hgbook

annotate es/undo.tex @ 374:c2ff82128ac5

find changesets by author, revision, files, or words in the commit message
Started translating undo.tex , added mistake and rollback
author Igor TAmara <igor@tamarapatino.org>
date Mon Oct 27 13:32:16 2008 -0500 (2008-10-27)
parents 2fb78d342e07
children d5f1049a79dd
rev   line source
igor@374 1 \chapter{Encontrar y arreglar sus equivocaciones}
jerojasro@343 2 \label{chap:undo}
jerojasro@343 3
igor@374 4 Errar es humano, pero tratar adecuadamente las consecuencias requiere
igor@374 5 un sistema de control de revisiones de primera categoría. En este
igor@374 6 capítulo, discutiremos algunas técnicas que puede usar cuando
igor@374 7 encuentra que hay un problema enraizado en su proyecto. Mercurial
igor@374 8 tiene unas características poderosas que le ayudarán a isolar las
igor@374 9 fuentes de los problemas, y a dar cuenta de ellas apropiadamente.
igor@374 10
igor@374 11 \section{Borrar la historia local}
igor@374 12
igor@374 13 \subsection{La consignación accidental}
igor@374 14
igor@374 15 Tengo el problema ocasional, pero persistente de teclear más rápido de
igor@374 16 lo que pienso, que aveces resulta en consignar un conjunto de cambios
igor@374 17 incompleto o simplemente malo. En mi caso, el conjunto de cambios
igor@374 18 incompleto consiste en que creé un nuevo archivo fuente, pero olvidé
igor@374 19 hacerle \hgcmd{add}. Un conjunto de cambios``simplemente malo'' no es
igor@374 20 tan común, pero sí resulta muy molesto.
igor@374 21
igor@374 22 \subsection{Retroceder una transacción}
jerojasro@343 23 \label{sec:undo:rollback}
jerojasro@343 24
igor@374 25 En la sección~\ref{sec:concepts:txn}, mencioné que Mercurial trata
igor@374 26 modificación a un repositorio como una \emph{transacción}. Cada vez
igor@374 27 que consigna un conjunto de cambios o lo jala de otro repositorio,
igor@374 28 Mercurial recuerda lo que hizo. Puede deshacer, o \emph{retroceder},
igor@374 29 exactamente una de tales acciones usando la orden \hgcmd{rollback}.
igor@374 30 (Ver en la sección~\ref{sec:undo:rollback-after-push} una anotación
igor@374 31 importante acerca del uso de esta orden.)
igor@374 32
igor@374 33 A continuación una equivocación que me sucede frecuentemente:
igor@374 34 consignar un cambio en el cual he creado un nuevo fichero, pero he
igor@374 35 olvidado hacerle \hgcmd{add}.
jerojasro@343 36 \interaction{rollback.commit}
igor@374 37 La salida de \hgcmd{status} después de la consignación confirma
igor@374 38 inmediatamente este error.
jerojasro@343 39 \interaction{rollback.status}
igor@374 40 La consignación capturó los cambios en el archivo \filename{a}, pero
igor@374 41 no el nuevo fichero \filename{b}. Si yo publicara este conjunto de
igor@374 42 cambios a un repositorio compartido con un colega, es bastante
igor@374 43 probable que algo en \filename{a} se refiriera a \filename{b}, el cual
igor@374 44 podría no estar presente cuando jalen mis cambios del repositorio. Me
igor@374 45 convertiría el sujeto de cierta indignación.
igor@374 46
igor@374 47 Como sea, la suerte me acompaña---Encontré mi error antes de publicar
igor@374 48 el conjunto de cambios. Uso la orden \hgcmd{rollback}, y Mercurial
igor@374 49 hace desaparecer el último conjunto de cambios.
jerojasro@343 50 \interaction{rollback.rollback}
igor@374 51 El conjunto de cambios ya no está en la historia del repositorio, y el
igor@374 52 directorio de trabajo cree que el fichero \filename{a} ha sido
igor@374 53 modificado. La consignación y el retroceso dejaron el directorio de
igor@374 54 trabajo exactamente como estaban antes de la consignación; el conjunto
igor@374 55 de cambios ha sido eliminado totlamente. Ahora puedo hacer \hgcmd{add}
igor@374 56 al fichero \filename{b}, y hacer de nuevo la consignación.
jerojasro@343 57 \interaction{rollback.add}
jerojasro@343 58
igor@374 59 \subsection{Erroneamente jalado}
igor@374 60
igor@374 61 Mantener ramas de desarrollo separadas de un proyecto en distintos
igor@374 62 repositorios es una práctica común con Mercurial. Su equipo de
igor@374 63 desarrollo puede tener un repositorio compartido para la versión ``0.9''
igor@374 64 y otra con cambios distintos para la versión ``1.0''.
igor@374 65
igor@374 66 Con este escenario, puede imaginar las consecuencias si tuviera un
igor@374 67 repositorio local ``0.9'', y jalara accidentalmente los cambios del
igor@374 68 repositorio compartido de la versión ``1.0'' en este. En el peor de
igor@374 69 los casos, por falta de atención, es posible que publique tales
igor@374 70 cambios en el árbol compartido ``0.9'', confundiendo a todo su equipo
igor@374 71 de trabajo(pero no se preocupe, volveremos a este terrorífico
igor@374 72 escenario posteriormente). En todo caso, es muy probable que usted se
igor@374 73 de cuenta inmediatamente, dado que Mercurial mostrará el URL de donde
igor@374 74 está jalando, o que vea jalando una sospechosa gran cantidad de
igor@374 75 cambios en el repositorio.
igor@374 76
igor@374 77 La orden \hgcmd{rollback} command will work nicely to expunge all of the
jerojasro@343 78 changesets that you just pulled. Mercurial groups all changes from
jerojasro@343 79 one \hgcmd{pull} into a single transaction, so one \hgcmd{rollback} is
jerojasro@343 80 all you need to undo this mistake.
jerojasro@343 81
jerojasro@343 82 \subsection{Rolling back is useless once you've pushed}
jerojasro@343 83 \label{sec:undo:rollback-after-push}
jerojasro@343 84
jerojasro@343 85 The value of the \hgcmd{rollback} command drops to zero once you've
jerojasro@343 86 pushed your changes to another repository. Rolling back a change
jerojasro@343 87 makes it disappear entirely, but \emph{only} in the repository in
jerojasro@343 88 which you perform the \hgcmd{rollback}. Because a rollback eliminates
jerojasro@343 89 history, there's no way for the disappearance of a change to propagate
jerojasro@343 90 between repositories.
jerojasro@343 91
jerojasro@343 92 If you've pushed a change to another repository---particularly if it's
jerojasro@343 93 a shared repository---it has essentially ``escaped into the wild,''
jerojasro@343 94 and you'll have to recover from your mistake in a different way. What
jerojasro@343 95 will happen if you push a changeset somewhere, then roll it back, then
jerojasro@343 96 pull from the repository you pushed to, is that the changeset will
jerojasro@343 97 reappear in your repository.
jerojasro@343 98
jerojasro@343 99 (If you absolutely know for sure that the change you want to roll back
jerojasro@343 100 is the most recent change in the repository that you pushed to,
jerojasro@343 101 \emph{and} you know that nobody else could have pulled it from that
jerojasro@343 102 repository, you can roll back the changeset there, too, but you really
jerojasro@343 103 should really not rely on this working reliably. If you do this,
jerojasro@343 104 sooner or later a change really will make it into a repository that
jerojasro@343 105 you don't directly control (or have forgotten about), and come back to
jerojasro@343 106 bite you.)
jerojasro@343 107
jerojasro@343 108 \subsection{You can only roll back once}
jerojasro@343 109
jerojasro@343 110 Mercurial stores exactly one transaction in its transaction log; that
jerojasro@343 111 transaction is the most recent one that occurred in the repository.
jerojasro@343 112 This means that you can only roll back one transaction. If you expect
jerojasro@343 113 to be able to roll back one transaction, then its predecessor, this is
jerojasro@343 114 not the behaviour you will get.
jerojasro@343 115 \interaction{rollback.twice}
jerojasro@343 116 Once you've rolled back one transaction in a repository, you can't
jerojasro@343 117 roll back again in that repository until you perform another commit or
jerojasro@343 118 pull.
jerojasro@343 119
jerojasro@343 120 \section{Reverting the mistaken change}
jerojasro@343 121
jerojasro@343 122 If you make a modification to a file, and decide that you really
jerojasro@343 123 didn't want to change the file at all, and you haven't yet committed
jerojasro@343 124 your changes, the \hgcmd{revert} command is the one you'll need. It
jerojasro@343 125 looks at the changeset that's the parent of the working directory, and
jerojasro@343 126 restores the contents of the file to their state as of that changeset.
jerojasro@343 127 (That's a long-winded way of saying that, in the normal case, it
jerojasro@343 128 undoes your modifications.)
jerojasro@343 129
jerojasro@343 130 Let's illustrate how the \hgcmd{revert} command works with yet another
jerojasro@343 131 small example. We'll begin by modifying a file that Mercurial is
jerojasro@343 132 already tracking.
jerojasro@343 133 \interaction{daily.revert.modify}
jerojasro@343 134 If we don't want that change, we can simply \hgcmd{revert} the file.
jerojasro@343 135 \interaction{daily.revert.unmodify}
jerojasro@343 136 The \hgcmd{revert} command provides us with an extra degree of safety
jerojasro@343 137 by saving our modified file with a \filename{.orig} extension.
jerojasro@343 138 \interaction{daily.revert.status}
jerojasro@343 139
jerojasro@343 140 Here is a summary of the cases that the \hgcmd{revert} command can
jerojasro@343 141 deal with. We will describe each of these in more detail in the
jerojasro@343 142 section that follows.
jerojasro@343 143 \begin{itemize}
jerojasro@343 144 \item If you modify a file, it will restore the file to its unmodified
jerojasro@343 145 state.
jerojasro@343 146 \item If you \hgcmd{add} a file, it will undo the ``added'' state of
jerojasro@343 147 the file, but leave the file itself untouched.
jerojasro@343 148 \item If you delete a file without telling Mercurial, it will restore
jerojasro@343 149 the file to its unmodified contents.
jerojasro@343 150 \item If you use the \hgcmd{remove} command to remove a file, it will
jerojasro@343 151 undo the ``removed'' state of the file, and restore the file to its
jerojasro@343 152 unmodified contents.
jerojasro@343 153 \end{itemize}
jerojasro@343 154
jerojasro@343 155 \subsection{File management errors}
jerojasro@343 156 \label{sec:undo:mgmt}
jerojasro@343 157
jerojasro@343 158 The \hgcmd{revert} command is useful for more than just modified
jerojasro@343 159 files. It lets you reverse the results of all of Mercurial's file
jerojasro@343 160 management commands---\hgcmd{add}, \hgcmd{remove}, and so on.
jerojasro@343 161
jerojasro@343 162 If you \hgcmd{add} a file, then decide that in fact you don't want
jerojasro@343 163 Mercurial to track it, use \hgcmd{revert} to undo the add. Don't
jerojasro@343 164 worry; Mercurial will not modify the file in any way. It will just
jerojasro@343 165 ``unmark'' the file.
jerojasro@343 166 \interaction{daily.revert.add}
jerojasro@343 167
jerojasro@343 168 Similarly, if you ask Mercurial to \hgcmd{remove} a file, you can use
jerojasro@343 169 \hgcmd{revert} to restore it to the contents it had as of the parent
jerojasro@343 170 of the working directory.
jerojasro@343 171 \interaction{daily.revert.remove}
jerojasro@343 172 This works just as well for a file that you deleted by hand, without
jerojasro@343 173 telling Mercurial (recall that in Mercurial terminology, this kind of
jerojasro@343 174 file is called ``missing'').
jerojasro@343 175 \interaction{daily.revert.missing}
jerojasro@343 176
jerojasro@343 177 If you revert a \hgcmd{copy}, the copied-to file remains in your
jerojasro@343 178 working directory afterwards, untracked. Since a copy doesn't affect
jerojasro@343 179 the copied-from file in any way, Mercurial doesn't do anything with
jerojasro@343 180 the copied-from file.
jerojasro@343 181 \interaction{daily.revert.copy}
jerojasro@343 182
jerojasro@343 183 \subsubsection{A slightly special case: reverting a rename}
jerojasro@343 184
jerojasro@343 185 If you \hgcmd{rename} a file, there is one small detail that
jerojasro@343 186 you should remember. When you \hgcmd{revert} a rename, it's not
jerojasro@343 187 enough to provide the name of the renamed-to file, as you can see
jerojasro@343 188 here.
jerojasro@343 189 \interaction{daily.revert.rename}
jerojasro@343 190 As you can see from the output of \hgcmd{status}, the renamed-to file
jerojasro@343 191 is no longer identified as added, but the renamed-\emph{from} file is
jerojasro@343 192 still removed! This is counter-intuitive (at least to me), but at
jerojasro@343 193 least it's easy to deal with.
jerojasro@343 194 \interaction{daily.revert.rename-orig}
jerojasro@343 195 So remember, to revert a \hgcmd{rename}, you must provide \emph{both}
jerojasro@343 196 the source and destination names.
jerojasro@343 197
jerojasro@343 198 % TODO: the output doesn't look like it will be removed!
jerojasro@343 199
jerojasro@343 200 (By the way, if you rename a file, then modify the renamed-to file,
jerojasro@343 201 then revert both components of the rename, when Mercurial restores the
jerojasro@343 202 file that was removed as part of the rename, it will be unmodified.
jerojasro@343 203 If you need the modifications in the renamed-to file to show up in the
jerojasro@343 204 renamed-from file, don't forget to copy them over.)
jerojasro@343 205
jerojasro@343 206 These fiddly aspects of reverting a rename arguably constitute a small
jerojasro@343 207 bug in Mercurial.
jerojasro@343 208
jerojasro@343 209 \section{Dealing with committed changes}
jerojasro@343 210
jerojasro@343 211 Consider a case where you have committed a change $a$, and another
jerojasro@343 212 change $b$ on top of it; you then realise that change $a$ was
jerojasro@343 213 incorrect. Mercurial lets you ``back out'' an entire changeset
jerojasro@343 214 automatically, and building blocks that let you reverse part of a
jerojasro@343 215 changeset by hand.
jerojasro@343 216
jerojasro@343 217 Before you read this section, here's something to keep in mind: the
jerojasro@343 218 \hgcmd{backout} command undoes changes by \emph{adding} history, not
jerojasro@343 219 by modifying or erasing it. It's the right tool to use if you're
jerojasro@343 220 fixing bugs, but not if you're trying to undo some change that has
jerojasro@343 221 catastrophic consequences. To deal with those, see
jerojasro@343 222 section~\ref{sec:undo:aaaiiieee}.
jerojasro@343 223
jerojasro@343 224 \subsection{Backing out a changeset}
jerojasro@343 225
jerojasro@343 226 The \hgcmd{backout} command lets you ``undo'' the effects of an entire
jerojasro@343 227 changeset in an automated fashion. Because Mercurial's history is
jerojasro@343 228 immutable, this command \emph{does not} get rid of the changeset you
jerojasro@343 229 want to undo. Instead, it creates a new changeset that
jerojasro@343 230 \emph{reverses} the effect of the to-be-undone changeset.
jerojasro@343 231
jerojasro@343 232 The operation of the \hgcmd{backout} command is a little intricate, so
jerojasro@343 233 let's illustrate it with some examples. First, we'll create a
jerojasro@343 234 repository with some simple changes.
jerojasro@343 235 \interaction{backout.init}
jerojasro@343 236
jerojasro@343 237 The \hgcmd{backout} command takes a single changeset ID as its
jerojasro@343 238 argument; this is the changeset to back out. Normally,
jerojasro@343 239 \hgcmd{backout} will drop you into a text editor to write a commit
jerojasro@343 240 message, so you can record why you're backing the change out. In this
jerojasro@343 241 example, we provide a commit message on the command line using the
jerojasro@343 242 \hgopt{backout}{-m} option.
jerojasro@343 243
jerojasro@343 244 \subsection{Backing out the tip changeset}
jerojasro@343 245
jerojasro@343 246 We're going to start by backing out the last changeset we committed.
jerojasro@343 247 \interaction{backout.simple}
jerojasro@343 248 You can see that the second line from \filename{myfile} is no longer
jerojasro@343 249 present. Taking a look at the output of \hgcmd{log} gives us an idea
jerojasro@343 250 of what the \hgcmd{backout} command has done.
jerojasro@343 251 \interaction{backout.simple.log}
jerojasro@343 252 Notice that the new changeset that \hgcmd{backout} has created is a
jerojasro@343 253 child of the changeset we backed out. It's easier to see this in
jerojasro@343 254 figure~\ref{fig:undo:backout}, which presents a graphical view of the
jerojasro@343 255 change history. As you can see, the history is nice and linear.
jerojasro@343 256
jerojasro@343 257 \begin{figure}[htb]
jerojasro@343 258 \centering
jerojasro@343 259 \grafix{undo-simple}
jerojasro@343 260 \caption{Backing out a change using the \hgcmd{backout} command}
jerojasro@343 261 \label{fig:undo:backout}
jerojasro@343 262 \end{figure}
jerojasro@343 263
jerojasro@343 264 \subsection{Backing out a non-tip change}
jerojasro@343 265
jerojasro@343 266 If you want to back out a change other than the last one you
jerojasro@343 267 committed, pass the \hgopt{backout}{--merge} option to the
jerojasro@343 268 \hgcmd{backout} command.
jerojasro@343 269 \interaction{backout.non-tip.clone}
jerojasro@343 270 This makes backing out any changeset a ``one-shot'' operation that's
jerojasro@343 271 usually simple and fast.
jerojasro@343 272 \interaction{backout.non-tip.backout}
jerojasro@343 273
jerojasro@343 274 If you take a look at the contents of \filename{myfile} after the
jerojasro@343 275 backout finishes, you'll see that the first and third changes are
jerojasro@343 276 present, but not the second.
jerojasro@343 277 \interaction{backout.non-tip.cat}
jerojasro@343 278
jerojasro@343 279 As the graphical history in figure~\ref{fig:undo:backout-non-tip}
jerojasro@343 280 illustrates, Mercurial actually commits \emph{two} changes in this
jerojasro@343 281 kind of situation (the box-shaped nodes are the ones that Mercurial
jerojasro@343 282 commits automatically). Before Mercurial begins the backout process,
jerojasro@343 283 it first remembers what the current parent of the working directory
jerojasro@343 284 is. It then backs out the target changeset, and commits that as a
jerojasro@343 285 changeset. Finally, it merges back to the previous parent of the
jerojasro@343 286 working directory, and commits the result of the merge.
jerojasro@343 287
jerojasro@343 288 % TODO: to me it looks like mercurial doesn't commit the second merge automatically!
jerojasro@343 289
jerojasro@343 290 \begin{figure}[htb]
jerojasro@343 291 \centering
jerojasro@343 292 \grafix{undo-non-tip}
jerojasro@343 293 \caption{Automated backout of a non-tip change using the \hgcmd{backout} command}
jerojasro@343 294 \label{fig:undo:backout-non-tip}
jerojasro@343 295 \end{figure}
jerojasro@343 296
jerojasro@343 297 The result is that you end up ``back where you were'', only with some
jerojasro@343 298 extra history that undoes the effect of the changeset you wanted to
jerojasro@343 299 back out.
jerojasro@343 300
jerojasro@343 301 \subsubsection{Always use the \hgopt{backout}{--merge} option}
jerojasro@343 302
jerojasro@343 303 In fact, since the \hgopt{backout}{--merge} option will do the ``right
jerojasro@343 304 thing'' whether or not the changeset you're backing out is the tip
jerojasro@343 305 (i.e.~it won't try to merge if it's backing out the tip, since there's
jerojasro@343 306 no need), you should \emph{always} use this option when you run the
jerojasro@343 307 \hgcmd{backout} command.
jerojasro@343 308
jerojasro@343 309 \subsection{Gaining more control of the backout process}
jerojasro@343 310
jerojasro@343 311 While I've recommended that you always use the
jerojasro@343 312 \hgopt{backout}{--merge} option when backing out a change, the
jerojasro@343 313 \hgcmd{backout} command lets you decide how to merge a backout
jerojasro@343 314 changeset. Taking control of the backout process by hand is something
jerojasro@343 315 you will rarely need to do, but it can be useful to understand what
jerojasro@343 316 the \hgcmd{backout} command is doing for you automatically. To
jerojasro@343 317 illustrate this, let's clone our first repository, but omit the
jerojasro@343 318 backout change that it contains.
jerojasro@343 319
jerojasro@343 320 \interaction{backout.manual.clone}
jerojasro@343 321 As with our earlier example, We'll commit a third changeset, then back
jerojasro@343 322 out its parent, and see what happens.
jerojasro@343 323 \interaction{backout.manual.backout}
jerojasro@343 324 Our new changeset is again a descendant of the changeset we backout
jerojasro@343 325 out; it's thus a new head, \emph{not} a descendant of the changeset
jerojasro@343 326 that was the tip. The \hgcmd{backout} command was quite explicit in
jerojasro@343 327 telling us this.
jerojasro@343 328 \interaction{backout.manual.log}
jerojasro@343 329
jerojasro@343 330 Again, it's easier to see what has happened by looking at a graph of
jerojasro@343 331 the revision history, in figure~\ref{fig:undo:backout-manual}. This
jerojasro@343 332 makes it clear that when we use \hgcmd{backout} to back out a change
jerojasro@343 333 other than the tip, Mercurial adds a new head to the repository (the
jerojasro@343 334 change it committed is box-shaped).
jerojasro@343 335
jerojasro@343 336 \begin{figure}[htb]
jerojasro@343 337 \centering
jerojasro@343 338 \grafix{undo-manual}
jerojasro@343 339 \caption{Backing out a change using the \hgcmd{backout} command}
jerojasro@343 340 \label{fig:undo:backout-manual}
jerojasro@343 341 \end{figure}
jerojasro@343 342
jerojasro@343 343 After the \hgcmd{backout} command has completed, it leaves the new
jerojasro@343 344 ``backout'' changeset as the parent of the working directory.
jerojasro@343 345 \interaction{backout.manual.parents}
jerojasro@343 346 Now we have two isolated sets of changes.
jerojasro@343 347 \interaction{backout.manual.heads}
jerojasro@343 348
jerojasro@343 349 Let's think about what we expect to see as the contents of
jerojasro@343 350 \filename{myfile} now. The first change should be present, because
jerojasro@343 351 we've never backed it out. The second change should be missing, as
jerojasro@343 352 that's the change we backed out. Since the history graph shows the
jerojasro@343 353 third change as a separate head, we \emph{don't} expect to see the
jerojasro@343 354 third change present in \filename{myfile}.
jerojasro@343 355 \interaction{backout.manual.cat}
jerojasro@343 356 To get the third change back into the file, we just do a normal merge
jerojasro@343 357 of our two heads.
jerojasro@343 358 \interaction{backout.manual.merge}
jerojasro@343 359 Afterwards, the graphical history of our repository looks like
jerojasro@343 360 figure~\ref{fig:undo:backout-manual-merge}.
jerojasro@343 361
jerojasro@343 362 \begin{figure}[htb]
jerojasro@343 363 \centering
jerojasro@343 364 \grafix{undo-manual-merge}
jerojasro@343 365 \caption{Manually merging a backout change}
jerojasro@343 366 \label{fig:undo:backout-manual-merge}
jerojasro@343 367 \end{figure}
jerojasro@343 368
jerojasro@343 369 \subsection{Why \hgcmd{backout} works as it does}
jerojasro@343 370
jerojasro@343 371 Here's a brief description of how the \hgcmd{backout} command works.
jerojasro@343 372 \begin{enumerate}
jerojasro@343 373 \item It ensures that the working directory is ``clean'', i.e.~that
jerojasro@343 374 the output of \hgcmd{status} would be empty.
jerojasro@343 375 \item It remembers the current parent of the working directory. Let's
jerojasro@343 376 call this changeset \texttt{orig}
jerojasro@343 377 \item It does the equivalent of a \hgcmd{update} to sync the working
jerojasro@343 378 directory to the changeset you want to back out. Let's call this
jerojasro@343 379 changeset \texttt{backout}
jerojasro@343 380 \item It finds the parent of that changeset. Let's call that
jerojasro@343 381 changeset \texttt{parent}.
jerojasro@343 382 \item For each file that the \texttt{backout} changeset affected, it
jerojasro@343 383 does the equivalent of a \hgcmdargs{revert}{-r parent} on that file,
jerojasro@343 384 to restore it to the contents it had before that changeset was
jerojasro@343 385 committed.
jerojasro@343 386 \item It commits the result as a new changeset. This changeset has
jerojasro@343 387 \texttt{backout} as its parent.
jerojasro@343 388 \item If you specify \hgopt{backout}{--merge} on the command line, it
jerojasro@343 389 merges with \texttt{orig}, and commits the result of the merge.
jerojasro@343 390 \end{enumerate}
jerojasro@343 391
jerojasro@343 392 An alternative way to implement the \hgcmd{backout} command would be
jerojasro@343 393 to \hgcmd{export} the to-be-backed-out changeset as a diff, then use
jerojasro@343 394 the \cmdopt{patch}{--reverse} option to the \command{patch} command to
jerojasro@343 395 reverse the effect of the change without fiddling with the working
jerojasro@343 396 directory. This sounds much simpler, but it would not work nearly as
jerojasro@343 397 well.
jerojasro@343 398
jerojasro@343 399 The reason that \hgcmd{backout} does an update, a commit, a merge, and
jerojasro@343 400 another commit is to give the merge machinery the best chance to do a
jerojasro@343 401 good job when dealing with all the changes \emph{between} the change
jerojasro@343 402 you're backing out and the current tip.
jerojasro@343 403
jerojasro@343 404 If you're backing out a changeset that's~100 revisions back in your
jerojasro@343 405 project's history, the chances that the \command{patch} command will
jerojasro@343 406 be able to apply a reverse diff cleanly are not good, because
jerojasro@343 407 intervening changes are likely to have ``broken the context'' that
jerojasro@343 408 \command{patch} uses to determine whether it can apply a patch (if
jerojasro@343 409 this sounds like gibberish, see \ref{sec:mq:patch} for a
jerojasro@343 410 discussion of the \command{patch} command). Also, Mercurial's merge
jerojasro@343 411 machinery will handle files and directories being renamed, permission
jerojasro@343 412 changes, and modifications to binary files, none of which
jerojasro@343 413 \command{patch} can deal with.
jerojasro@343 414
jerojasro@343 415 \section{Changes that should never have been}
jerojasro@343 416 \label{sec:undo:aaaiiieee}
jerojasro@343 417
jerojasro@343 418 Most of the time, the \hgcmd{backout} command is exactly what you need
jerojasro@343 419 if you want to undo the effects of a change. It leaves a permanent
jerojasro@343 420 record of exactly what you did, both when committing the original
jerojasro@343 421 changeset and when you cleaned up after it.
jerojasro@343 422
jerojasro@343 423 On rare occasions, though, you may find that you've committed a change
jerojasro@343 424 that really should not be present in the repository at all. For
jerojasro@343 425 example, it would be very unusual, and usually considered a mistake,
jerojasro@343 426 to commit a software project's object files as well as its source
jerojasro@343 427 files. Object files have almost no intrinsic value, and they're
jerojasro@343 428 \emph{big}, so they increase the size of the repository and the amount
jerojasro@343 429 of time it takes to clone or pull changes.
jerojasro@343 430
jerojasro@343 431 Before I discuss the options that you have if you commit a ``brown
jerojasro@343 432 paper bag'' change (the kind that's so bad that you want to pull a
jerojasro@343 433 brown paper bag over your head), let me first discuss some approaches
jerojasro@343 434 that probably won't work.
jerojasro@343 435
jerojasro@343 436 Since Mercurial treats history as accumulative---every change builds
jerojasro@343 437 on top of all changes that preceded it---you generally can't just make
jerojasro@343 438 disastrous changes disappear. The one exception is when you've just
jerojasro@343 439 committed a change, and it hasn't been pushed or pulled into another
jerojasro@343 440 repository. That's when you can safely use the \hgcmd{rollback}
jerojasro@343 441 command, as I detailed in section~\ref{sec:undo:rollback}.
jerojasro@343 442
jerojasro@343 443 After you've pushed a bad change to another repository, you
jerojasro@343 444 \emph{could} still use \hgcmd{rollback} to make your local copy of the
jerojasro@343 445 change disappear, but it won't have the consequences you want. The
jerojasro@343 446 change will still be present in the remote repository, so it will
jerojasro@343 447 reappear in your local repository the next time you pull.
jerojasro@343 448
jerojasro@343 449 If a situation like this arises, and you know which repositories your
jerojasro@343 450 bad change has propagated into, you can \emph{try} to get rid of the
jerojasro@343 451 changeefrom \emph{every} one of those repositories. This is, of
jerojasro@343 452 course, not a satisfactory solution: if you miss even a single
jerojasro@343 453 repository while you're expunging, the change is still ``in the
jerojasro@343 454 wild'', and could propagate further.
jerojasro@343 455
jerojasro@343 456 If you've committed one or more changes \emph{after} the change that
jerojasro@343 457 you'd like to see disappear, your options are further reduced.
jerojasro@343 458 Mercurial doesn't provide a way to ``punch a hole'' in history,
jerojasro@343 459 leaving changesets intact.
jerojasro@343 460
jerojasro@343 461 XXX This needs filling out. The \texttt{hg-replay} script in the
jerojasro@343 462 \texttt{examples} directory works, but doesn't handle merge
jerojasro@343 463 changesets. Kind of an important omission.
jerojasro@343 464
jerojasro@343 465 \subsection{Protect yourself from ``escaped'' changes}
jerojasro@343 466
jerojasro@343 467 If you've committed some changes to your local repository and they've
jerojasro@343 468 been pushed or pulled somewhere else, this isn't necessarily a
jerojasro@343 469 disaster. You can protect yourself ahead of time against some classes
jerojasro@343 470 of bad changeset. This is particularly easy if your team usually
jerojasro@343 471 pulls changes from a central repository.
jerojasro@343 472
jerojasro@343 473 By configuring some hooks on that repository to validate incoming
jerojasro@343 474 changesets (see chapter~\ref{chap:hook}), you can automatically
jerojasro@343 475 prevent some kinds of bad changeset from being pushed to the central
jerojasro@343 476 repository at all. With such a configuration in place, some kinds of
jerojasro@343 477 bad changeset will naturally tend to ``die out'' because they can't
jerojasro@343 478 propagate into the central repository. Better yet, this happens
jerojasro@343 479 without any need for explicit intervention.
jerojasro@343 480
jerojasro@343 481 For instance, an incoming change hook that verifies that a changeset
jerojasro@343 482 will actually compile can prevent people from inadvertantly ``breaking
jerojasro@343 483 the build''.
jerojasro@343 484
jerojasro@343 485 \section{Finding the source of a bug}
jerojasro@343 486 \label{sec:undo:bisect}
jerojasro@343 487
jerojasro@343 488 While it's all very well to be able to back out a changeset that
jerojasro@343 489 introduced a bug, this requires that you know which changeset to back
jerojasro@343 490 out. Mercurial provides an invaluable command, called
jerojasro@343 491 \hgcmd{bisect}, that helps you to automate this process and accomplish
jerojasro@343 492 it very efficiently.
jerojasro@343 493
jerojasro@343 494 The idea behind the \hgcmd{bisect} command is that a changeset has
jerojasro@343 495 introduced some change of behaviour that you can identify with a
jerojasro@343 496 simple binary test. You don't know which piece of code introduced the
jerojasro@343 497 change, but you know how to test for the presence of the bug. The
jerojasro@343 498 \hgcmd{bisect} command uses your test to direct its search for the
jerojasro@343 499 changeset that introduced the code that caused the bug.
jerojasro@343 500
jerojasro@343 501 Here are a few scenarios to help you understand how you might apply
jerojasro@343 502 this command.
jerojasro@343 503 \begin{itemize}
jerojasro@343 504 \item The most recent version of your software has a bug that you
jerojasro@343 505 remember wasn't present a few weeks ago, but you don't know when it
jerojasro@343 506 was introduced. Here, your binary test checks for the presence of
jerojasro@343 507 that bug.
jerojasro@343 508 \item You fixed a bug in a rush, and now it's time to close the entry
jerojasro@343 509 in your team's bug database. The bug database requires a changeset
jerojasro@343 510 ID when you close an entry, but you don't remember which changeset
jerojasro@343 511 you fixed the bug in. Once again, your binary test checks for the
jerojasro@343 512 presence of the bug.
jerojasro@343 513 \item Your software works correctly, but runs~15\% slower than the
jerojasro@343 514 last time you measured it. You want to know which changeset
jerojasro@343 515 introduced the performance regression. In this case, your binary
jerojasro@343 516 test measures the performance of your software, to see whether it's
jerojasro@343 517 ``fast'' or ``slow''.
jerojasro@343 518 \item The sizes of the components of your project that you ship
jerojasro@343 519 exploded recently, and you suspect that something changed in the way
jerojasro@343 520 you build your project.
jerojasro@343 521 \end{itemize}
jerojasro@343 522
jerojasro@343 523 From these examples, it should be clear that the \hgcmd{bisect}
jerojasro@343 524 command is not useful only for finding the sources of bugs. You can
jerojasro@343 525 use it to find any ``emergent property'' of a repository (anything
jerojasro@343 526 that you can't find from a simple text search of the files in the
jerojasro@343 527 tree) for which you can write a binary test.
jerojasro@343 528
jerojasro@343 529 We'll introduce a little bit of terminology here, just to make it
jerojasro@343 530 clear which parts of the search process are your responsibility, and
jerojasro@343 531 which are Mercurial's. A \emph{test} is something that \emph{you} run
jerojasro@343 532 when \hgcmd{bisect} chooses a changeset. A \emph{probe} is what
jerojasro@343 533 \hgcmd{bisect} runs to tell whether a revision is good. Finally,
jerojasro@343 534 we'll use the word ``bisect'', as both a noun and a verb, to stand in
jerojasro@343 535 for the phrase ``search using the \hgcmd{bisect} command.
jerojasro@343 536
jerojasro@343 537 One simple way to automate the searching process would be simply to
jerojasro@343 538 probe every changeset. However, this scales poorly. If it took ten
jerojasro@343 539 minutes to test a single changeset, and you had 10,000 changesets in
jerojasro@343 540 your repository, the exhaustive approach would take on average~35
jerojasro@343 541 \emph{days} to find the changeset that introduced a bug. Even if you
jerojasro@343 542 knew that the bug was introduced by one of the last 500 changesets,
jerojasro@343 543 and limited your search to those, you'd still be looking at over 40
jerojasro@343 544 hours to find the changeset that introduced your bug.
jerojasro@343 545
jerojasro@343 546 What the \hgcmd{bisect} command does is use its knowledge of the
jerojasro@343 547 ``shape'' of your project's revision history to perform a search in
jerojasro@343 548 time proportional to the \emph{logarithm} of the number of changesets
jerojasro@343 549 to check (the kind of search it performs is called a dichotomic
jerojasro@343 550 search). With this approach, searching through 10,000 changesets will
jerojasro@343 551 take less than three hours, even at ten minutes per test (the search
jerojasro@343 552 will require about 14 tests). Limit your search to the last hundred
jerojasro@343 553 changesets, and it will take only about an hour (roughly seven tests).
jerojasro@343 554
jerojasro@343 555 The \hgcmd{bisect} command is aware of the ``branchy'' nature of a
jerojasro@343 556 Mercurial project's revision history, so it has no problems dealing
jerojasro@343 557 with branches, merges, or multiple heads in a repoository. It can
jerojasro@343 558 prune entire branches of history with a single probe, which is how it
jerojasro@343 559 operates so efficiently.
jerojasro@343 560
jerojasro@343 561 \subsection{Using the \hgcmd{bisect} command}
jerojasro@343 562
jerojasro@343 563 Here's an example of \hgcmd{bisect} in action.
jerojasro@343 564
jerojasro@343 565 \begin{note}
jerojasro@343 566 In versions 0.9.5 and earlier of Mercurial, \hgcmd{bisect} was not a
jerojasro@343 567 core command: it was distributed with Mercurial as an extension.
jerojasro@343 568 This section describes the built-in command, not the old extension.
jerojasro@343 569 \end{note}
jerojasro@343 570
jerojasro@343 571 Now let's create a repository, so that we can try out the
jerojasro@343 572 \hgcmd{bisect} command in isolation.
jerojasro@343 573 \interaction{bisect.init}
jerojasro@343 574 We'll simulate a project that has a bug in it in a simple-minded way:
jerojasro@343 575 create trivial changes in a loop, and nominate one specific change
jerojasro@343 576 that will have the ``bug''. This loop creates 35 changesets, each
jerojasro@343 577 adding a single file to the repository. We'll represent our ``bug''
jerojasro@343 578 with a file that contains the text ``i have a gub''.
jerojasro@343 579 \interaction{bisect.commits}
jerojasro@343 580
jerojasro@343 581 The next thing that we'd like to do is figure out how to use the
jerojasro@343 582 \hgcmd{bisect} command. We can use Mercurial's normal built-in help
jerojasro@343 583 mechanism for this.
jerojasro@343 584 \interaction{bisect.help}
jerojasro@343 585
jerojasro@343 586 The \hgcmd{bisect} command works in steps. Each step proceeds as follows.
jerojasro@343 587 \begin{enumerate}
jerojasro@343 588 \item You run your binary test.
jerojasro@343 589 \begin{itemize}
jerojasro@343 590 \item If the test succeeded, you tell \hgcmd{bisect} by running the
jerojasro@343 591 \hgcmdargs{bisect}{good} command.
jerojasro@343 592 \item If it failed, run the \hgcmdargs{bisect}{--bad} command.
jerojasro@343 593 \end{itemize}
jerojasro@343 594 \item The command uses your information to decide which changeset to
jerojasro@343 595 test next.
jerojasro@343 596 \item It updates the working directory to that changeset, and the
jerojasro@343 597 process begins again.
jerojasro@343 598 \end{enumerate}
jerojasro@343 599 The process ends when \hgcmd{bisect} identifies a unique changeset
jerojasro@343 600 that marks the point where your test transitioned from ``succeeding''
jerojasro@343 601 to ``failing''.
jerojasro@343 602
jerojasro@343 603 To start the search, we must run the \hgcmdargs{bisect}{--reset} command.
jerojasro@343 604 \interaction{bisect.search.init}
jerojasro@343 605
jerojasro@343 606 In our case, the binary test we use is simple: we check to see if any
jerojasro@343 607 file in the repository contains the string ``i have a gub''. If it
jerojasro@343 608 does, this changeset contains the change that ``caused the bug''. By
jerojasro@343 609 convention, a changeset that has the property we're searching for is
jerojasro@343 610 ``bad'', while one that doesn't is ``good''.
jerojasro@343 611
jerojasro@343 612 Most of the time, the revision to which the working directory is
jerojasro@343 613 synced (usually the tip) already exhibits the problem introduced by
jerojasro@343 614 the buggy change, so we'll mark it as ``bad''.
jerojasro@343 615 \interaction{bisect.search.bad-init}
jerojasro@343 616
jerojasro@343 617 Our next task is to nominate a changeset that we know \emph{doesn't}
jerojasro@343 618 have the bug; the \hgcmd{bisect} command will ``bracket'' its search
jerojasro@343 619 between the first pair of good and bad changesets. In our case, we
jerojasro@343 620 know that revision~10 didn't have the bug. (I'll have more words
jerojasro@343 621 about choosing the first ``good'' changeset later.)
jerojasro@343 622 \interaction{bisect.search.good-init}
jerojasro@343 623
jerojasro@343 624 Notice that this command printed some output.
jerojasro@343 625 \begin{itemize}
jerojasro@343 626 \item It told us how many changesets it must consider before it can
jerojasro@343 627 identify the one that introduced the bug, and how many tests that
jerojasro@343 628 will require.
jerojasro@343 629 \item It updated the working directory to the next changeset to test,
jerojasro@343 630 and told us which changeset it's testing.
jerojasro@343 631 \end{itemize}
jerojasro@343 632
jerojasro@343 633 We now run our test in the working directory. We use the
jerojasro@343 634 \command{grep} command to see if our ``bad'' file is present in the
jerojasro@343 635 working directory. If it is, this revision is bad; if not, this
jerojasro@343 636 revision is good.
jerojasro@343 637 \interaction{bisect.search.step1}
jerojasro@343 638
jerojasro@343 639 This test looks like a perfect candidate for automation, so let's turn
jerojasro@343 640 it into a shell function.
jerojasro@343 641 \interaction{bisect.search.mytest}
jerojasro@343 642 We can now run an entire test step with a single command,
jerojasro@343 643 \texttt{mytest}.
jerojasro@343 644 \interaction{bisect.search.step2}
jerojasro@343 645 A few more invocations of our canned test step command, and we're
jerojasro@343 646 done.
jerojasro@343 647 \interaction{bisect.search.rest}
jerojasro@343 648
jerojasro@343 649 Even though we had~40 changesets to search through, the \hgcmd{bisect}
jerojasro@343 650 command let us find the changeset that introduced our ``bug'' with
jerojasro@343 651 only five tests. Because the number of tests that the \hgcmd{bisect}
jerojasro@343 652 command performs grows logarithmically with the number of changesets to
jerojasro@343 653 search, the advantage that it has over the ``brute force'' search
jerojasro@343 654 approach increases with every changeset you add.
jerojasro@343 655
jerojasro@343 656 \subsection{Cleaning up after your search}
jerojasro@343 657
jerojasro@343 658 When you're finished using the \hgcmd{bisect} command in a
jerojasro@343 659 repository, you can use the \hgcmdargs{bisect}{reset} command to drop
jerojasro@343 660 the information it was using to drive your search. The command
jerojasro@343 661 doesn't use much space, so it doesn't matter if you forget to run this
jerojasro@343 662 command. However, \hgcmd{bisect} won't let you start a new search in
jerojasro@343 663 that repository until you do a \hgcmdargs{bisect}{reset}.
jerojasro@343 664 \interaction{bisect.search.reset}
jerojasro@343 665
jerojasro@343 666 \section{Tips for finding bugs effectively}
jerojasro@343 667
jerojasro@343 668 \subsection{Give consistent input}
jerojasro@343 669
jerojasro@343 670 The \hgcmd{bisect} command requires that you correctly report the
jerojasro@343 671 result of every test you perform. If you tell it that a test failed
jerojasro@343 672 when it really succeeded, it \emph{might} be able to detect the
jerojasro@343 673 inconsistency. If it can identify an inconsistency in your reports,
jerojasro@343 674 it will tell you that a particular changeset is both good and bad.
jerojasro@343 675 However, it can't do this perfectly; it's about as likely to report
jerojasro@343 676 the wrong changeset as the source of the bug.
jerojasro@343 677
jerojasro@343 678 \subsection{Automate as much as possible}
jerojasro@343 679
jerojasro@343 680 When I started using the \hgcmd{bisect} command, I tried a few times
jerojasro@343 681 to run my tests by hand, on the command line. This is an approach
jerojasro@343 682 that I, at least, am not suited to. After a few tries, I found that I
jerojasro@343 683 was making enough mistakes that I was having to restart my searches
jerojasro@343 684 several times before finally getting correct results.
jerojasro@343 685
jerojasro@343 686 My initial problems with driving the \hgcmd{bisect} command by hand
jerojasro@343 687 occurred even with simple searches on small repositories; if the
jerojasro@343 688 problem you're looking for is more subtle, or the number of tests that
jerojasro@343 689 \hgcmd{bisect} must perform increases, the likelihood of operator
jerojasro@343 690 error ruining the search is much higher. Once I started automating my
jerojasro@343 691 tests, I had much better results.
jerojasro@343 692
jerojasro@343 693 The key to automated testing is twofold:
jerojasro@343 694 \begin{itemize}
jerojasro@343 695 \item always test for the same symptom, and
jerojasro@343 696 \item always feed consistent input to the \hgcmd{bisect} command.
jerojasro@343 697 \end{itemize}
jerojasro@343 698 In my tutorial example above, the \command{grep} command tests for the
jerojasro@343 699 symptom, and the \texttt{if} statement takes the result of this check
jerojasro@343 700 and ensures that we always feed the same input to the \hgcmd{bisect}
jerojasro@343 701 command. The \texttt{mytest} function marries these together in a
jerojasro@343 702 reproducible way, so that every test is uniform and consistent.
jerojasro@343 703
jerojasro@343 704 \subsection{Check your results}
jerojasro@343 705
jerojasro@343 706 Because the output of a \hgcmd{bisect} search is only as good as the
jerojasro@343 707 input you give it, don't take the changeset it reports as the
jerojasro@343 708 absolute truth. A simple way to cross-check its report is to manually
jerojasro@343 709 run your test at each of the following changesets:
jerojasro@343 710 \begin{itemize}
jerojasro@343 711 \item The changeset that it reports as the first bad revision. Your
jerojasro@343 712 test should still report this as bad.
jerojasro@343 713 \item The parent of that changeset (either parent, if it's a merge).
jerojasro@343 714 Your test should report this changeset as good.
jerojasro@343 715 \item A child of that changeset. Your test should report this
jerojasro@343 716 changeset as bad.
jerojasro@343 717 \end{itemize}
jerojasro@343 718
jerojasro@343 719 \subsection{Beware interference between bugs}
jerojasro@343 720
jerojasro@343 721 It's possible that your search for one bug could be disrupted by the
jerojasro@343 722 presence of another. For example, let's say your software crashes at
jerojasro@343 723 revision 100, and worked correctly at revision 50. Unknown to you,
jerojasro@343 724 someone else introduced a different crashing bug at revision 60, and
jerojasro@343 725 fixed it at revision 80. This could distort your results in one of
jerojasro@343 726 several ways.
jerojasro@343 727
jerojasro@343 728 It is possible that this other bug completely ``masks'' yours, which
jerojasro@343 729 is to say that it occurs before your bug has a chance to manifest
jerojasro@343 730 itself. If you can't avoid that other bug (for example, it prevents
jerojasro@343 731 your project from building), and so can't tell whether your bug is
jerojasro@343 732 present in a particular changeset, the \hgcmd{bisect} command cannot
jerojasro@343 733 help you directly. Instead, you can mark a changeset as untested by
jerojasro@343 734 running \hgcmdargs{bisect}{--skip}.
jerojasro@343 735
jerojasro@343 736 A different problem could arise if your test for a bug's presence is
jerojasro@343 737 not specific enough. If you check for ``my program crashes'', then
jerojasro@343 738 both your crashing bug and an unrelated crashing bug that masks it
jerojasro@343 739 will look like the same thing, and mislead \hgcmd{bisect}.
jerojasro@343 740
jerojasro@343 741 Another useful situation in which to use \hgcmdargs{bisect}{--skip} is
jerojasro@343 742 if you can't test a revision because your project was in a broken and
jerojasro@343 743 hence untestable state at that revision, perhaps because someone
jerojasro@343 744 checked in a change that prevented the project from building.
jerojasro@343 745
jerojasro@343 746 \subsection{Bracket your search lazily}
jerojasro@343 747
jerojasro@343 748 Choosing the first ``good'' and ``bad'' changesets that will mark the
jerojasro@343 749 end points of your search is often easy, but it bears a little
jerojasro@343 750 discussion nevertheless. From the perspective of \hgcmd{bisect}, the
jerojasro@343 751 ``newest'' changeset is conventionally ``bad'', and the older
jerojasro@343 752 changeset is ``good''.
jerojasro@343 753
jerojasro@343 754 If you're having trouble remembering when a suitable ``good'' change
jerojasro@343 755 was, so that you can tell \hgcmd{bisect}, you could do worse than
jerojasro@343 756 testing changesets at random. Just remember to eliminate contenders
jerojasro@343 757 that can't possibly exhibit the bug (perhaps because the feature with
jerojasro@343 758 the bug isn't present yet) and those where another problem masks the
jerojasro@343 759 bug (as I discussed above).
jerojasro@343 760
jerojasro@343 761 Even if you end up ``early'' by thousands of changesets or months of
jerojasro@343 762 history, you will only add a handful of tests to the total number that
jerojasro@343 763 \hgcmd{bisect} must perform, thanks to its logarithmic behaviour.
jerojasro@343 764
jerojasro@343 765 %%% Local Variables:
jerojasro@343 766 %%% mode: latex
jerojasro@343 767 %%% TeX-master: "00book"
jerojasro@343 768 %%% End: