hgbook
view en/concepts.tex @ 115:b74102b56df5
Wow! Lots more work detailing the working directory, merging, etc.
| author | Bryan O'Sullivan <bos@serpentine.com> |
|---|---|
| date | Mon Nov 13 16:19:48 2006 -0800 (2006-11-13) |
| parents | a0f57b3e677e |
| children | ca99f247899e |
line source
1 \chapter{Behind the scenes}
2 \label{chap:concepts}
4 Unlike many revision control systems, the concepts upon which
5 Mercurial is built are simple enough that it's easy to understand how
6 the software really works. Knowing this certainly isn't necessary,
7 but I find it useful to have a ``mental model'' of what's going on.
9 This understanding gives me confidence that Mercurial has been
10 carefully designed to be both \emph{safe} and \emph{efficient}. And
11 just as importantly, if it's easy for me to retain a good idea of what
12 the software is doing when I perform a revision control task, I'm less
13 likely to be surprised by its behaviour.
15 In this chapter, we'll initially cover the core concepts behind
16 Mercurial's design, then continue to discuss some of the interesting
17 details of its implementation.
19 \section{Mercurial's historical record}
21 \subsection{Tracking the history of a single file}
23 When Mercurial tracks modifications to a file, it stores the history
24 of that file in a metadata object called a \emph{filelog}. Each entry
25 in the filelog contains enough information to reconstruct one revision
26 of the file that is being tracked. Filelogs are stored as files in
27 the \sdirname{.hg/data} directory. A filelog contains two kinds of
28 information: revision data, and an index to help Mercurial to find a
29 revision efficiently.
31 A file that is large, or has a lot of history, has its filelog stored
32 in separate data (``\texttt{.d}'' suffix) and index (``\texttt{.i}''
33 suffix) files. For small files without much history, the revision
34 data and index are combined in a single ``\texttt{.i}'' file. The
35 correspondence between a file in the working directory and the filelog
36 that tracks its history in the repository is illustrated in
37 figure~\ref{fig:concepts:filelog}.
39 \begin{figure}[ht]
40 \centering
41 \grafix{filelog}
42 \caption{Relationships between files in working directory and
43 filelogs in repository}
44 \label{fig:concepts:filelog}
45 \end{figure}
47 \subsection{Managing tracked files}
49 Mercurial uses a structure called a \emph{manifest} to collect
50 together information about the files that it tracks. Each entry in
51 the manifest contains information about the files present in a single
52 changeset. An entry records which files are present in the changeset,
53 the revision of each file, and a few other pieces of file metadata.
55 \subsection{Recording changeset information}
57 The \emph{changelog} contains information about each changeset. Each
58 revision records who committed a change, the changeset comment, other
59 pieces of changeset-related information, and the revision of the
60 manifest to use.
62 \subsection{Relationships between revisions}
64 Within a changelog, a manifest, or a filelog, each revision stores a
65 pointer to its immediate parent (or to its two parents, if it's a
66 merge revision). As I mentioned above, there are also relationships
67 between revisions \emph{across} these structures, and they are
68 hierarchical in nature.
70 For every changeset in a repository, there is exactly one revision
71 stored in the changelog. Each revision of the changelog contains a
72 pointer to a single revision of the manifest. A revision of the
73 manifest stores a pointer to a single revision of each filelog tracked
74 when that changeset was created. These relationships are illustrated
75 in figure~\ref{fig:concepts:metadata}.
77 \begin{figure}[ht]
78 \centering
79 \grafix{metadata}
80 \caption{Metadata relationships}
81 \label{fig:concepts:metadata}
82 \end{figure}
84 As the illustration shows, there is \emph{not} a ``one to one''
85 relationship between revisions in the changelog, manifest, or filelog.
86 If the manifest hasn't changed between two changesets, the changelog
87 entries for those changesets will point to the same revision of the
88 manifest. If a file that Mercurial tracks hasn't changed between two
89 changesets, the entry for that file in the two revisions of the
90 manifest will point to the same revision of its filelog.
92 \section{Safe, efficient storage}
94 The underpinnings of changelogs, manifests, and filelogs are provided
95 by a single structure called the \emph{revlog}.
97 \subsection{Efficient storage}
99 The revlog provides efficient storage of revisions using a
100 \emph{delta} mechanism. Instead of storing a complete copy of a file
101 for each revision, it stores the changes needed to transform an older
102 revision into the new revision. For many kinds of file data, these
103 deltas are typically a fraction of a percent of the size of a full
104 copy of a file.
106 Some obsolete revision control systems can only work with deltas of
107 text files. They must either store binary files as complete snapshots
108 or encoded into a text representation, both of which are wasteful
109 approaches. Mercurial can efficiently handle deltas of files with
110 arbitrary binary contents; it doesn't need to treat text as special.
112 \subsection{Safe operation}
114 Mercurial only ever \emph{appends} data to the end of a revlog file.
115 It never modifies a section of a file after it has written it. This
116 is both more robust and efficient than schemes that need to modify or
117 rewrite data.
119 In addition, Mercurial treats every write as part of a
120 \emph{transaction} that can span a number of files. A transaction is
121 \emph{atomic}: either the entire transaction succeeds and its effects
122 are all visible to readers in one go, or the whole thing is undone.
123 This guarantee of atomicity means that if you're running two copies of
124 Mercurial, where one is reading data and one is writing it, the reader
125 will never see a partially written result that might confuse it.
127 The fact that Mercurial only appends to files makes it easier to
128 provide this transactional guarantee. The easier it is to do stuff
129 like this, the more confident you should be that it's done correctly.
131 \subsection{Fast retrieval}
133 Mercurial cleverly avoids a pitfall common to all earlier
134 revision control systems: the problem of \emph{inefficient retrieval}.
135 Most revision control systems store the contents of a revision as an
136 incremental series of modifications against a ``snapshot''. To
137 reconstruct a specific revision, you must first read the snapshot, and
138 then every one of the revisions between the snapshot and your target
139 revision. The more history that a file accumulates, the more
140 revisions you must read, hence the longer it takes to reconstruct a
141 particular revision.
143 \begin{figure}[ht]
144 \centering
145 \grafix{snapshot}
146 \caption{Snapshot of a revlog, with incremental deltas}
147 \label{fig:concepts:snapshot}
148 \end{figure}
150 The innovation that Mercurial applies to this problem is simple but
151 effective. Once the cumulative amount of delta information stored
152 since the last snapshot exceeds a fixed threshold, it stores a new
153 snapshot (compressed, of course), instead of another delta. This
154 makes it possible to reconstruct \emph{any} revision of a file
155 quickly. This approach works so well that it has since been copied by
156 several other revision control systems.
158 Figure~\ref{fig:concepts:snapshot} illustrates the idea. In an entry
159 in a revlog's index file, Mercurial stores the range of entries from
160 the data file that it must read to reconstruct a particular revision.
162 \subsubsection{Aside: the influence of video compression}
164 If you're familiar with video compression or have ever watched a TV
165 feed through a digital cable or satellite service, you may know that
166 most video compression schemes store each frame of video as a delta
167 against its predecessor frame. In addition, these schemes use
168 ``lossy'' compression techniques to increase the compression ratio, so
169 visual errors accumulate over the course of a number of inter-frame
170 deltas.
172 Because it's possible for a video stream to ``drop out'' occasionally
173 due to signal glitches, and to limit the accumulation of artefacts
174 introduced by the lossy compression process, video encoders
175 periodically insert a complete frame (called a ``key frame'') into the
176 video stream; the next delta is generated against that frame. This
177 means that if the video signal gets interrupted, it will resume once
178 the next key frame is received. Also, the accumulation of encoding
179 errors restarts anew with each key frame.
181 \subsection{Identification and strong integrity}
183 Along with delta or snapshot information, a revlog entry contains a
184 cryptographic hash of the data that it represents. This makes it
185 difficult to forge the contents of a revision, and easy to detect
186 accidental corruption.
188 Hashes provide more than a mere check against corruption; they are
189 used as the identifiers for revisions. The changeset identification
190 hashes that you see as an end user are from revisions of the
191 changelog. Although filelogs and the manifest also use hashes,
192 Mercurial only uses these behind the scenes.
194 Mercurial verifies that hashes are correct when it retrieves file
195 revisions and when it pulls changes from another repository. If it
196 encounters an integrity problem, it will complain and stop whatever
197 it's doing.
199 In addition to the effect it has on retrieval efficiency, Mercurial's
200 use of periodic snapshots makes it more robust against partial data
201 corruption. If a revlog becomes partly corrupted due to a hardware
202 error or system bug, it's often possible to reconstruct some or most
203 revisions from the uncorrupted sections of the revlog, both before and
204 after the corrupted section. This would not be possible with a
205 delta-only storage model.
207 \section{Revision history, branching,
208 and merging}
210 Every entry in a Mercurial revlog knows the identity of its immediate
211 ancestor revision, usually referred to as its \emph{parent}. In fact,
212 a revision contains room for not one parent, but two. Mercurial uses
213 a special hash, called the ``null ID'', to represent the idea ``there
214 is no parent here''. This hash is simply a string of zeroes.
216 In figure~\ref{fig:concepts:revlog}, you can see an example of the
217 conceptual structure of a revlog. Filelogs, manifests, and changelogs
218 all have this same structure; they differ only in the kind of data
219 stored in each delta or snapshot.
221 The first revision in a revlog (at the bottom of the image) has the
222 null ID in both of its parent slots. For a ``normal'' revision, its
223 first parent slot contains the ID of its parent revision, and its
224 second contains the null ID, indicating that the revision has only one
225 real parent. Any two revisions that have the same parent ID are
226 branches. A revision that represents a merge between branches has two
227 normal revision IDs in its parent slots.
229 \begin{figure}[ht]
230 \centering
231 \grafix{revlog}
232 \caption{}
233 \label{fig:concepts:revlog}
234 \end{figure}
236 \section{The working directory}
238 In the working directory, Mercurial stores a snapshot of the files
239 from the repository as of a particular changeset.
241 The working directory ``knows'' which changeset it contains. When you
242 update the working directory to contain a particular changeset,
243 Mercurial looks up the appropriate revision of the manifest to find
244 out which files it was tracking at the time that changeset was
245 committed, and which revision of each file was then current. It then
246 recreates a copy of each of those files, with the same contents it had
247 when the changeset was committed.
249 The \emph{dirstate} contains Mercurial's knowledge of the working
250 directory. This details which changeset the working directory is
251 updated to, and all of the files that Mercurial is tracking in the
252 working directory.
254 Just as a revision of a revlog has room for two parents, so that it
255 can represent either a normal revision (with one parent) or a merge of
256 two earlier revisions, the dirstate has slots for two parents. When
257 you use the \hgcmd{update} command, the changeset that you update to
258 is stored in the ``first parent'' slot, and the null ID in the second.
259 When you \hgcmd{merge} with another changeset, the first parent
260 remains unchanged, and the second parent is filled in with the
261 changeset you're merging with. The \hgcmd{parents} command tells you
262 what the parents of the dirstate are.
264 \subsection{What happens when you commit}
266 The dirstate stores parent information for more than just book-keeping
267 purposes. Mercurial uses the parents of the dirstate as \emph{the
268 parents of a new changeset} when you perform a commit.
270 \begin{figure}[ht]
271 \centering
272 \grafix{wdir}
273 \caption{The working directory can have two parents}
274 \label{fig:concepts:wdir}
275 \end{figure}
277 Figure~\ref{fig:concepts:wdir} shows the normal state of the working
278 directory, where it has a single changeset as parent. That changeset
279 is the \emph{tip}, the newest changeset in the repository that has no
280 children.
282 \begin{figure}[ht]
283 \centering
284 \grafix{wdir-after-commit}
285 \caption{The working directory gains new parents after a commit}
286 \label{fig:concepts:wdir-after-commit}
287 \end{figure}
289 It's useful to think of the working directory as ``the changeset I'm
290 about to commit''. Any files that you tell Mercurial that you've
291 added, removed, renamed, or copied will be reflected in that
292 changeset, as will modifications to any files that Mercurial is
293 already tracking; the new changeset will have the parents of the
294 working directory as its parents.
296 After a commit, Mercurial will update the parents of the working
297 directory, so that the first parent is the ID of the new changeset,
298 and the second is the null ID. This is shown in
299 figure~\ref{fig:concepts:wdir-after-commit}. Mercurial doesn't touch
300 any of the files in the working directory when you commit; it just
301 modifies the dirstate to note its new parents.
303 \subsection{Creating a new head}
305 It's perfectly normal to update the working directory to a changeset
306 other than the current tip. For example, you might want to know what
307 your project looked like last Tuesday, or you could be looking through
308 changesets to see which one introduced a bug. In cases like this, the
309 natural thing to do is update the working directory to the changeset
310 you're interested in, and then examine the files in the working
311 directory directly to see their contents as they werea when you
312 committed that changeset. The effect of this is shown in
313 figure~\ref{fig:concepts:wdir-pre-branch}.
315 \begin{figure}[ht]
316 \centering
317 \grafix{wdir-pre-branch}
318 \caption{The working directory, updated to an older changeset}
319 \label{fig:concepts:wdir-pre-branch}
320 \end{figure}
322 Having updated the working directory to an older changeset, what
323 happens if you make some changes, and then commit? Mercurial behaves
324 in the same way as I outlined above. The parents of the working
325 directory become the parents of the new changeset. This new changeset
326 has no children, so it becomes the new tip. And the repository now
327 contains two changesets that have no children; we call these
328 \emph{heads}. You can see the structure that this creates in
329 figure~\ref{fig:concepts:wdir-branch}.
331 \begin{figure}[ht]
332 \centering
333 \grafix{wdir-branch}
334 \caption{After a commit made while synced to an older changeset}
335 \label{fig:concepts:wdir-branch}
336 \end{figure}
338 \begin{note}
339 If you're new to Mercurial, you should keep in mind a common
340 ``error'', which is to use the \hgcmd{pull} command without any
341 options. By default, the \hgcmd{pull} command \emph{does not}
342 update the working directory, so you'll bring new changesets into
343 your repository, but the working directory will stay synced at the
344 same changeset as before the pull. If you make some changes and
345 commit afterwards, you'll thus create a new head, because your
346 working directory isn't synced to whatever the current tip is.
348 I put the word ``error'' in quotes because all that you need to do
349 to rectify this situation is \hgcmd{merge}, then \hgcmd{commit}. In
350 other words, this almost never has negative consequences; it just
351 surprises people. I'll discuss other ways to avoid this behaviour,
352 and why Mercurial behaves in this initially surprising way, later
353 on.
354 \end{note}
356 \subsection{Merging heads}
358 When you run the \hgcmd{merge} command, Mercurial leaves the first
359 parent of the working directory unchanged, and sets the second parent
360 to the changeset you're merging with, as shown in
361 figure~\ref{fig:concepts:wdir-merge}.
363 \begin{figure}[ht]
364 \centering
365 \grafix{wdir-merge}
366 \caption{Merging two hehads}
367 \label{fig:concepts:wdir-merge}
368 \end{figure}
370 Mercurial also has to modify the working directory, to merge the files
371 managed in the two changesets. Simplified a little, the merging
372 process goes like this, for every file in the manifests of both
373 changesets.
374 \begin{itemize}
375 \item If neither changeset has modified a file, do nothing with that
376 file.
377 \item If one changeset has modified a file, and the other hasn't,
378 create the modified copy of the file in the working directory.
379 \item If one changeset has removed a file, and the other hasn't (or
380 has also deleted it), delete the file from the working directory.
381 \item If one changeset has removed a file, but the other has modified
382 the file, ask the user what to do: keep the modified file, or remove
383 it?
384 \item If both changesets have modified a file, invoke an external
385 merge program to choose the new contents for the merged file. This
386 may require input from the user.
387 \item If one changeset has modified a file, and the other has renamed
388 or copied the file, make sure that the changes follow the new name
389 of the file.
390 \end{itemize}
391 There are more details---merging has plenty of corner cases---but
392 these are the most common choices that are involved in a merge. As
393 you can see, most cases are completely automatic, and indeed most
394 merges finish automatically, without requiring your input to resolve
395 any conflicts.
397 When you're thinking about what happens when you commit after a merge,
398 once again the working directory is ``the changeset I'm about to
399 commit''. After the \hgcmd{merge} command completes, the working
400 directory has two parents; these will become the parents of the new
401 changeset.
403 Mercurial lets you perform multiple merges, but you must commit the
404 results of each individual merge as you go. This is necessary because
405 Mercurial only tracks two parents for both revisions and the working
406 directory. While it would be technically possible to merge multiple
407 changesets at once, the prospect of user confusion and making a
408 terrible mess of a merge immediately becomes overwhelming.
410 \section{Other interesting design features}
412 In the sections above, I've tried to highlight some of the most
413 important aspects of Mercurial's design, to illustrate that it pays
414 careful attention to reliability and performance. However, the
415 attention to detail doesn't stop there. There are a number of other
416 aspects of Mercurial's construction that I personally find
417 interesting. I'll detail a few of them here, separate from the ``big
418 ticket'' items above, so that if you're interested, you can gain a
419 better idea of the amount of thinking that goes into a well-designed
420 system.
422 \subsection{Clever compression}
424 When appropriate, Mercurial will store both snapshots and deltas in
425 compressed form. It does this by always \emph{trying to} compress a
426 snapshot or delta, but only storing the compressed version if it's
427 smaller than the uncompressed version.
429 This means that Mercurial does ``the right thing'' when storing a file
430 whose native form is compressed, such as a \texttt{zip} archive or a
431 JPEG image. When these types of files are compressed a second time,
432 the resulting file is usually bigger than the once-compressed form,
433 and so Mercurial will store the plain \texttt{zip} or JPEG.
435 Deltas between revisions of a compressed file are usually larger than
436 snapshots of the file, and Mercurial again does ``the right thing'' in
437 these cases. It finds that such a delta exceeds the threshold at
438 which it should store a complete snapshot of the file, so it stores
439 the snapshot, again saving space compared to a naive delta-only
440 approach.
442 \subsubsection{Network recompression}
444 When storing revisions on disk, Mercurial uses the ``deflate''
445 compression algorithm (the same one used by the popular \texttt{zip}
446 archive format), which balances good speed with a respectable
447 compression ratio. However, when transmitting revision data over a
448 network connection, Mercurial uncompresses the compressed revision
449 data.
451 If the connection is over HTTP, Mercurial recompresses the entire
452 stream of data using a compression algorithm that gives a etter
453 compression ratio (the Burrows-Wheeler algorithm from the widely used
454 \texttt{bzip2} compression package). This combination of algorithm
455 and compression of the entire stream (instead of a revision at a time)
456 substantially reduces the number of bytes to be transferred, yielding
457 better network performance over almost all kinds of network.
459 (If the connection is over \command{ssh}, Mercurial \emph{doesn't}
460 recompress the stream, because \command{ssh} can already do this
461 itself.)
463 \subsection{Read/write ordering and atomicity}
465 Appending to files isn't the whole story when it comes to guaranteeing
466 that a reader won't see a partial write. If you recall
467 figure~\ref{fig:concepts:metadata}, revisions in the changelog point to
468 revisions in the manifest, and revisions in the manifest point to
469 revisions in filelogs. This hierarchy is deliberate.
471 A writer starts a transaction by writing filelog and manifest data,
472 and doesn't write any changelog data until those are finished. A
473 reader starts by reading changelog data, then manifest data, followed
474 by filelog data.
476 Since the writer has always finished writing filelog and manifest data
477 before it writes to the changelog, a reader will never read a pointer
478 to a partially written manifest revision from the changelog, and it will
479 never read a pointer to a partially written filelog revision from the
480 manifest.
482 \subsection{Concurrent access}
484 The read/write ordering and atomicity guarantees mean that Mercurial
485 never needs to \emph{lock} a repository when it's reading data, even
486 if the repository is being written to while the read is occurring.
487 This has a big effect on scalability; you can have an arbitrary number
488 of Mercurial processes safely reading data from a repository safely
489 all at once, no matter whether it's being written to or not.
491 The lockless nature of reading means that if you're sharing a
492 repository on a multi-user system, you don't need to grant other local
493 users permission to \emph{write} to your repository in order for them
494 to be able to clone it or pull changes from it; they only need
495 \emph{read} permission. (This is \emph{not} a common feature among
496 revision control systems, so don't take it for granted! Most require
497 readers to be able to lock a repository to access it safely, and this
498 requires write permission on at least one directory, which of course
499 makes for all kinds of nasty and annoying security and administrative
500 problems.)
502 Mercurial uses locks to ensure that only one process can write to a
503 repository at a time (the locking mechanism is safe even over
504 filesystems that are notoriously hostile to locking, such as NFS). If
505 a repository is locked, a writer will wait for a while to retry if the
506 repository becomes unlocked, but if the repository remains locked for
507 too long, the process attempting to write will time out after a while.
508 This means that your daily automated scripts won't get stuck forever
509 and pile up if a system crashes unnoticed, for example. (Yes, the
510 timeout is configurable, from zero to infinity.)
512 \subsubsection{Safe dirstate access}
514 As with revision data, Mercurial doesn't take a lock to read the
515 dirstate file; it does acquire a lock to write it. To avoid the
516 possibility of reading a partially written copy of the dirstate file,
517 Mercurial writes to a file with a unique name in the same directory as
518 the dirstate file, then renames the temporary file atomically to
519 \filename{dirstate}. The file named \filename{dirstate} is thus
520 guaranteed to be complete, not partially written.
522 \subsection{Avoiding seeks}
524 Critical to Mercurial's performance is the avoidance of seeks of the
525 disk head, since any seek is far more expensive than even a
526 comparatively large read operation.
528 This is why, for example, the dirstate is stored in a single file. If
529 there were a dirstate file per directory that Mercurial tracked, the
530 disk would seek once per directory. Instead, Mercurial reads the
531 entire single dirstate file in one step.
533 Mercurial also uses a ``copy on write'' scheme when cloning a
534 repository on local storage. Instead of copying every revlog file
535 from the old repository into the new repository, it makes a ``hard
536 link'', which is a shorthand way to say ``these two names point to the
537 same file''. When Mercurial is about to write to one of a revlog's
538 files, it checks to see if the number of names pointing at the file is
539 greater than one. If it is, more than one repository is using the
540 file, so Mercurial makes a new copy of the file that is private to
541 this repository.
543 A few revision control developers have pointed out that this idea of
544 making a complete private copy of a file is not very efficient in its
545 use of storage. While this is true, storage is cheap, and this method
546 gives the highest performance while deferring most book-keeping to the
547 operating system. An alternative scheme would most likely reduce
548 performance and increase the complexity of the software, each of which
549 is much more important to the ``feel'' of day-to-day use.
551 \subsection{Other contents of the dirstate}
553 Because Mercurial doesn't force you to tell it when you're modifying a
554 file, it uses the dirstate to store some extra information so it can
555 determine efficiently whether you have modified a file. For each file
556 in the working directory, it stores the time that it last modified the
557 file itself, and the size of the file at that time.
559 When you explicitly \hgcmd{add}, \hgcmd{remove}, \hgcmd{rename} or
560 \hgcmd{copy} files, the dirstate is updated each time.
562 When Mercurial is checking the states of files in the working
563 directory, it first checks a file's modification time. If that has
564 not changed, the file must not have been modified. If the file's size
565 has changed, the file must have been modified. If the modification
566 time has changed, but the size has not, only then does Mercurial need
567 to read the actual contents of the file to see if they've changed.
568 Storing these few extra pieces of information dramatically reduces the
569 amount of data that Mercurial needs to read, which yields large
570 performance improvements compared to other revision control systems.
572 %%% Local Variables:
573 %%% mode: latex
574 %%% TeX-master: "00book"
575 %%% End: