rev |
line source |
jerojasro@393
|
1 \chapter{Tras bambalinas}
|
jerojasro@343
|
2 \label{chap:concepts}
|
jerojasro@343
|
3
|
jerojasro@393
|
4 A diferencia de varios sistemas de control de revisiones, los
|
jerojasro@393
|
5 conceptos en los que se fundamenta Mercurial son lo suficientemente
|
jerojasro@393
|
6 simples como para entender fácilmente cómo funciona el software.
|
jerojasro@393
|
7 Saber esto no es necesario, pero considero útil tener un ``modelo
|
jerojasro@393
|
8 mental'' de qué es lo que sucede.
|
jerojasro@393
|
9
|
jerojasro@393
|
10 Comprender esto me da la confianza de que Mercurial ha sido
|
jerojasro@393
|
11 cuidadosamente diseñado para ser tanto \emph{seguro} como
|
jerojasro@393
|
12 \emph{eficiente}. Y tal vez con la misma importancia, si es fácil
|
jerojasro@393
|
13 para mí hacerme a una idea adecuada de qué está haciendo el software
|
jerojasro@393
|
14 cuando llevo a cabo una tarea relacionada con control de revisiones,
|
jerojasro@393
|
15 es menos probable que me sosprenda su comportamiento.
|
jerojasro@393
|
16
|
jerojasro@393
|
17 En este capítulo, cubriremos inicialmente los conceptos centrales
|
jerojasro@393
|
18 del diseño de Mercurial, y luego discutiremos algunos detalles
|
jerojasro@393
|
19 interesantes de su implementación.
|
jerojasro@393
|
20
|
jerojasro@393
|
21 \section{Registro del historial de Mercurial}
|
jerojasro@393
|
22
|
jerojasro@393
|
23 \subsection{Seguir el historial de un único fichero}
|
jerojasro@393
|
24
|
jerojasro@393
|
25 Cuando Mercurial sigue las modificaciones a un fichero, guarda el
|
jerojasro@393
|
26 historial de dicho fichero en un objeto de metadatos llamado
|
jerojasro@395
|
27 \emph{filelog}\ndt{Fichero de registro}. Cada entrada en el fichero
|
jerojasro@395
|
28 de registro contiene suficiente información para reconstruir una
|
jerojasro@395
|
29 revisión del fichero que se está siguiendo. Los ficheros de registro
|
jerojasro@395
|
30 son almacenados como ficheros el el directorio
|
jerojasro@395
|
31 \sdirname{.hg/store/data}. Un fichero de registro contiene dos tipos
|
jerojasro@395
|
32 de información: datos de revisiones, y un índice para ayudar a
|
jerojasro@395
|
33 Mercurial a buscar revisiones eficientemente.
|
jerojasro@395
|
34
|
jerojasro@395
|
35 El fichero de registro de un fichero grande, o con un historial muy
|
jerojasro@395
|
36 largo, es guardado como ficheros separados para datos (sufijo
|
jerojasro@395
|
37 ``\texttt{.d}'') y para el índice (sufijo ``\texttt{.i}''). Para
|
jerojasro@395
|
38 ficheros pequeños con un historial pequeño, los datos de revisiones y
|
jerojasro@395
|
39 el índice son combinados en un único fichero ``\texttt{.i}''. La
|
jerojasro@395
|
40 correspondencia entre un fichero en el directorio de trabajo y el
|
jerojasro@395
|
41 fichero de registro que hace seguimiento a su historial en el
|
jerojasro@395
|
42 repositorio se ilustra en la figura~\ref{fig:concepts:filelog}.
|
jerojasro@343
|
43
|
jerojasro@343
|
44 \begin{figure}[ht]
|
jerojasro@343
|
45 \centering
|
jerojasro@343
|
46 \grafix{filelog}
|
jerojasro@343
|
47 \caption{Relationships between files in working directory and
|
jerojasro@343
|
48 filelogs in repository}
|
jerojasro@343
|
49 \label{fig:concepts:filelog}
|
jerojasro@343
|
50 \end{figure}
|
jerojasro@343
|
51
|
jerojasro@343
|
52 \subsection{Managing tracked files}
|
jerojasro@343
|
53
|
jerojasro@343
|
54 Mercurial uses a structure called a \emph{manifest} to collect
|
jerojasro@343
|
55 together information about the files that it tracks. Each entry in
|
jerojasro@343
|
56 the manifest contains information about the files present in a single
|
jerojasro@343
|
57 changeset. An entry records which files are present in the changeset,
|
jerojasro@343
|
58 the revision of each file, and a few other pieces of file metadata.
|
jerojasro@343
|
59
|
jerojasro@343
|
60 \subsection{Recording changeset information}
|
jerojasro@343
|
61
|
jerojasro@343
|
62 The \emph{changelog} contains information about each changeset. Each
|
jerojasro@343
|
63 revision records who committed a change, the changeset comment, other
|
jerojasro@343
|
64 pieces of changeset-related information, and the revision of the
|
jerojasro@343
|
65 manifest to use.
|
jerojasro@343
|
66
|
jerojasro@343
|
67 \subsection{Relationships between revisions}
|
jerojasro@343
|
68
|
jerojasro@343
|
69 Within a changelog, a manifest, or a filelog, each revision stores a
|
jerojasro@343
|
70 pointer to its immediate parent (or to its two parents, if it's a
|
jerojasro@343
|
71 merge revision). As I mentioned above, there are also relationships
|
jerojasro@343
|
72 between revisions \emph{across} these structures, and they are
|
jerojasro@343
|
73 hierarchical in nature.
|
jerojasro@343
|
74
|
jerojasro@343
|
75 For every changeset in a repository, there is exactly one revision
|
jerojasro@343
|
76 stored in the changelog. Each revision of the changelog contains a
|
jerojasro@343
|
77 pointer to a single revision of the manifest. A revision of the
|
jerojasro@343
|
78 manifest stores a pointer to a single revision of each filelog tracked
|
jerojasro@343
|
79 when that changeset was created. These relationships are illustrated
|
jerojasro@343
|
80 in figure~\ref{fig:concepts:metadata}.
|
jerojasro@343
|
81
|
jerojasro@343
|
82 \begin{figure}[ht]
|
jerojasro@343
|
83 \centering
|
jerojasro@343
|
84 \grafix{metadata}
|
jerojasro@343
|
85 \caption{Metadata relationships}
|
jerojasro@343
|
86 \label{fig:concepts:metadata}
|
jerojasro@343
|
87 \end{figure}
|
jerojasro@343
|
88
|
jerojasro@343
|
89 As the illustration shows, there is \emph{not} a ``one to one''
|
jerojasro@343
|
90 relationship between revisions in the changelog, manifest, or filelog.
|
jerojasro@343
|
91 If the manifest hasn't changed between two changesets, the changelog
|
jerojasro@343
|
92 entries for those changesets will point to the same revision of the
|
jerojasro@343
|
93 manifest. If a file that Mercurial tracks hasn't changed between two
|
jerojasro@343
|
94 changesets, the entry for that file in the two revisions of the
|
jerojasro@343
|
95 manifest will point to the same revision of its filelog.
|
jerojasro@343
|
96
|
jerojasro@343
|
97 \section{Safe, efficient storage}
|
jerojasro@343
|
98
|
jerojasro@343
|
99 The underpinnings of changelogs, manifests, and filelogs are provided
|
jerojasro@343
|
100 by a single structure called the \emph{revlog}.
|
jerojasro@343
|
101
|
jerojasro@343
|
102 \subsection{Efficient storage}
|
jerojasro@343
|
103
|
jerojasro@343
|
104 The revlog provides efficient storage of revisions using a
|
jerojasro@343
|
105 \emph{delta} mechanism. Instead of storing a complete copy of a file
|
jerojasro@343
|
106 for each revision, it stores the changes needed to transform an older
|
jerojasro@343
|
107 revision into the new revision. For many kinds of file data, these
|
jerojasro@343
|
108 deltas are typically a fraction of a percent of the size of a full
|
jerojasro@343
|
109 copy of a file.
|
jerojasro@343
|
110
|
jerojasro@343
|
111 Some obsolete revision control systems can only work with deltas of
|
jerojasro@343
|
112 text files. They must either store binary files as complete snapshots
|
jerojasro@343
|
113 or encoded into a text representation, both of which are wasteful
|
jerojasro@343
|
114 approaches. Mercurial can efficiently handle deltas of files with
|
jerojasro@343
|
115 arbitrary binary contents; it doesn't need to treat text as special.
|
jerojasro@343
|
116
|
jerojasro@343
|
117 \subsection{Safe operation}
|
jerojasro@343
|
118 \label{sec:concepts:txn}
|
jerojasro@343
|
119
|
jerojasro@343
|
120 Mercurial only ever \emph{appends} data to the end of a revlog file.
|
jerojasro@343
|
121 It never modifies a section of a file after it has written it. This
|
jerojasro@343
|
122 is both more robust and efficient than schemes that need to modify or
|
jerojasro@343
|
123 rewrite data.
|
jerojasro@343
|
124
|
jerojasro@343
|
125 In addition, Mercurial treats every write as part of a
|
jerojasro@343
|
126 \emph{transaction} that can span a number of files. A transaction is
|
jerojasro@343
|
127 \emph{atomic}: either the entire transaction succeeds and its effects
|
jerojasro@343
|
128 are all visible to readers in one go, or the whole thing is undone.
|
jerojasro@343
|
129 This guarantee of atomicity means that if you're running two copies of
|
jerojasro@343
|
130 Mercurial, where one is reading data and one is writing it, the reader
|
jerojasro@343
|
131 will never see a partially written result that might confuse it.
|
jerojasro@343
|
132
|
jerojasro@343
|
133 The fact that Mercurial only appends to files makes it easier to
|
jerojasro@343
|
134 provide this transactional guarantee. The easier it is to do stuff
|
jerojasro@343
|
135 like this, the more confident you should be that it's done correctly.
|
jerojasro@343
|
136
|
jerojasro@343
|
137 \subsection{Fast retrieval}
|
jerojasro@343
|
138
|
jerojasro@343
|
139 Mercurial cleverly avoids a pitfall common to all earlier
|
jerojasro@343
|
140 revision control systems: the problem of \emph{inefficient retrieval}.
|
jerojasro@343
|
141 Most revision control systems store the contents of a revision as an
|
jerojasro@343
|
142 incremental series of modifications against a ``snapshot''. To
|
jerojasro@343
|
143 reconstruct a specific revision, you must first read the snapshot, and
|
jerojasro@343
|
144 then every one of the revisions between the snapshot and your target
|
jerojasro@343
|
145 revision. The more history that a file accumulates, the more
|
jerojasro@343
|
146 revisions you must read, hence the longer it takes to reconstruct a
|
jerojasro@343
|
147 particular revision.
|
jerojasro@343
|
148
|
jerojasro@343
|
149 \begin{figure}[ht]
|
jerojasro@343
|
150 \centering
|
jerojasro@343
|
151 \grafix{snapshot}
|
jerojasro@343
|
152 \caption{Snapshot of a revlog, with incremental deltas}
|
jerojasro@343
|
153 \label{fig:concepts:snapshot}
|
jerojasro@343
|
154 \end{figure}
|
jerojasro@343
|
155
|
jerojasro@343
|
156 The innovation that Mercurial applies to this problem is simple but
|
jerojasro@343
|
157 effective. Once the cumulative amount of delta information stored
|
jerojasro@343
|
158 since the last snapshot exceeds a fixed threshold, it stores a new
|
jerojasro@343
|
159 snapshot (compressed, of course), instead of another delta. This
|
jerojasro@343
|
160 makes it possible to reconstruct \emph{any} revision of a file
|
jerojasro@343
|
161 quickly. This approach works so well that it has since been copied by
|
jerojasro@343
|
162 several other revision control systems.
|
jerojasro@343
|
163
|
jerojasro@343
|
164 Figure~\ref{fig:concepts:snapshot} illustrates the idea. In an entry
|
jerojasro@343
|
165 in a revlog's index file, Mercurial stores the range of entries from
|
jerojasro@343
|
166 the data file that it must read to reconstruct a particular revision.
|
jerojasro@343
|
167
|
jerojasro@343
|
168 \subsubsection{Aside: the influence of video compression}
|
jerojasro@343
|
169
|
jerojasro@343
|
170 If you're familiar with video compression or have ever watched a TV
|
jerojasro@343
|
171 feed through a digital cable or satellite service, you may know that
|
jerojasro@343
|
172 most video compression schemes store each frame of video as a delta
|
jerojasro@343
|
173 against its predecessor frame. In addition, these schemes use
|
jerojasro@343
|
174 ``lossy'' compression techniques to increase the compression ratio, so
|
jerojasro@343
|
175 visual errors accumulate over the course of a number of inter-frame
|
jerojasro@343
|
176 deltas.
|
jerojasro@343
|
177
|
jerojasro@343
|
178 Because it's possible for a video stream to ``drop out'' occasionally
|
jerojasro@343
|
179 due to signal glitches, and to limit the accumulation of artefacts
|
jerojasro@343
|
180 introduced by the lossy compression process, video encoders
|
jerojasro@343
|
181 periodically insert a complete frame (called a ``key frame'') into the
|
jerojasro@343
|
182 video stream; the next delta is generated against that frame. This
|
jerojasro@343
|
183 means that if the video signal gets interrupted, it will resume once
|
jerojasro@343
|
184 the next key frame is received. Also, the accumulation of encoding
|
jerojasro@343
|
185 errors restarts anew with each key frame.
|
jerojasro@343
|
186
|
jerojasro@343
|
187 \subsection{Identification and strong integrity}
|
jerojasro@343
|
188
|
jerojasro@343
|
189 Along with delta or snapshot information, a revlog entry contains a
|
jerojasro@343
|
190 cryptographic hash of the data that it represents. This makes it
|
jerojasro@343
|
191 difficult to forge the contents of a revision, and easy to detect
|
jerojasro@343
|
192 accidental corruption.
|
jerojasro@343
|
193
|
jerojasro@343
|
194 Hashes provide more than a mere check against corruption; they are
|
jerojasro@343
|
195 used as the identifiers for revisions. The changeset identification
|
jerojasro@343
|
196 hashes that you see as an end user are from revisions of the
|
jerojasro@343
|
197 changelog. Although filelogs and the manifest also use hashes,
|
jerojasro@343
|
198 Mercurial only uses these behind the scenes.
|
jerojasro@343
|
199
|
jerojasro@343
|
200 Mercurial verifies that hashes are correct when it retrieves file
|
jerojasro@343
|
201 revisions and when it pulls changes from another repository. If it
|
jerojasro@343
|
202 encounters an integrity problem, it will complain and stop whatever
|
jerojasro@343
|
203 it's doing.
|
jerojasro@343
|
204
|
jerojasro@343
|
205 In addition to the effect it has on retrieval efficiency, Mercurial's
|
jerojasro@343
|
206 use of periodic snapshots makes it more robust against partial data
|
jerojasro@343
|
207 corruption. If a revlog becomes partly corrupted due to a hardware
|
jerojasro@343
|
208 error or system bug, it's often possible to reconstruct some or most
|
jerojasro@343
|
209 revisions from the uncorrupted sections of the revlog, both before and
|
jerojasro@343
|
210 after the corrupted section. This would not be possible with a
|
jerojasro@343
|
211 delta-only storage model.
|
jerojasro@343
|
212
|
jerojasro@343
|
213 \section{Revision history, branching,
|
jerojasro@343
|
214 and merging}
|
jerojasro@343
|
215
|
jerojasro@343
|
216 Every entry in a Mercurial revlog knows the identity of its immediate
|
jerojasro@343
|
217 ancestor revision, usually referred to as its \emph{parent}. In fact,
|
jerojasro@343
|
218 a revision contains room for not one parent, but two. Mercurial uses
|
jerojasro@343
|
219 a special hash, called the ``null ID'', to represent the idea ``there
|
jerojasro@343
|
220 is no parent here''. This hash is simply a string of zeroes.
|
jerojasro@343
|
221
|
jerojasro@343
|
222 In figure~\ref{fig:concepts:revlog}, you can see an example of the
|
jerojasro@343
|
223 conceptual structure of a revlog. Filelogs, manifests, and changelogs
|
jerojasro@343
|
224 all have this same structure; they differ only in the kind of data
|
jerojasro@343
|
225 stored in each delta or snapshot.
|
jerojasro@343
|
226
|
jerojasro@343
|
227 The first revision in a revlog (at the bottom of the image) has the
|
jerojasro@343
|
228 null ID in both of its parent slots. For a ``normal'' revision, its
|
jerojasro@343
|
229 first parent slot contains the ID of its parent revision, and its
|
jerojasro@343
|
230 second contains the null ID, indicating that the revision has only one
|
jerojasro@343
|
231 real parent. Any two revisions that have the same parent ID are
|
jerojasro@343
|
232 branches. A revision that represents a merge between branches has two
|
jerojasro@343
|
233 normal revision IDs in its parent slots.
|
jerojasro@343
|
234
|
jerojasro@343
|
235 \begin{figure}[ht]
|
jerojasro@343
|
236 \centering
|
jerojasro@343
|
237 \grafix{revlog}
|
jerojasro@343
|
238 \caption{}
|
jerojasro@343
|
239 \label{fig:concepts:revlog}
|
jerojasro@343
|
240 \end{figure}
|
jerojasro@343
|
241
|
jerojasro@343
|
242 \section{The working directory}
|
jerojasro@343
|
243
|
jerojasro@343
|
244 In the working directory, Mercurial stores a snapshot of the files
|
jerojasro@343
|
245 from the repository as of a particular changeset.
|
jerojasro@343
|
246
|
jerojasro@343
|
247 The working directory ``knows'' which changeset it contains. When you
|
jerojasro@343
|
248 update the working directory to contain a particular changeset,
|
jerojasro@343
|
249 Mercurial looks up the appropriate revision of the manifest to find
|
jerojasro@343
|
250 out which files it was tracking at the time that changeset was
|
jerojasro@343
|
251 committed, and which revision of each file was then current. It then
|
jerojasro@343
|
252 recreates a copy of each of those files, with the same contents it had
|
jerojasro@343
|
253 when the changeset was committed.
|
jerojasro@343
|
254
|
jerojasro@343
|
255 The \emph{dirstate} contains Mercurial's knowledge of the working
|
jerojasro@343
|
256 directory. This details which changeset the working directory is
|
jerojasro@343
|
257 updated to, and all of the files that Mercurial is tracking in the
|
jerojasro@343
|
258 working directory.
|
jerojasro@343
|
259
|
jerojasro@343
|
260 Just as a revision of a revlog has room for two parents, so that it
|
jerojasro@343
|
261 can represent either a normal revision (with one parent) or a merge of
|
jerojasro@343
|
262 two earlier revisions, the dirstate has slots for two parents. When
|
jerojasro@343
|
263 you use the \hgcmd{update} command, the changeset that you update to
|
jerojasro@343
|
264 is stored in the ``first parent'' slot, and the null ID in the second.
|
jerojasro@343
|
265 When you \hgcmd{merge} with another changeset, the first parent
|
jerojasro@343
|
266 remains unchanged, and the second parent is filled in with the
|
jerojasro@343
|
267 changeset you're merging with. The \hgcmd{parents} command tells you
|
jerojasro@343
|
268 what the parents of the dirstate are.
|
jerojasro@343
|
269
|
jerojasro@343
|
270 \subsection{What happens when you commit}
|
jerojasro@343
|
271
|
jerojasro@343
|
272 The dirstate stores parent information for more than just book-keeping
|
jerojasro@343
|
273 purposes. Mercurial uses the parents of the dirstate as \emph{the
|
jerojasro@343
|
274 parents of a new changeset} when you perform a commit.
|
jerojasro@343
|
275
|
jerojasro@343
|
276 \begin{figure}[ht]
|
jerojasro@343
|
277 \centering
|
jerojasro@343
|
278 \grafix{wdir}
|
jerojasro@343
|
279 \caption{The working directory can have two parents}
|
jerojasro@343
|
280 \label{fig:concepts:wdir}
|
jerojasro@343
|
281 \end{figure}
|
jerojasro@343
|
282
|
jerojasro@343
|
283 Figure~\ref{fig:concepts:wdir} shows the normal state of the working
|
jerojasro@343
|
284 directory, where it has a single changeset as parent. That changeset
|
jerojasro@343
|
285 is the \emph{tip}, the newest changeset in the repository that has no
|
jerojasro@343
|
286 children.
|
jerojasro@343
|
287
|
jerojasro@343
|
288 \begin{figure}[ht]
|
jerojasro@343
|
289 \centering
|
jerojasro@343
|
290 \grafix{wdir-after-commit}
|
jerojasro@343
|
291 \caption{The working directory gains new parents after a commit}
|
jerojasro@343
|
292 \label{fig:concepts:wdir-after-commit}
|
jerojasro@343
|
293 \end{figure}
|
jerojasro@343
|
294
|
jerojasro@343
|
295 It's useful to think of the working directory as ``the changeset I'm
|
jerojasro@343
|
296 about to commit''. Any files that you tell Mercurial that you've
|
jerojasro@343
|
297 added, removed, renamed, or copied will be reflected in that
|
jerojasro@343
|
298 changeset, as will modifications to any files that Mercurial is
|
jerojasro@343
|
299 already tracking; the new changeset will have the parents of the
|
jerojasro@343
|
300 working directory as its parents.
|
jerojasro@343
|
301
|
jerojasro@343
|
302 After a commit, Mercurial will update the parents of the working
|
jerojasro@343
|
303 directory, so that the first parent is the ID of the new changeset,
|
jerojasro@343
|
304 and the second is the null ID. This is shown in
|
jerojasro@343
|
305 figure~\ref{fig:concepts:wdir-after-commit}. Mercurial doesn't touch
|
jerojasro@343
|
306 any of the files in the working directory when you commit; it just
|
jerojasro@343
|
307 modifies the dirstate to note its new parents.
|
jerojasro@343
|
308
|
jerojasro@343
|
309 \subsection{Creating a new head}
|
jerojasro@343
|
310
|
jerojasro@343
|
311 It's perfectly normal to update the working directory to a changeset
|
jerojasro@343
|
312 other than the current tip. For example, you might want to know what
|
jerojasro@343
|
313 your project looked like last Tuesday, or you could be looking through
|
jerojasro@343
|
314 changesets to see which one introduced a bug. In cases like this, the
|
jerojasro@343
|
315 natural thing to do is update the working directory to the changeset
|
jerojasro@343
|
316 you're interested in, and then examine the files in the working
|
jerojasro@343
|
317 directory directly to see their contents as they werea when you
|
jerojasro@343
|
318 committed that changeset. The effect of this is shown in
|
jerojasro@343
|
319 figure~\ref{fig:concepts:wdir-pre-branch}.
|
jerojasro@343
|
320
|
jerojasro@343
|
321 \begin{figure}[ht]
|
jerojasro@343
|
322 \centering
|
jerojasro@343
|
323 \grafix{wdir-pre-branch}
|
jerojasro@343
|
324 \caption{The working directory, updated to an older changeset}
|
jerojasro@343
|
325 \label{fig:concepts:wdir-pre-branch}
|
jerojasro@343
|
326 \end{figure}
|
jerojasro@343
|
327
|
jerojasro@343
|
328 Having updated the working directory to an older changeset, what
|
jerojasro@343
|
329 happens if you make some changes, and then commit? Mercurial behaves
|
jerojasro@343
|
330 in the same way as I outlined above. The parents of the working
|
jerojasro@343
|
331 directory become the parents of the new changeset. This new changeset
|
jerojasro@343
|
332 has no children, so it becomes the new tip. And the repository now
|
jerojasro@343
|
333 contains two changesets that have no children; we call these
|
jerojasro@343
|
334 \emph{heads}. You can see the structure that this creates in
|
jerojasro@343
|
335 figure~\ref{fig:concepts:wdir-branch}.
|
jerojasro@343
|
336
|
jerojasro@343
|
337 \begin{figure}[ht]
|
jerojasro@343
|
338 \centering
|
jerojasro@343
|
339 \grafix{wdir-branch}
|
jerojasro@343
|
340 \caption{After a commit made while synced to an older changeset}
|
jerojasro@343
|
341 \label{fig:concepts:wdir-branch}
|
jerojasro@343
|
342 \end{figure}
|
jerojasro@343
|
343
|
jerojasro@343
|
344 \begin{note}
|
jerojasro@343
|
345 If you're new to Mercurial, you should keep in mind a common
|
jerojasro@343
|
346 ``error'', which is to use the \hgcmd{pull} command without any
|
jerojasro@343
|
347 options. By default, the \hgcmd{pull} command \emph{does not}
|
jerojasro@343
|
348 update the working directory, so you'll bring new changesets into
|
jerojasro@343
|
349 your repository, but the working directory will stay synced at the
|
jerojasro@343
|
350 same changeset as before the pull. If you make some changes and
|
jerojasro@343
|
351 commit afterwards, you'll thus create a new head, because your
|
jerojasro@343
|
352 working directory isn't synced to whatever the current tip is.
|
jerojasro@343
|
353
|
jerojasro@343
|
354 I put the word ``error'' in quotes because all that you need to do
|
jerojasro@343
|
355 to rectify this situation is \hgcmd{merge}, then \hgcmd{commit}. In
|
jerojasro@343
|
356 other words, this almost never has negative consequences; it just
|
jerojasro@343
|
357 surprises people. I'll discuss other ways to avoid this behaviour,
|
jerojasro@343
|
358 and why Mercurial behaves in this initially surprising way, later
|
jerojasro@343
|
359 on.
|
jerojasro@343
|
360 \end{note}
|
jerojasro@343
|
361
|
jerojasro@343
|
362 \subsection{Merging heads}
|
jerojasro@343
|
363
|
jerojasro@343
|
364 When you run the \hgcmd{merge} command, Mercurial leaves the first
|
jerojasro@343
|
365 parent of the working directory unchanged, and sets the second parent
|
jerojasro@343
|
366 to the changeset you're merging with, as shown in
|
jerojasro@343
|
367 figure~\ref{fig:concepts:wdir-merge}.
|
jerojasro@343
|
368
|
jerojasro@343
|
369 \begin{figure}[ht]
|
jerojasro@343
|
370 \centering
|
jerojasro@343
|
371 \grafix{wdir-merge}
|
jerojasro@343
|
372 \caption{Merging two heads}
|
jerojasro@343
|
373 \label{fig:concepts:wdir-merge}
|
jerojasro@343
|
374 \end{figure}
|
jerojasro@343
|
375
|
jerojasro@343
|
376 Mercurial also has to modify the working directory, to merge the files
|
jerojasro@343
|
377 managed in the two changesets. Simplified a little, the merging
|
jerojasro@343
|
378 process goes like this, for every file in the manifests of both
|
jerojasro@343
|
379 changesets.
|
jerojasro@343
|
380 \begin{itemize}
|
jerojasro@343
|
381 \item If neither changeset has modified a file, do nothing with that
|
jerojasro@343
|
382 file.
|
jerojasro@343
|
383 \item If one changeset has modified a file, and the other hasn't,
|
jerojasro@343
|
384 create the modified copy of the file in the working directory.
|
jerojasro@343
|
385 \item If one changeset has removed a file, and the other hasn't (or
|
jerojasro@343
|
386 has also deleted it), delete the file from the working directory.
|
jerojasro@343
|
387 \item If one changeset has removed a file, but the other has modified
|
jerojasro@343
|
388 the file, ask the user what to do: keep the modified file, or remove
|
jerojasro@343
|
389 it?
|
jerojasro@343
|
390 \item If both changesets have modified a file, invoke an external
|
jerojasro@343
|
391 merge program to choose the new contents for the merged file. This
|
jerojasro@343
|
392 may require input from the user.
|
jerojasro@343
|
393 \item If one changeset has modified a file, and the other has renamed
|
jerojasro@343
|
394 or copied the file, make sure that the changes follow the new name
|
jerojasro@343
|
395 of the file.
|
jerojasro@343
|
396 \end{itemize}
|
jerojasro@343
|
397 There are more details---merging has plenty of corner cases---but
|
jerojasro@343
|
398 these are the most common choices that are involved in a merge. As
|
jerojasro@343
|
399 you can see, most cases are completely automatic, and indeed most
|
jerojasro@343
|
400 merges finish automatically, without requiring your input to resolve
|
jerojasro@343
|
401 any conflicts.
|
jerojasro@343
|
402
|
jerojasro@343
|
403 When you're thinking about what happens when you commit after a merge,
|
jerojasro@343
|
404 once again the working directory is ``the changeset I'm about to
|
jerojasro@343
|
405 commit''. After the \hgcmd{merge} command completes, the working
|
jerojasro@343
|
406 directory has two parents; these will become the parents of the new
|
jerojasro@343
|
407 changeset.
|
jerojasro@343
|
408
|
jerojasro@343
|
409 Mercurial lets you perform multiple merges, but you must commit the
|
jerojasro@343
|
410 results of each individual merge as you go. This is necessary because
|
jerojasro@343
|
411 Mercurial only tracks two parents for both revisions and the working
|
jerojasro@343
|
412 directory. While it would be technically possible to merge multiple
|
jerojasro@343
|
413 changesets at once, the prospect of user confusion and making a
|
jerojasro@343
|
414 terrible mess of a merge immediately becomes overwhelming.
|
jerojasro@343
|
415
|
jerojasro@343
|
416 \section{Other interesting design features}
|
jerojasro@343
|
417
|
jerojasro@343
|
418 In the sections above, I've tried to highlight some of the most
|
jerojasro@343
|
419 important aspects of Mercurial's design, to illustrate that it pays
|
jerojasro@343
|
420 careful attention to reliability and performance. However, the
|
jerojasro@343
|
421 attention to detail doesn't stop there. There are a number of other
|
jerojasro@343
|
422 aspects of Mercurial's construction that I personally find
|
jerojasro@343
|
423 interesting. I'll detail a few of them here, separate from the ``big
|
jerojasro@343
|
424 ticket'' items above, so that if you're interested, you can gain a
|
jerojasro@343
|
425 better idea of the amount of thinking that goes into a well-designed
|
jerojasro@343
|
426 system.
|
jerojasro@343
|
427
|
jerojasro@343
|
428 \subsection{Clever compression}
|
jerojasro@343
|
429
|
jerojasro@343
|
430 When appropriate, Mercurial will store both snapshots and deltas in
|
jerojasro@343
|
431 compressed form. It does this by always \emph{trying to} compress a
|
jerojasro@343
|
432 snapshot or delta, but only storing the compressed version if it's
|
jerojasro@343
|
433 smaller than the uncompressed version.
|
jerojasro@343
|
434
|
jerojasro@343
|
435 This means that Mercurial does ``the right thing'' when storing a file
|
jerojasro@343
|
436 whose native form is compressed, such as a \texttt{zip} archive or a
|
jerojasro@343
|
437 JPEG image. When these types of files are compressed a second time,
|
jerojasro@343
|
438 the resulting file is usually bigger than the once-compressed form,
|
jerojasro@343
|
439 and so Mercurial will store the plain \texttt{zip} or JPEG.
|
jerojasro@343
|
440
|
jerojasro@343
|
441 Deltas between revisions of a compressed file are usually larger than
|
jerojasro@343
|
442 snapshots of the file, and Mercurial again does ``the right thing'' in
|
jerojasro@343
|
443 these cases. It finds that such a delta exceeds the threshold at
|
jerojasro@343
|
444 which it should store a complete snapshot of the file, so it stores
|
jerojasro@343
|
445 the snapshot, again saving space compared to a naive delta-only
|
jerojasro@343
|
446 approach.
|
jerojasro@343
|
447
|
jerojasro@343
|
448 \subsubsection{Network recompression}
|
jerojasro@343
|
449
|
jerojasro@343
|
450 When storing revisions on disk, Mercurial uses the ``deflate''
|
jerojasro@343
|
451 compression algorithm (the same one used by the popular \texttt{zip}
|
jerojasro@343
|
452 archive format), which balances good speed with a respectable
|
jerojasro@343
|
453 compression ratio. However, when transmitting revision data over a
|
jerojasro@343
|
454 network connection, Mercurial uncompresses the compressed revision
|
jerojasro@343
|
455 data.
|
jerojasro@343
|
456
|
jerojasro@343
|
457 If the connection is over HTTP, Mercurial recompresses the entire
|
jerojasro@343
|
458 stream of data using a compression algorithm that gives a better
|
jerojasro@343
|
459 compression ratio (the Burrows-Wheeler algorithm from the widely used
|
jerojasro@343
|
460 \texttt{bzip2} compression package). This combination of algorithm
|
jerojasro@343
|
461 and compression of the entire stream (instead of a revision at a time)
|
jerojasro@343
|
462 substantially reduces the number of bytes to be transferred, yielding
|
jerojasro@343
|
463 better network performance over almost all kinds of network.
|
jerojasro@343
|
464
|
jerojasro@343
|
465 (If the connection is over \command{ssh}, Mercurial \emph{doesn't}
|
jerojasro@343
|
466 recompress the stream, because \command{ssh} can already do this
|
jerojasro@343
|
467 itself.)
|
jerojasro@343
|
468
|
jerojasro@343
|
469 \subsection{Read/write ordering and atomicity}
|
jerojasro@343
|
470
|
jerojasro@343
|
471 Appending to files isn't the whole story when it comes to guaranteeing
|
jerojasro@343
|
472 that a reader won't see a partial write. If you recall
|
jerojasro@343
|
473 figure~\ref{fig:concepts:metadata}, revisions in the changelog point to
|
jerojasro@343
|
474 revisions in the manifest, and revisions in the manifest point to
|
jerojasro@343
|
475 revisions in filelogs. This hierarchy is deliberate.
|
jerojasro@343
|
476
|
jerojasro@343
|
477 A writer starts a transaction by writing filelog and manifest data,
|
jerojasro@343
|
478 and doesn't write any changelog data until those are finished. A
|
jerojasro@343
|
479 reader starts by reading changelog data, then manifest data, followed
|
jerojasro@343
|
480 by filelog data.
|
jerojasro@343
|
481
|
jerojasro@343
|
482 Since the writer has always finished writing filelog and manifest data
|
jerojasro@343
|
483 before it writes to the changelog, a reader will never read a pointer
|
jerojasro@343
|
484 to a partially written manifest revision from the changelog, and it will
|
jerojasro@343
|
485 never read a pointer to a partially written filelog revision from the
|
jerojasro@343
|
486 manifest.
|
jerojasro@343
|
487
|
jerojasro@343
|
488 \subsection{Concurrent access}
|
jerojasro@343
|
489
|
jerojasro@343
|
490 The read/write ordering and atomicity guarantees mean that Mercurial
|
jerojasro@343
|
491 never needs to \emph{lock} a repository when it's reading data, even
|
jerojasro@343
|
492 if the repository is being written to while the read is occurring.
|
jerojasro@343
|
493 This has a big effect on scalability; you can have an arbitrary number
|
jerojasro@343
|
494 of Mercurial processes safely reading data from a repository safely
|
jerojasro@343
|
495 all at once, no matter whether it's being written to or not.
|
jerojasro@343
|
496
|
jerojasro@343
|
497 The lockless nature of reading means that if you're sharing a
|
jerojasro@343
|
498 repository on a multi-user system, you don't need to grant other local
|
jerojasro@343
|
499 users permission to \emph{write} to your repository in order for them
|
jerojasro@343
|
500 to be able to clone it or pull changes from it; they only need
|
jerojasro@343
|
501 \emph{read} permission. (This is \emph{not} a common feature among
|
jerojasro@343
|
502 revision control systems, so don't take it for granted! Most require
|
jerojasro@343
|
503 readers to be able to lock a repository to access it safely, and this
|
jerojasro@343
|
504 requires write permission on at least one directory, which of course
|
jerojasro@343
|
505 makes for all kinds of nasty and annoying security and administrative
|
jerojasro@343
|
506 problems.)
|
jerojasro@343
|
507
|
jerojasro@343
|
508 Mercurial uses locks to ensure that only one process can write to a
|
jerojasro@343
|
509 repository at a time (the locking mechanism is safe even over
|
jerojasro@343
|
510 filesystems that are notoriously hostile to locking, such as NFS). If
|
jerojasro@343
|
511 a repository is locked, a writer will wait for a while to retry if the
|
jerojasro@343
|
512 repository becomes unlocked, but if the repository remains locked for
|
jerojasro@343
|
513 too long, the process attempting to write will time out after a while.
|
jerojasro@343
|
514 This means that your daily automated scripts won't get stuck forever
|
jerojasro@343
|
515 and pile up if a system crashes unnoticed, for example. (Yes, the
|
jerojasro@343
|
516 timeout is configurable, from zero to infinity.)
|
jerojasro@343
|
517
|
jerojasro@343
|
518 \subsubsection{Safe dirstate access}
|
jerojasro@343
|
519
|
jerojasro@343
|
520 As with revision data, Mercurial doesn't take a lock to read the
|
jerojasro@343
|
521 dirstate file; it does acquire a lock to write it. To avoid the
|
jerojasro@343
|
522 possibility of reading a partially written copy of the dirstate file,
|
jerojasro@343
|
523 Mercurial writes to a file with a unique name in the same directory as
|
jerojasro@343
|
524 the dirstate file, then renames the temporary file atomically to
|
jerojasro@343
|
525 \filename{dirstate}. The file named \filename{dirstate} is thus
|
jerojasro@343
|
526 guaranteed to be complete, not partially written.
|
jerojasro@343
|
527
|
jerojasro@343
|
528 \subsection{Avoiding seeks}
|
jerojasro@343
|
529
|
jerojasro@343
|
530 Critical to Mercurial's performance is the avoidance of seeks of the
|
jerojasro@343
|
531 disk head, since any seek is far more expensive than even a
|
jerojasro@343
|
532 comparatively large read operation.
|
jerojasro@343
|
533
|
jerojasro@343
|
534 This is why, for example, the dirstate is stored in a single file. If
|
jerojasro@343
|
535 there were a dirstate file per directory that Mercurial tracked, the
|
jerojasro@343
|
536 disk would seek once per directory. Instead, Mercurial reads the
|
jerojasro@343
|
537 entire single dirstate file in one step.
|
jerojasro@343
|
538
|
jerojasro@343
|
539 Mercurial also uses a ``copy on write'' scheme when cloning a
|
jerojasro@343
|
540 repository on local storage. Instead of copying every revlog file
|
jerojasro@343
|
541 from the old repository into the new repository, it makes a ``hard
|
jerojasro@343
|
542 link'', which is a shorthand way to say ``these two names point to the
|
jerojasro@343
|
543 same file''. When Mercurial is about to write to one of a revlog's
|
jerojasro@343
|
544 files, it checks to see if the number of names pointing at the file is
|
jerojasro@343
|
545 greater than one. If it is, more than one repository is using the
|
jerojasro@343
|
546 file, so Mercurial makes a new copy of the file that is private to
|
jerojasro@343
|
547 this repository.
|
jerojasro@343
|
548
|
jerojasro@343
|
549 A few revision control developers have pointed out that this idea of
|
jerojasro@343
|
550 making a complete private copy of a file is not very efficient in its
|
jerojasro@343
|
551 use of storage. While this is true, storage is cheap, and this method
|
jerojasro@343
|
552 gives the highest performance while deferring most book-keeping to the
|
jerojasro@343
|
553 operating system. An alternative scheme would most likely reduce
|
jerojasro@343
|
554 performance and increase the complexity of the software, each of which
|
jerojasro@343
|
555 is much more important to the ``feel'' of day-to-day use.
|
jerojasro@343
|
556
|
jerojasro@343
|
557 \subsection{Other contents of the dirstate}
|
jerojasro@343
|
558
|
jerojasro@343
|
559 Because Mercurial doesn't force you to tell it when you're modifying a
|
jerojasro@343
|
560 file, it uses the dirstate to store some extra information so it can
|
jerojasro@343
|
561 determine efficiently whether you have modified a file. For each file
|
jerojasro@343
|
562 in the working directory, it stores the time that it last modified the
|
jerojasro@343
|
563 file itself, and the size of the file at that time.
|
jerojasro@343
|
564
|
jerojasro@343
|
565 When you explicitly \hgcmd{add}, \hgcmd{remove}, \hgcmd{rename} or
|
jerojasro@343
|
566 \hgcmd{copy} files, Mercurial updates the dirstate so that it knows
|
jerojasro@343
|
567 what to do with those files when you commit.
|
jerojasro@343
|
568
|
jerojasro@343
|
569 When Mercurial is checking the states of files in the working
|
jerojasro@343
|
570 directory, it first checks a file's modification time. If that has
|
jerojasro@343
|
571 not changed, the file must not have been modified. If the file's size
|
jerojasro@343
|
572 has changed, the file must have been modified. If the modification
|
jerojasro@343
|
573 time has changed, but the size has not, only then does Mercurial need
|
jerojasro@343
|
574 to read the actual contents of the file to see if they've changed.
|
jerojasro@343
|
575 Storing these few extra pieces of information dramatically reduces the
|
jerojasro@343
|
576 amount of data that Mercurial needs to read, which yields large
|
jerojasro@343
|
577 performance improvements compared to other revision control systems.
|
jerojasro@343
|
578
|
jerojasro@343
|
579 %%% Local Variables:
|
jerojasro@343
|
580 %%% mode: latex
|
jerojasro@343
|
581 %%% TeX-master: "00book"
|
jerojasro@343
|
582 %%% End:
|