mummy | 2010-07-26 09:19:39 -0400 (Mon, 26 Jul 2010) | 1 line
421 The budget might not even stretch to the Arctic :-(
422 ------------------------------------------------------------------------
423 r4 | mummy | 2010-07-26 09:17:46 -0400 (Mon, 26 Jul 2010) | 1 line
425 Budget cuts may force us to do another dry run in the Arctic.
426 ------------------------------------------------------------------------
427 r3 | mummy | 2010-07-26 09:14:14 -0400 (Mon, 26 Jul 2010) | 1 line
429 Converting document to wiki-formatted text.
430 ------------------------------------------------------------------------
431 r2 | mummy | 2010-07-26 09:11:55 -0400 (Mon, 26 Jul 2010) | 1 line
433 Or put it down near the Face of Cydonia?
434 ------------------------------------------------------------------------
435 r1 | mummy | 2010-07-26 09:08:23 -0400 (Mon, 26 Jul 2010) | 1 line
437 Send the probe to Mons Olympus?
438 ------------------------------------------------------------------------</span>
442 Subversion displays a summary of all the changes made to the project so far.
443 This list includes the
444 <a href="glossary.html#revision-number">revision number</a>,
445 the name of the person who made the change,
446 the date the change was made,
447 and whatever comment the user provided when the change was submitted.
449 the <code>explore</code> project is currently at revision 6,
450 and all changes so far have been made by the Mummy.
454 Notice how detailed the comments on the updates are.
455 Good comments are as important in version control as they are in coding.
456 Without them, it can be very difficult to figure out who did what, when, and why.
457 We can use comments like "Changed things" and "Fixed it" if we want,
458 or even no comments at all,
459 but we'll only be making more work for our future selves.
463 <h3>Numbering Versions</h3>
466 Another thing to notice is that the revision number applies to the whole repository,
467 not to a particular file.
468 When we talk about "version 61" we mean
469 "the state of all files and directories at that point."
470 Older version control systems like CVS gave each file a new version number when it was updated,
471 which meant that version 38 of one file could correspond in time to version 17 of another
472 (<a href="#f:version_numbering">Figure 7</a>).
473 Experience shows that
474 global version numbers that apply to everything in the repository
475 are easier to manage than
476 per-file version numbers,
477 so that's what Subversion uses.
480 <figure id="f:version_numbering">
481 <img src="svn/version_numbering.png" alt="Version Numbering Schemes" />
482 <figcaption>Figure 7: Version Numbering Schemes</figcaption>
487 A couple of cubicles away,
488 Wolfman also runs <code>svn checkout</code>
489 to get a working copy of the repository.
490 He also gets version 6,
491 so the files on his machine are the same as the files on Dracula's.
492 While he is looking through the files,
493 Dracula decides to add some information to the repository about Jupiter's moons.
494 Using his favorite editor,
495 he creates a file in the <code>jupiter</code> directory called <code>moons.txt</code>,
496 and fills it with information about Io, Europa, Ganymede, and Callisto:
499 <pre src="svn/moons_initial.txt">
500 Name Orbital Radius Orbital Period Mass Radius
501 Io 421.6 1.769138 893.2 1821.6
502 Europa 670.9 3.551181 480.0 1560.8
503 Ganymede 1070.4 7.154553 1481.9 2631.2
504 Calisto 1882.7 16.689018 1075.9 2410.3
508 After double-checking his data,
509 he wants to commit the file to the repository so that everyone else on the project can see it.
510 The first step is to add the file to his working copy using <code>svn add</code>:
514 $ <span class="in">svn add jupiter/moons.txt</span>
515 <span class="out">A jupiter/moons.txt</span>
519 Adding a file is not the same as creating it—he has already done that.
521 the <code>svn add</code> command tells Subversion to add the file to
522 the list of things it's supposed to manage.
524 particularly in programming projects,
525 to have backup files or intermediate files in a directory
526 that aren't worth storing in the repository.
527 This is why version control requires us to explicitly tell it which files are to be managed.
531 Once he has told Subversion to add the file,
532 Dracula can go ahead and commit his changes to the repository.
533 He uses the <code>-m</code> flag to provide a one-line message explaining what he's doing;
535 Subversion would open his default editor
536 so that he could type in something longer.
540 $ <span class="in">svn commit -m "Some basic facts about the Galilean moons of Jupiter." jupiter/moons.txt</span>
541 <span class="out">Adding jupiter/moons.txt
542 Transmitting file data .
543 Committed revision 7.</span>
547 When Dracula runs the <code>svn commit</code> command,
548 Subversion establishes a connection to the server,
549 copies over his changes,
550 and updates the revision number from 6 to 7
551 (<a href="#f:updated_repo">Figure 8</a>).
554 <figure id="f:updated_repo">
555 <img src="svn/updated_repo.png" alt="Updated Repository" />
556 <figcaption>Figure 8: Updated Repository</figcaption>
560 <h3>When <em>Not</em> to Use Version Control</h3>
563 Despite the rapidly decreasing cost of storage,
564 it is still possible to run out of disk space.
566 people can easy go through 2 TB/month if they're not careful.
567 Since version control tools usually store revisions in terms of lines,
568 with binary data files,
569 they end up essentially storing every revision separately.
571 (it's what we'd be doing anyway),
572 but it means version control isn't doing what it likes to do,
573 and the repository can get very large very quickly.
574 Another concern is that if very old data will no longer be used,
575 it can be nice to archive or delete old data files.
576 This is not possible if our data is version controlled:
577 information can only be added to a repository,
578 so it can only ever increase in size.
583 <p id="a:define-head">
585 Wolfman uses <code>svn update</code> to update his working copy.
586 It tells him that a new file has been added
587 and brings his working copy up to date with version 7 of the repository,
588 because this is now the most recent revision
589 (also called the <a href="glossary.html#head">head</a>).
590 <code>svn update</code> updates an existing working copy,
591 rather than checking out a new one.
592 While <code>svn checkout</code> is usually only run once per project per machine,
593 <code>svn update</code> may be run many times a day.
597 Looking in the new file <code>jupiter/moons.txt</code>,
598 Wolfman notices that Dracula has misspelled "Callisto"
599 (it is supposed to have two L's.)
600 Wolfman edits that line of the file:
603 <pre src="svn/moons_spelling.txt">
604 Name Orbital Radius Orbital Period Mass Radius
605 Io 421.6 1.769138 893.2 1821.6
606 Europa 670.9 3.551181 480.0 1560.8
607 Ganymede 1070.4 7.154553 1481.9 2631.2
608 <span class="highlight">Callisto 1882.7 16.689018 1075.9 2410.3</span>
612 He also adds a line about Amalthea,
613 which he thinks might be an interesting place to send a probe
614 despite its small size:
617 <pre src="svn/moons_amalthea.txt">
618 Name Orbital Radius Orbital Period Mass Radius
619 <span class="highlight">Amalthea 181.4 0.498179 0.075 125.0</span>
620 Io 421.6 1.769138 893.2 1821.6
621 Europa 670.9 3.551181 480.0 1560.8
622 Ganymede 1070.4 7.154553 1481.9 2631.2
623 Callisto 1882.7 16.689018 1075.9 2410.3
628 he uses the <code>svn status</code> command to check that he hasn't accidentally changed anything else:
632 $ <span class="in">svn status</span>
633 <span class="out">M jupiter/moons.txt</span>
637 and then runs <code>svn commit</code>.
638 Since has hasn't used the <code>-m</code> flag to provide a message on the command line,
639 Subversion launches his default editor and shows him:
644 --This line, and those below, will be ignored--
650 He changes this to be
654 1. Fixed typo in moon's name: 'Calisto' -> 'Callisto'.
655 2. Added information about Amalthea.
656 --This line, and those below, will be ignored--
662 When he saves this temporary file and exits the editor,
663 Subversion commits his changes:
667 <span class="out">Sending jupiter/moons.txt
668 Transmitting file data .
669 Committed revision 8.</span>
673 Note that since Wolfman didn't specify a particular file to commit,
674 Subversion commits <em>all</em> of his changes.
675 This is why he ran the <code>svn status</code> command first.
679 <h3>Which Editor?</h3>
681 If you don't have a default editor set up,
682 Subversion will probably open an editor called Vi.
684 type escape-colon-w-q-! to exit
685 and hope it never happens again.
689 <div class="box" id="b:basics:transaction">
690 <h3>Working With Multiple Files</h3>
693 Our example only includes one file,
694 but version control can work on any number of files at once.
696 if Wolfman noticed that a dozen data files had the same incorrect header,
697 he could change it in all 12 files,
698 then commit all those changes at once.
699 This is actually the best way to work:
700 every logical change to the project should be a single commit,
701 and every commit should include everything involved in one logical change.
708 Dracula wants to synchronize with Wolfman's work.
709 Before updating his working copy with <code>svn update</code>,
711 he checks to see if he has made any changes locally
712 by running <code>svn diff</code>.
714 it compares what's in his working copy to what he got the last time he updated.
715 There are no differences,
716 so there's no output:
720 $ <span class="in">svn diff</span>
725 To compare his working copy to the master,
726 Dracula uses <code>svn diff -r HEAD</code>.
727 The <code>-r</code> flag is used to specify a revision,
728 while <code>HEAD</code> means
729 "<a href="#a:define-head">the latest version of the master</a>".
733 $ <span class="in">svn diff -r HEAD</span>
734 <span class="out">--- moons.txt(revision 8)
735 +++ moons.txt(working copy)
737 Name Orbital Radius Orbital Period Mass Radius
738 +Amalthea 181.4 0.498179 0.075 125.0
739 Io 421.6 1.769138 893.2 1821.6
740 Europa 670.9 3.551181 480.0 1560.8
741 Ganymede 1070.4 7.154553 1481.9 2631.2
742 -Calisto 1882.7 16.689018 1075.9 2410.3
743 +Callisto 1882.7 16.689018 1075.9 2410.3
748 After looking over the changes,
749 Dracula goes ahead and does the update.
753 <h3>Reading a Diff</h3>
756 The output of <code>diff</code> is cryptic even by Unix standards.
761 --- moons.txt(revision 9)
762 +++ moons.txt(working copy)
766 signal that '-' will be used to show content from revision 9
767 and '+' to show content from the user's working copy.
768 The next line, with the '@' markers,
769 indicates where lines were inserted or removed.
770 This isn't really intended for human consumption:
771 editors and other tools can use this information
772 to replay a series of edits against a file.
776 The most important parts of what follows are the lines marked with '+' and '-',
777 which show insertions and deletions respectively.
779 we can see that the line for Amalthea was inserted,
780 and that the line for Callisto was changed
781 (which is indicated by an add and a delete right next to one another).
782 Many editors and other tools can display diffs like this in a two-column display,
783 highlighting changes.
789 <h3>Nothing's Perfekt</h3>
792 Version control systems do have one important shortcoming.
793 While it is easy for them to find, display, and merge differences in text files,
794 images, MP3s, PDFs, or Microsoft Word or Excel files aren't stored as text—they
795 use specialized binary data formats.
796 Most version control systems don't know how to deal with these formats,
797 so all they can say is, "These files differ."
798 Reconciling those differences will probably require use of an auxiliary tool,
799 such as an audio editor
800 or Microsoft Word's "Compare and Merge" utility.
805 <h3>Diffing Other Files</h3>
808 <code>svn diff</code> mimics the behavior of
809 the Unix <code>diff</code> command,
810 which can be used to compare any two files.
811 Given these two files:
816 <th><code>left.txt</code></th>
817 <th><code>right.txt</code></th>
839 <code>diff</code>'s output is:
842 $ <span class="in">diff left.txt right.txt</span>
843 <span class="out">2a3
850 > strontium</span>
855 This is a very common workflow,
856 and is the basic heartbeat of most developers' days.
863 Update our working copy
864 so that we have any changes other people have committed.
872 Commit our changes to the repository
873 so that other people can get them.
879 It's worth noticing here how important Wolfman's comments about his changes were.
880 It's hard to see the difference between "Calisto" with one 'L' and "Callisto" with two,
881 even if the line containing the difference has been highlighted.
882 Without Wolfman's comments,
883 Dracula might have wasted time wondering what the difference was.
888 Wolfman should probably have committed his two changes separately,
889 since there's no logical connection between
890 fixing a typo in Callisto's name
891 and adding information about Amalthea to the same file.
892 Just as a function or program should do one job and one job only,
893 a single commit to version control should have a single logical purpose so that it's easier to find,
895 and if necessary undo later on.
899 <h3>Who Did What?</h3>
902 One other very useful command is <code>svn blame</code>,
903 which shows when each line in the file was last changed
908 $ <span class="in">svn blame moons.txt</span>
909 <span class="out"> 14 dracula Name Orbital Radius Orbital Period Mass Radius
910 14 dracula (10**3 km) (days) (10**20 kg) (km)
911 14 dracula Amalthea 181.4 0.498179 0.075 131 x 73 x 67
912 9 mummy Io 421.6 1.769138 893.2 1821.6
913 9 mummy Europa 670.9 3.551181 480.0 1560.8
914 9 mummy Ganymede 1070.4 7.154553 1481.9 2631.2
915 14 dracula Callisto 1882.7 16.689018 1075.9 2410.3
916 14 dracula Himalia 11460 250.5662 0.095 85.0
917 14 dracula Elara 11740 259.6528 0.008 40.0</span>
921 If you are ever wondering who to talk to about a change,
923 <code>svn blame</code> is a good place to start.
927 <div class="keypoints">
930 <li>Version control is a better way to manage shared files than email or shared folders.</li>
931 <li>The master copy is stored in a repository.</li>
932 <li>Nobody ever edits the master directory: instead, each person edits a local working copy.</li>
933 <li>People share changes by committing them to the master or updating their local copy from the master.</li>
934 <li>The version control system prevents people from overwriting each other's work by forcing them to merge concurrent changes before committing.</li>
935 <li>It also keeps a complete history of changes made to the master so that old versions can be recovered reliably.</li>
936 <li>Version control systems work best with text files, but can also handle binary files such as images and Word documents.</li>
937 <li>Every repository is identified by a URL.</li>
938 <li>Working copies of different repositories may not overlap.</li>
939 <li>Each changed to the master copy is identified by a unique revision number.</li>
940 <li>Revisions identify snapshots of the entire repository, not changes to individual files.</li>
941 <li>Each change should be commented to make the history more readable.</li>
942 <li>Commits are transactions: either all changes are successfully committed, or none are.</li>
943 <li>The basic workflow for version control is update-change-commit.</li>
944 <li><code>svn add <em>things</em></code> tells Subversion to start managing particular files or directories.</li>
945 <li><code>svn checkout <em>url</em></code> checks out a working copy of a repository.</li>
946 <li><code>svn commit -m "<em>message</em>" <em>things</em></code> sends changes to the repository.</li>
947 <li><code>svn diff</code> compares the current state of a working copy to the state after the most recent update.</li>
948 <li><code>svn diff -r HEAD</code> compares the current state of a working copy to the state of the master copy.</li>
949 <li><code>svn history</code> shows the history of a working copy.</li>
950 <li><code>svn status</code> shows the status of a working copy.</li>
951 <li><code>svn update</code> updates a working copy from the repository.</li>
955 <div class="challenges">
961 Using the repository URL, user ID, and password provided by the instructor,
962 perform the following actions:
965 Check out a working copy of the repository.
968 Create a text file called <em>your_id</em>.txt
969 (using your user ID instead of <em>your_id</em>)
970 and write a three-line biography of yourself in it.
973 Add this file to your working copy.
976 Commit your changes to the repository.
979 Update your working copy to get other people's biographies.
982 Examine the change log to see
983 the order in which people added their biographies
990 What does the command <code>svn diff -r 14</code> do?
991 What does it do if there have only been 10 changes to the repository?
996 Unix <code>diff</code> and <code>svn diff</code> compare files line by line.
997 Why doesn't this work for MP3 audio files?
1005 <section id="s:merge">
1006 <h2>Merging Conflicts</h2>
1008 <div class="understand">
1009 <h3>Learning Objectives</h3>
1011 <li>Explain what causes conflicts to occur and how to tell when one has occurred.</li>
1012 <li>Resolve a conflict.</li>
1013 <li>Identify the auxiliary files created when a conflict occurs.</li>
1018 Dracula and Wolfman have both synchronized their working copies of <code>explore</code>
1019 with version 8 of the repository.
1020 Dracula now edits his copy to change Amalthea's radius
1021 from a single number to a triple to reflect its irregular shape:
1024 <pre src="svn/moons_dracula_triple.txt">
1025 Name Orbital Radius Orbital Period Mass Radius
1026 <span class="highlight">Amalthea 181.4 0.498179 0.075 131 x 73 x 67</span>
1027 Io 421.6 1.769138 893.2 1821.6
1028 Europa 670.9 3.551181 480.0 1560.8
1029 Ganymede 1070.4 7.154553 1481.9 2631.2
1030 Callisto 1882.7 16.689018 1075.9 2410.3
1033 <p class="continue">
1034 He then commits his work,
1035 creating revision 9 of the repository
1036 (<a href="#f:after_dracula_commits">Figure 9</a>).
1039 <figure id="f:after_dracula_commits">
1040 <img src="svn/after_dracula_commits.png" alt="After Dracula Commits" />
1041 <figcaption>Figure 9: After Dracula Commits</figcaption>
1045 But while he is doing this,
1046 Wolfman is editing <em>his</em> copy
1047 to add information about two other minor moons,
1051 <pre src="svn/moons_wolfman_extras.txt">
1052 Name Orbital Radius Orbital Period Mass Radius
1053 Amalthea 181.4 0.498179 0.075 131
1054 Io 421.6 1.769138 893.2 1821.6
1055 Europa 670.9 3.551181 480.0 1560.8
1056 Ganymede 1070.4 7.154553 1481.9 2631.2
1057 Callisto 1882.7 16.689018 1075.9 2410.3
1058 <span class="highlight">Himalia 11460 250.5662 0.095 85.0
1059 Elara 11740 259.6528 0.008 40.0</span>
1063 When Wolfman tries to commit his changes to the repository,
1064 Subversion won't let him:
1068 $ <span class="in">svn commit -m "Added data for Himalia, Elara"</span>
1069 <span class="out">Sending jupiter/moons.txt
1070 svn: Commit failed (details follow):
1071 svn: File or directory 'moons.txt' is out of date; try updating
1072 svn: resource out of date; try updating</span>
1075 <p class="continue">
1077 Wolfman's changes were based on revision 8,
1078 but the repository is now at revision 9,
1079 and the file that Wolfman is trying to overwrite
1080 is different in the later revision.
1082 one of version control's main jobs is to make sure that
1083 people don't trample on each other's work.)
1084 Wolfman has to update his working copy to get Dracula's changes before he can commit.
1086 Dracula edited a line that Wolfman didn't change,
1087 so Subversion can merge the differences automatically.
1091 This does <em>not</em> mean that Wolfman's changes have been committed to the repository:
1092 Subversion only does that when it's ordered to.
1093 Wolfman's changes are still in his working copy,
1094 and <em>only</em> in his working copy.
1095 But since Wolfman's version of the file now includes
1096 the lines that Dracula added,
1097 Wolfman can go ahead and commit them as usual to create revision 10
1098 (<a href="#f:merge_without_conflict">Figure 10</a>).
1101 <figure id="f:merge_without_conflict">
1102 <img src="svn/merge_without_conflict.png" alt="Merging Without Conflict" />
1103 <figcaption>Figure 10: Merging Without Conflict</figcaption>
1107 Wolfman's working copy is now in sync with the master,
1108 but Dracula's is one behind at revision 9.
1110 they independently decide to add measurement units
1111 to the columns in <code>moons.txt</code>.
1112 Wolfman is quicker off the mark this time;
1113 he adds a line to the file:
1116 <pre src="svn/moons_wolfman_units.txt">
1117 Name Orbital Radius Orbital Period Mass Radius
1118 <span class="highlight"> (10**3 km) (days) (10**20 kg) (km)</span>
1119 Amalthea 181.4 0.498179 0.075 131 x 73 x 67
1120 Io 421.6 1.769138 893.2 1821.6
1121 Europa 670.9 3.551181 480.0 1560.8
1122 Ganymede 1070.4 7.154553 1481.9 2631.2
1123 Callisto 1882.7 16.689018 1075.9 2410.3
1124 Himalia 11460 250.5662 0.095 85.0
1125 Elara 11740 259.6528 0.008 40.0
1128 <p class="continue">
1129 and commits it to create revision 11.
1130 While he is doing this,
1132 Dracula inserts a different line at the top of the file:
1135 <pre src="svn/moons_dracula_units.txt">
1136 Name Orbital Radius Orbital Period Mass Radius
1137 <span class="highlight"> * 10^3 km * days * 10^20 kg * km</span>
1138 Amalthea 181.4 0.498179 0.075 131 x 73 x 67
1139 Io 421.6 1.769138 893.2 1821.6
1140 Europa 670.9 3.551181 480.0 1560.8
1141 Ganymede 1070.4 7.154553 1481.9 2631.2
1142 Callisto 1882.7 16.689018 1075.9 2410.3
1143 Himalia 11460 250.5662 0.095 85.0
1144 Elara 11740 259.6528 0.008 40.0
1149 when Dracula tries to commit,
1150 Subversion tells him he can't.
1152 when Dracula does updates his working copy,
1153 he doesn't just get the line Wolfman added to create revision 11
1154 (<a href="#f:merge_with_conflict">Figure 11</a>).
1157 <figure id="f:merge_with_conflict">
1158 <img src="svn/merge_with_conflict.png" alt="Merge With Conflict" />
1159 <figcaption>Figure 11: Merge With Conflict</figcaption>
1163 There is an actual conflict in the file,
1164 so Subversion asks Dracula what he wants to do:
1167 <pre src="svn/moons_dracula_conflict.txt">
1168 $ <span class="in">svn update</span>
1169 <span class="out">Conflict discovered in 'jupiter/moons.txt'.
1170 Select: (p) postpone, (df) diff-full, (e) edit,
1171 (mc) mine-conflict, (tc) theirs-conflict,
1172 (s) show all options:</span>
1176 Dracula choose <code>p</code> for "postpone",
1177 which tells Subversion that he'll deal with the problem later.
1178 Once the update is finished,
1179 he opens <code>moons.txt</code> in his editor and sees:
1183 Name Orbital Radius Orbital Period Mass
1184 +<<<<<<< .mine
1185 + * 10^3 km * days * 10^20 kg
1187 + (10**3 km) (days) (10**20 kg)
1188 +>>>>>>> .r11
1189 Amalthea 181.4 0.498179 0.074
1190 Io 421.6 1.769138 893.2
1191 Europa 670.9 3.551181 480.0
1192 Ganymede 1070.4 7.154553 1481.9
1193 Callisto 1882.7 16.689018 1075.9
1196 <p class="continue">
1198 Subversion has inserted
1199 <a href="glossary.html#conflict-marker">conflict markers</a>
1200 in <code>moons.txt</code>
1201 wherever there is a conflict.
1202 The line <code><<<<<<< .mine</code> shows the start of the conflict,
1203 and is followed by the lines from the local copy of the file.
1204 The separator <code>=======</code> is then
1205 followed by the lines from the repository's file that are in conflict with that section,
1206 while <code>>>>>>>> .r11</code> marks the end of the conflict.
1210 Before he can commit,
1211 Dracula has to edit his copy of the file to get rid of those markers.
1215 <pre src="svn/moons_dracula_resolved.txt">
1216 Name Orbital Radius Orbital Period Mass Radius
1217 <span class="highlight"> (10^3 km) (days) (10^20 kg) (km)</span>
1218 Amalthea 181.4 0.498179 0.075 131 x 73 x 67
1219 Io 421.6 1.769138 893.2 1821.6
1220 Europa 670.9 3.551181 480.0 1560.8
1221 Ganymede 1070.4 7.154553 1481.9 2631.2
1222 Callisto 1882.7 16.689018 1075.9 2410.3
1223 Himalia 11460 250.5662 0.095 85.0
1224 Elara 11740 259.6528 0.008 40.0
1227 <p class="continue">
1228 then uses the <code>svn resolved</code> command to tell Subversion that
1229 he has fixed the problem.
1230 Subversion will now let him commit to create revision 12.
1234 <h3>Auxiliary Files</h3>
1237 When Dracula did his update and Subversion detected the conflict in <code>moons.txt</code>,
1238 it created three temporary files to help Dracula resolve it.
1239 The first is called <code>moons.txt.r9</code>;
1240 it is the file as it was in Dracula's local copy
1241 before he started making changes,
1242 i.e., the common ancestor for his work
1243 and whatever he is in conflict with.
1247 The second file is <code>moons.txt.r11</code>.
1248 This is the most up-to-date revision from the repository—the
1249 file as it is including Wolfman's changes.
1250 The third temporary file, <code>moons.txt.mine</code>,
1251 is the file as it was in Dracula's working copy before he did the Subversion update.
1255 Subversion creates these auxiliary files primarily
1256 to help people merge conflicts in binary files.
1257 It wouldn't make sense to insert <code><<<<<<<</code>
1258 and <code>>>>>>>></code> characters into an image file
1259 (it would almost certainly result in a corrupted image).
1260 The <code>svn resolved</code> command deletes these three extra files
1261 as well as telling Subversion that the conflict has been taken care of.
1267 Some power users prefer to work with interpolated conflict markers directly,
1268 but for the rest of us,
1269 there are several tools for displaying differences and helping to merge them,
1270 including <a href="http://diffuse.sourceforge.net/">Diffuse</a> and <a href="http://winmerge.org/">WinMerge</a>.
1271 If Dracula launches Diffuse,
1272 it displays his file,
1273 the common base that he and Wolfman were working from,
1274 and Wolfman's file in a three-pane view
1275 (<a href="#f:diff_viewer">Figure 12</a>):
1278 <figure id="f:diff_viewer">
1279 <img src="svn/diff_viewer.png" alt="A Difference Viewer" />
1280 <figcaption>Figure 12: A Difference Viewer</figcaption>
1283 <p class="continue">
1284 Dracula can use the buttons to merge changes from either of the edited versions
1285 into the common ancestor,
1286 or edit the central pane directly.
1289 he uses <code>svn resolved</code> and <code>svn commit</code>
1290 to create revision 12 of the repository.
1294 In this case, the conflict was small and easy to fix.
1295 However, if two or more people on a team are repeatedly creating conflicts for one another,
1296 it's usually a signal of deeper communication problems:
1297 either they aren't talking as often as they should, or their responsibilities overlap.
1299 the version control system can help the team find and fix these issues
1300 so that it will be more productive in future.
1304 <h3>Working With Multiple Files</h3>
1307 As mentioned <a href="#a:transaction">earlier</a>,
1308 every logical change to a project should result in a single commit,
1309 and every commit should represent one logical change.
1310 This is especially true when resolving conflicts:
1311 the work done to reconcile one person's changes with another are often complicated,
1312 so it should be a single entry in the project's history,
1313 with other, later, changes coming after it.
1318 <div class="keypoints">
1321 <li>Conflicts must be resolved before a commit can be completed.</li>
1322 <li>Subversion puts markers in text files to show regions of conflict.</li>
1323 <li>For each conflicted file, Subversion creates auxiliary files containing the common parent, the master version, and the local version.</li>
1324 <li><code>svn resolve <em>files</em></code> tells Subversion that conflicts have been resolved.</li>
1328 <div class="challenges">
1332 If you are working in a group,
1333 partner with someone who has also wrote a biography for themselves
1334 for the previous section's challenges.
1339 Both partners use <code>svn update</code>
1340 to make sure their working copies are up to date
1341 and that there are no local changes.
1344 The first partner edits her biography and commits the changes.
1347 The second partner edits her copy of the file
1348 (<em>without</em> having updated to get the first partner's changes),
1349 then tries to <code>svn commit</code>.
1352 Once the second partner has resolved the conflict,
1353 she commits her changes.
1356 Repeat these four steps with roles reversed.
1361 If you are working on your own,
1362 you can simulate the steps above
1363 by checking out a second copy of the project into a new directory.
1365 this cannot overlap any existing checked-out copies.)
1366 Edit your biography in one copy and commit those changes,
1367 then switch to the other copy and edit the same file
1374 <section id="s:rollback">
1375 <h2>Recovering Old Versions</h2>
1377 <div class="understand">
1378 <h3>Learning Objectives</h3>
1380 <li>Discard changes made to a working copy.</li>
1381 <li>Recover an old version of a file.</li>
1382 <li>Explain what branches are and when they are used.</li>
1387 Now that we have seen how to merge files and resolve conflicts,
1388 we can look at how to use version control as an "infinite undo".
1389 Suppose that when Wolfman starts work late one night,
1390 his copy of <code>explore</code> is in sync with the head at revision 12.
1391 He decides to edit the file <code>moons.txt</code>;
1392 unfortunately, he forgot that there was a full moon,
1393 so his changes don't make a lot of sense:
1396 <pre src="svn/poetry.txt">
1397 Just one moon can make me growl
1398 Four would make me want to howl
1403 When he's back in human form the next day,
1404 he wants to undo his changes.
1405 Without version control, his choices would be grim:
1406 he could try to edit them back into their original state by hand
1407 (which for some reason hardly ever seems to work),
1408 or ask his colleagues to send him their copies of the files
1409 (which is almost as embarrassing as chasing the neighbor's cat when in wolf form).
1413 Since he's using Subversion, though,
1414 and hasn't committed his work to the repository,
1415 all he has to do is <a href="glossary.html#revert">revert</a> his local changes.
1416 <code>svn revert</code> simply throws away local changes to files
1417 and puts things back the way they were before those changes were made.
1418 This is a purely local operation:
1419 since Subversion stores the history of the project inside every working copy,
1420 Wolfman doesn't need to be connected to the network to do this.
1425 Wolfman uses <code>svn diff</code> <em>without</em> the <code>-r HEAD</code> flag
1426 to take a look at the differences between his file
1427 and the master copy in the repository.
1428 Since he doesn't want to keep his changes,
1429 his next command is <code>svn revert moons.txt</code>.
1433 $ <span class="in">cd jupiter</span>
1434 $ <span class="in">svn revert moons.txt</span>
1435 <span class="out">Reverted moons.txt</span>
1439 What if someone <em>has</em> committed their changes,
1440 but still wants to undo them?
1442 suppose Dracula decides that the numbers in <code>moons.txt</code> would look better with commas.
1443 He edits the file to put them in:
1446 <pre src="svn/moons_commas.txt">
1447 Name Orbital Radius Orbital Period Mass Radius
1448 (10^3 km) (days) (10^20 kg) (km)
1449 Amalthea 181.4 0.498179 0.075 131 x 73 x 67
1450 Io 421.6 1.769138 893.2 1<span class="highlight">,</span>821.6
1451 Europa 670.9 3.551181 480.0 1<span class="highlight">,</span>560.8
1452 Ganymede 1<span class="highlight">,</span>070.4 7.154553 1<span class="highlight">,</span>481.9 2<span class="highlight">,</span>631.2
1453 Callisto 1<span class="highlight">,</span>882.7 16.689018 1<span class="highlight">,</span>075.9 2<span class="highlight">,</span>410.3
1454 Himalia 11<span class="highlight">,</span>460 250.5662 0.095 85.0
1455 Elara 11<span class="highlight">,</span>740 259.6528 0.008 40.0
1458 <p class="continue">
1459 then commits his changes to create revision 13.
1460 A little while later,
1461 the Mummy sees the change and orders Dracula to put things back the way they were.
1462 What should Dracula do?
1466 We can draw the sequence of events leading up to revision 13
1467 as shown in <a href="#f:before_undoing">Figure 13</a>:
1470 <figure id="f:before_undoing">
1471 <img src="svn/before_undoing.png" alt="Before Undoing" />
1472 <figcaption>Figure 13: Before Undoing</figcaption>
1475 <p class="continue">
1476 Dracula wants to erase revision 13 from the repository,
1477 but he can't actually do that:
1478 once a change is in the repository,
1480 What he can do instead is merge the old revision with the current revision
1481 to create a new revision
1482 (<a href="#f:merging_history">Figure 14</a>).
1485 <figure id="f:merging_history">
1486 <img src="svn/merging_history.png" alt="Merging History" />
1487 <figcaption>Figure 14: Merging History</figcaption>
1490 <p class="continue">
1491 This is exactly like merging changes made by two different people;
1492 the only difference is that the "other person" is his past self.
1497 Dracula must merge revision 12 (the one before his change)
1498 with revision 13 (the current head revision)
1499 using <code>svn merge</code>:
1503 $ <span class="in">svn merge -r HEAD:12 moons.txt</span>
1504 <span class="out">-- Reverse-merging r13 into 'moons.txt'
1508 <p class="continue">
1509 The <code>-r</code> flag specifies the range of revisions to merge:
1510 to undo the changes from revision 12 to revision 13,
1511 he uses either <code>13:12</code> or <code>HEAD:12</code>
1512 (since he is going backward in time from the most recent revision to revision 12).
1513 This is called a <a href="glossary.html#reverse-merge">reverse</a> merge
1514 because he's going backward in time.
1518 After he runs this command,
1519 he must run <code>svn commit</code> to save the changes to the repository.
1520 This creates a new revision, number 14,
1521 rather than erasing revision 13.
1523 the changes he made to create revision 13 are still there
1524 if he can ever convince the Mummy that numbers should have commas.
1528 <h3>Another Way to Do It</h3>
1531 Another way to recover a particular version of a particular file
1532 is to use the <code>svn copy</code> command.
1533 If the URL of our repository is
1534 <code>https://universal.software-carpentry.org/explore</code>,
1539 $ <span class="in">svn copy https://universal.software-carpentry.org/explore/mission.txt@120 ./mission.txt</span>
1542 <p class="continue">
1543 copies the file <code>mission.txt</code> as it was in revision 120
1544 into our working directory
1545 (overwriting whatever <code>mission.txt</code> file we currently have,
1548 using <code>svn copy</code> brings along the file's history as well,
1549 so that future <code>svn log</code> operations will show
1550 how <code>mission.txt</code> was resurrected.
1555 Merging can be used to recover older revisions of files,
1556 not just the most recent,
1557 and to recover many files or directories at a time.
1558 The most frequent use, though,
1559 is to manage parallel streams of development in large projects.
1560 This is outside the scope of this chapter,
1561 but the basic idea is simple.
1565 Suppose that Universal Missions has just released a new program
1566 for designing interplanetary voyages.
1567 Dracula and Wolfman are supposed to add some features
1568 that were left out of the first release because time ran short.
1570 Frankenstein and the Mummy are doing technical support:
1571 their job is to fix any bugs that users find.
1575 All sorts of things could go wrong
1576 if both teams tried to work on the same code at the same time.
1578 Dracula and Wolfman might want to make large changes
1579 to the structure of the code
1580 in order to make it easier to add new features,
1581 while Frankenstein and the Mummy want to make as few changes as possible
1582 so as not to introduce new bugs while fixing old ones.
1586 The usual way to handle this situation is
1587 to create a <a href="glossary.html#branch">branch</a>
1588 in the repository for each major sub-project
1589 (<a href="#f:branch_merge">Figure 15</a>).
1590 While Wolfman and Dracula work on
1591 the <a href="glossary.html#main-line">main line</a>,
1592 Frankenstein and the Mummy create a branch,
1593 which is just another copy of the repository's files and directories
1594 that is also under version control.
1595 They can work in their branch without disturbing Wolfman and Dracula and vice versa:
1598 <figure id="f:branch_merge">
1599 <img src="svn/branch_merge.png" alt="Branching and Merging" />
1600 <figcaption>Figure 15: Branching and Merging</figcaption>
1604 Branches in version control repositories are often described as "parallel universes".
1605 Each branch starts off as a clone of the project at some moment in time
1606 (typically each time the software is released,
1607 or whenever work starts on a major new feature).
1608 Changes made to a branch only affect that branch,
1609 just as changes made to the files in one directory don't affect files in other directories.
1611 the branch and the main line are both stored in the same repository,
1612 so their revision numbers are always in step.
1616 If someone decides that a bug fix in one branch should also be made in another,
1617 all they have to do is merge the files in question.
1618 This is exactly like merging an old version of a file with the current one,
1619 but instead of going backward in time,
1620 the change is brought sideways from one branch to another.
1624 Branching helps projects scale up by letting sub-teams work independently,
1625 but too many branches can cause as many problems as they solve.
1626 Karl Fogel's excellent book
1627 <a href="bib.html#fogel-producing-oss"><cite>Producing Open Source Software</cite></a>,
1628 and Laura Wingerd and Christopher Seiwald's paper
1629 "<a href="bib.html#wingerd-seiwald-scm">High-level Best Practices in Software Configuration Management</a>",
1630 talk about branches in much more detail.
1631 Projects usually don't need to do this until they have a dozen or more developers,
1632 or until several versions of their software are in simultaneous use,
1633 but using branches is a key part of switching from software carpentry to software engineering.
1636 <div class="keypoints">
1639 <li>Old versions of files can be recovered by merging their old state with their current state.</li>
1640 <li>Recovering an old version of a file does not erase the intervening changes.</li>
1641 <li>Use branches to support parallel independent development.</li>
1642 <li><code>svn revert</code> undoes local changes to files.</li>
1643 <li><code>svn merge</code> merges two revisions of a file.</li>
1647 <div class="challenges">
1652 Explain what the command:
1654 svn diff -r 240:261 fish.dat
1656 does, and when you might want to run it.
1660 Suppose that a file called <code>mission.txt</code>
1661 existed in revision 90 of a repository,
1662 but had been deleted in revision 91.
1663 What two commands could we use to recover it?
1671 <section id="s:setup">
1672 <h2>Setting up a Repository</h2>
1674 <div class="understand">
1675 <h3>Learning Objectives</h3>
1677 <li>How to create a repository.</li>
1682 It is finally time to see how to create a repository.
1684 we will keep the master copy of our work in a repository
1685 on a server that we can access from other machines on the internet.
1686 That master copy consists of files and directories that no-one ever edits directly.
1687 Instead, a copy of Subversion running on that machine
1688 manages updates for us and watches for conflicts.
1689 Our working copy is a mirror image of the master sitting on our computer.
1690 When our Subversion client needs to communicate with the master,
1691 it exchanges data with the copy of Subversion running on the server.
1695 To make this to work, we need four things:
1701 The repository itself.
1702 It's not enough to create an empty directory and start filling it with files:
1703 Subversion needs to create a lot of other structure
1704 in order to keep track of old revisions, who made what changes, and so on.
1708 The full URL of the repository.
1709 This includes the URL of the server
1710 and the path to the repository on that machine.
1711 (The second part is needed because a single server can,
1713 host many repositories.)
1717 Permission to read or write the master copy.
1718 Many open source projects give the whole world permission to read from their repository,
1719 but very few allow strangers to write to it:
1720 there are just too many possibilities for abuse.
1721 Somehow, we have to set up a password or something like it
1722 so that users can prove who they are.
1726 A working copy of the repository on our computer.
1727 Once the first three things are in place,
1728 this just means running the <code>checkout</code> command.
1734 To keep things simple,
1735 we will start by creating a repository on the machine that we're working on.
1736 This won't let us share our work with other people,
1737 but it <em>will</em> allow us to save the history of our work as we go along.
1741 The command to create a repository is <code>svnadmin create</code>,
1742 followed by the path to the repository.
1743 If we want to create a repository called <code>missions_repo</code>
1744 directly under our home directory,
1745 we just <code>cd</code> to get home
1746 and run <code>svnadmin create missions_repo</code>.
1747 This command creates a directory called <code>missions_repo</code> to hold our repository,
1748 and fills it with various files that Subversion uses
1749 to keep track of the project's history:
1753 $ <span class="in">cd</span>
1754 $ <span class="in">svnadmin create missions_repo</span>
1755 $ <span class="in">ls -F missions_repo</span>
1756 <span class="out">README.txt conf/ db/ format hooks/ locks/</span>
1759 <p class="continue">
1760 We should <em>never</em> edit any of this directly,
1761 since it will almost certainly make the repository unusable.
1763 we should use <code>svn checkout</code>
1764 to get a working copy of this repository.
1765 If our home directory is <code>/users/mummy</code>,
1766 then the full path to the repository we just created is <code>/users/mummy/missions_repo</code>,
1767 so we run <code>svn checkout file:///users/mummy/missions missions_working</code>.
1772 the second argument,
1773 <code>missions_working</code>,
1774 specifies where the working copy is to be put.
1775 The first argument is the URL of our repository,
1776 and it has two parts.
1777 <code>/users/mummy/missions_repo</code> is the path to repository directory.
1778 <code>file://</code> specifies the <a href="glossary.html#protocol">protocol</a>
1779 that Subversion will use to communicate with the repository—in this case,
1780 it says that the repository is part of the local machine's filesystem.
1781 (Notice that the protocol ends in two slashes,
1782 while the absolute path to the repository starts with a slash,
1783 making three in total.
1784 A very common mistake is to type only two, since that's what web URLs normally have.)
1788 When we're doing a checkout,
1789 it is <em>very</em> important that we provide the second argument,
1790 which specifies the name of the directory we want the working copy to be put in.
1792 Subversion will try to use the name of the repository,
1793 <code>missions_repo</code>,
1794 as the name of the working copy.
1795 Since we're in the directory that contains the repository,
1796 this means that Subversion will try to overwrite the repository with a working copy.
1798 there isn't much risk of our sanity being torn to shreds,
1799 but this could ruin our repository.
1803 To avoid this problem,
1804 most people create a sub-directory in their account called something like <code>repos</code>,
1805 and then create their repositories in that.
1807 we could create our repository in <code>/users/mummy/repos/missions</code>,
1808 then check out a working copy as <code>/users/mummy/missions</code>.
1809 This practice makes both names easier to read.
1813 The obvious next step is to put our repository on a server,
1814 rather than on our personal machine.
1816 we should <em>always</em> do this
1817 so that we don't lose the history of our project
1818 if our laptop is damaged or stolen.
1819 A departmental server is also much more likely to be backed up regularly
1820 than our personal machine…
1824 Creating a repository on a server is simple:
1825 just log in and go through the steps described above.
1826 Accessing that repository from another machine
1827 is also straightforward.
1828 If the machine's address is <code>serv.euphoric.edu</code>,
1829 and our user ID is <code>dracula</code>,
1830 the URL of the repository will be something like:
1834 svn+ssh://dracula@serv.euphoric.edu/home/dracula/repos/missions
1838 Reading from left to right:
1843 <code>svn+ssh</code> is the protocol that Subversion uses to connect to the server
1845 a combination of Subversion's own protocol
1846 and <a href="shell.html#s:ssh">SSH</a>);
1849 <code>dracula@serv.euphoric.edu</code> identifies the server and who we are
1850 (just like an email address);
1854 <code>/home/dracula/repos/missions</code> is the absolutely path of the repository
1859 <p id="a:only_user">
1860 That's fine if you are the only person using the repository,
1861 but if you want to share it with others,
1862 you need to worry about security.
1863 As we discuss in the lesson on <a href="web.html">web programming</a>,
1864 as soon as you provide a service on the internet,
1865 there's the possibility that someone may try to attack your system through it.
1866 Rather than trying to learn enough system administration skills
1867 to set things up safely,
1868 it is usually easier to:
1874 ask your department's system administrator to set it up for you;
1878 use a hosting service like <a href="http://www.sf.net">SourceForge</a>,
1879 <a href="http://code.google.com">Google Code</a>,
1880 <a href="https://github.com/">GitHub</a>,
1881 or <a href="https://bitbucket.org/">BitBucket</a>; or
1885 spend a few dollars a month on a commercial hosting service
1886 that provides web-based GUIs for creating and managing repositories.
1892 If you choose the second or third option,
1893 please check with whoever handles intellectual property at your institution
1894 to make sure that putting your work on a commercially-operated machine
1895 that is probably in some other legal jurisdiction
1896 isn't going to cause trouble.
1897 Many people assume that it's "just OK",
1898 while others act as if not having asked will be an acceptable defence later on.
1900 neither is true…
1903 <div class="keypoints">
1906 <li><code>svnadmin create <em>name</em></code> creates a new repository.</li>
1907 <li>Repositories can be hosted locally, on local (departmental) servers, on hosting services, or on their owners' own domains.</li>
1911 <div class="challenges">
1917 Create a Subversion repository called <code>trials_repo</code>
1918 in your home directory.
1919 Check out a working copy in a directory called <code>trials_working</code>
1920 (also in your home directory).
1921 Add a couple of text files,
1923 and then use <code>svn info trials_working</code>
1924 to see what Subversion tells you about your working copy.
1928 We said <a href="#a:only_user">above</a> that
1929 you might be the only person using a particular repository.
1930 When and why is version control worth using
1931 if no-one else is working on a project with you?
1935 There are many ways to organize repositories.
1936 Some of the most common are to create one repository for:
1938 <li>each person</li>
1940 <li>all the work done on one grant</li>
1941 <li>all the work done on one project</li>
1942 <li>the entire lab (which is shared by everyone in the lab)</li>
1943 <li>the entire department (typically with a top-level directory for each person or project in the department)</li>
1945 What activities does each one make easy or hard?
1946 Which of these would you prefer, and why?
1954 <section id="s:provenance">
1957 <div class="understand">
1958 <h3>Understand:</h3>
1960 <li>What data provenance is.</li>
1961 <li>How to embed version numbers and other information in files managed by version control.</li>
1962 <li>How to record version information about a program in its output.</li>
1968 the <a href="glossary.html#provenance">provenance</a> of a work
1969 is the history of who owned it, when, and where.
1971 it's the record of how a particular result came to be:
1972 what raw data was processed by what version of what program to create which intermediate files,
1973 what was used to turn those files into which figures of which papers,
1978 One of the big benefits of using version control is that
1979 it lets us track the provenance of scientific data automatically.
1981 suppose we have a text file <code>combustion.dat</code> in a Subversion repository.
1982 Run the following two commands:
1986 $ svn propset svn:keywords Revision combustion.dat
1987 $ svn commit -m "Turning on the 'Revision' keyword" combustion.dat
1990 <p class="continue">
1991 This does nothing by itself,
1992 but now open the file in an editor
1993 and add the following line somewhere near the top:
2001 The <code>$Revision:$</code> string means something special to Subversion.
2002 Save the file, and commit the change:
2006 $ svn commit -m "Inserting the 'Revision' keyword" combustion.dat
2010 When we open the file again,
2011 we'll see that Subversion has changed that line to something like:
2018 <p class="continue">
2019 i.e., it has inserted the version number
2020 after the colon and before the closing <code>$</code>.
2021 If we edit the file again—e.g., add a couple of lines with random numbers—and
2023 the line is updated again to:
2031 Here's what just happened.
2032 First, Subversion allows uss to add
2033 <a href="glossary.html#property-subversion">properties</a>
2034 to files and and directories.
2035 These properties aren't stored in the files or directories themselves,
2036 but in Subversion's database.
2037 One of those properties,
2038 <code>svn:keywords</code>,
2039 tells Subversion to look in files that are being changed
2040 for strings of the form <code>$propertyname: …$</code>,
2041 where <code>propertyname</code> is a string like <code>Revision</code> or <code>Author</code>.
2042 (About half a dozen such strings are supported.)
2046 If it sees such a string,
2047 Subversion rewrites it as the commit is taking place to replace <code>…</code>
2048 with the current version number,
2049 the name of the person making the change,
2050 or whatever else the property's name tells it to do.
2051 We only have to add the string to the file once;
2053 Subversion updates it for you every time the file changes.
2057 Putting the version number in the file this way can be pretty handy.
2058 If you copy the file to another machine,
2060 it carries its version number with it,
2061 so you can tell which version you have even if it's outside version control.
2062 We'll see some more useful things we can do with this information <a href="python.html">later</a>.
2066 We can use this trick with shell scripts too,
2067 or with almost any other kind of program.
2068 Let's go back to Nelle Nemo's data processing from
2069 the lesson on the <a href="shell.html">shell</a>.
2070 Suppose she writes a shell script called <code>gooclean</code>
2071 to tidy up data files.
2072 Her first version looks like this:
2076 # gooclean: clean up a single data file
2077 goonorm -b 0 100 < $1 | goofilter -x --enlarge 2.0 > cleaned-$1
2080 <p class="continue">
2082 it runs <code>goonorm</code> and then <code>goofilter</code> with some fixed parameters
2083 and creates an output file called <code>cleaned-something.dat</code>
2084 (if the input file's name was <code>something.dat</code>).
2085 Assuming that '#' is the comment character for her output files,
2086 she could instead write:
2090 # gooclean: clean up a single data file
2091 <span class="highlight">echo "# gooclean $Revision:$" > cleaned-$1</span>
2092 goonorm -b 0 100 < $1 | goofilter -x --enlarge 2.0 <span class="highlight">>></span> cleaned-$1
2095 <p class="continue">
2096 then set the <code>svn:keywords</code> property
2097 and commit the file to insert the revision number,
2102 # gooclean: clean up a single data file
2103 <span class="highlight">echo "# gooclean $Revision: 487$" > cleaned-$1</span>
2104 goonorm -b 0 100 < $1 | goofilter -x --enlarge 2.0 <span class="highlight">>></span> cleaned-$1
2109 each time this script is run it will:
2116 # gooclean $Revision: 487$
2122 append whatever the pipline containing <code>goonorm</code> and <code>goofilter</code>
2123 would have put in the file originally.
2124 (The double redirection <code>>></code> means "append to" rather than "overwrite".)
2128 <p class="continue">
2130 the output of this shell script will always record
2131 exactly what version of the script produced it.
2132 This isn't enough to reproduce the output—we would need to record
2133 the version numbers of the input files and the <code>goonorm</code> and <code>goofilter</code> programs,
2134 and the values of the parameters those programs used
2135 in order to do that—but it's an important and useful first step.
2138 <div class="keypoints">
2141 <li><code>$Keyword: …$</code> in a file can be filled in with a property value each time the file is committed.</li>
2142 <li>Put version numbers in programs' output to establish provenance for data.</li>
2143 <li><code>svn propset svn:keywords <em>property</em> <em>files</em></code> tells Subversion to start filling in property values.</li>
2147 <div class="challenges">
2153 Add <code>$Id:$</code> to a file,
2154 use <code>svn propset</code> to set the corresponding property,
2155 and then commit a change to the file.
2156 What value does Subversion fill in for this keyword?
2157 When would you use this rather than <code>Revision</code> or <code>Author</code>?
2161 What does the <code>svn:ignore</code> property do when applied to a directory?
2162 When would you use it?
2171 <section id="s:summary">
2176 <a href="bib.html#mccullough-reproducibility">McCullough, McGeary, and Harrison</a>
2177 analyzed several years of
2178 the data and code archive of <cite>Journal of Money, Credit, and Banking</cite>,
2179 a prestigious journal with a mandatory archiving policy.
2180 Of 266 articles published during that time,
2181 193 were empirical and should have had data and code deposited in the archive.
2183 only 69 actually had anything in the archive;
2184 Excluding eleven articles that only had data,
2185 and seven that required software or other resources they did not have,
2186 McCullough et al. were only able to replicate 14 of the remaining 186 articles.
2187 This doesn't mean that the other 92% were wrong,
2188 but it does mean there is no practical way to tell.
2193 version control doesn't making computational research reproducible.
2194 It <em>does</em> help,
2196 and also eliminates the frustration and wasted time caused by
2197 trying to figure out which emailed copy of a file,
2198 or which of a dozen directories or USB drives,
2200 And while correlation doesn't imply causality,
2201 there is certainly a strong correlation between
2202 knowing enough about good computational practices to use version control
2203 and knowing how to do other things right as well.
2207 {% endblock content %}