C. Reference manual for Administrative files

Inside the repository, in the directory `$CVSROOT/CVSROOT', there are a number of supportive files for CVS. You can use CVS in a limited fashion without any of them, but if they are set up properly they can help make life easier. For a discussion of how to edit them, see The administrative files.

The most important of these files is the `modules' file, which defines the modules inside the repository.

C.1 The modules file

The `modules' file records your definitions of names for collections of source code. CVS will use these definitions if you use CVS to update the modules file (use normal commands like add, commit, etc).

The `modules' file may contain blank lines and comments (lines beginning with `#') as well as module definitions. Long lines can be continued on the next line by specifying a backslash (`\') as the last character on the line.

There are three basic types of modules: alias modules, regular modules, and ampersand modules. The difference between them is the way that they map files in the repository to files in the working directory. In all of the following examples, the top-level repository contains a directory called `first-dir', which contains two files, `file1' and `file2', and a directory `sdir'. `first-dir/sdir' contains a file `sfile'.

C.1.1 Alias modules

Alias modules are the simplest kind of module:

mname -a aliases…

This represents the simplest way of defining a module mname. The `-a' flags the definition as a simple alias: CVS will treat any use of mname (as a command argument) as if the list of names aliases had been specified instead. aliases may contain either other module names or paths. When you use paths in aliases, checkout creates all intermediate directories in the working directory, just as if the path had been specified explicitly in the CVS arguments.

For example, if the modules file contains:

amodule -a first-dir

then the following two commands are equivalent:

$ cvs co amodule
$ cvs co first-dir

and they each would provide output such as:

cvs checkout: Updating first-dir
U first-dir/file1
U first-dir/file2
cvs checkout: Updating first-dir/sdir
U first-dir/sdir/sfile

C.1.2 Regular modules

mname [ options ] dir [ files… ]

In the simplest case, this form of module definition reduces to `mname dir'. This defines all the files in directory dir as module mname. dir is a relative path (from $CVSROOT) to a directory of source in the source repository. In this case, on checkout, a single directory called mname is created as a working directory; no intermediate directory levels are used by default, even if dir was a path involving several directory levels.

For example, if a module is defined by:

regmodule first-dir

then regmodule will contain the files from first-dir:

$ cvs co regmodule
cvs checkout: Updating regmodule
U regmodule/file1
U regmodule/file2
cvs checkout: Updating regmodule/sdir
U regmodule/sdir/sfile
$

By explicitly specifying files in the module definition after dir, you can select particular files from directory dir. Here is an example:

regfiles first-dir/sdir sfile

With this definition, getting the regfiles module will create a single working directory `regfiles' containing the file listed, which comes from a directory deeper in the CVS source repository:

$ cvs co regfiles
U regfiles/sfile
$

C.1.3 Ampersand modules

A module definition can refer to other modules by including `&module' in its definition.

mname [ options ] &module…

Then getting the module creates a subdirectory for each such module, in the directory containing the module. For example, if modules contains

ampermod &first-dir

then a checkout will create an ampermod directory which contains a directory called first-dir, which in turns contains all the directories and files which live there. For example, the command

$ cvs co ampermod

will create the following files:

ampermod/first-dir/file1
ampermod/first-dir/file2
ampermod/first-dir/sdir/sfile

There is one quirk/bug: the messages that CVS prints omit the `ampermod', and thus do not correctly display the location to which it is checking out the files:

$ cvs co ampermod
cvs checkout: Updating first-dir
U first-dir/file1
U first-dir/file2
cvs checkout: Updating first-dir/sdir
U first-dir/sdir/sfile
$

Do not rely on this buggy behavior; it may get fixed in a future release of CVS.

C.1.4 Excluding directories

An alias module may exclude particular directories from other modules by using an exclamation mark (`!') before the name of each directory to be excluded.

For example, if the modules file contains:

exmodule -a !first-dir/sdir first-dir

then checking out the module `exmodule' will check out everything in `first-dir' except any files in the subdirectory `first-dir/sdir'.

C.1.5 Module options

Either regular modules or ampersand modules can contain options, which supply additional information concerning the module.

-d name

Name the working directory something other than the module name.

-e prog

Specify a program prog to run whenever files in a module are exported. prog runs with a single argument, the module name.

-o prog

Specify a program prog to run whenever files in a module are checked out. prog runs with a single argument, the module name. See How the modules file “program options” programs are run for information on how prog is called.

-s status

Assign a status to the module. When the module file is printed with `cvs checkout -s' the modules are sorted according to primarily module status, and secondarily according to the module name. This option has no other meaning. You can use this option for several things besides status: for instance, list the person that is responsible for this module.

-t prog

Specify a program prog to run whenever files in a module are tagged with rtag. prog runs with two arguments: the module name and the symbolic tag specified to rtag. It is not run when tag is executed. Generally you will find that the `taginfo' file is a better solution (see section Taginfo).

You should also see see section How the modules file “program options” programs are run about how the “program options” programs are run.

C.1.6 How the modules file “program options” programs are run

For checkout, rtag, and export, the program is server-based, and as such the following applies:-

If using remote access methods (pserver, ext, etc.), CVS will execute this program on the server from a temporary directory. The path is searched for this program.

If using “local access” (on a local or remote NFS file system, i.e. repository set just to a path), the program will be executed from the newly checked-out tree, if found there, or alternatively searched for in the path if not.

The programs are all run after the operation has effectively completed.

C.2 The cvswrappers file

Wrappers refers to a CVS feature which lets you control certain settings based on the name of the file which is being operated on. The settings are `-k' for binary files, and `-m' for nonmergeable text files.

The `-m' option specifies the merge methodology that should be used when a non-binary file is updated. MERGE means the usual CVS behavior: try to merge the files. COPY means that cvs update will refuse to merge files, as it also does for files specified as binary with `-kb' (but if the file is specified as binary, there is no need to specify `-m 'COPY''). CVS will provide the user with the two versions of the files, and require the user using mechanisms outside CVS, to insert any necessary changes.

WARNING: do not use COPY with CVS 1.9 or earlier - such versions of CVS will copy one version of your file over the other, wiping out the previous contents. The `-m' wrapper option only affects behavior when merging is done on update; it does not affect how files are stored. See Handling binary files, for more on binary files.

The basic format of the file `cvswrappers' is:

wildcard     [option value][option value]...

where option is one of
-m           update methodology      value: MERGE or COPY
-k           keyword expansion       value: expansion mode

and value is a single-quote delimited value.

For example, the following command imports a directory, treating files whose name ends in `.exe' as binary:

cvs import -I ! -W "*.exe -k 'b'" first-dir vendortag reltag

C.3 The Trigger Scripts

Several of the administrative files support triggers, or the launching external scripts or programs at specific times before or after particular events. The individual files are discussed in the later sections, The commit support files and Taginfo, but some of the common elements are discussed here.

All the trigger scripts are launched in a copy of the user sandbox being committed, on the server, in client-server mode. In local mode, the scripts are actually launched directly from the user sandbox directory being committed. For most intents and purposes, the same scripts can be run in both locations without alteration.

C.3.1 The common syntax

The administrative files such as `commitinfo', `loginfo', `rcsinfo', `verifymsg', etc., all have a common format. The purpose of the files are described later on. The common syntax is described here.

Each line contains the following:

• A regular expression. This is a basic regular expression in the syntax used by GNU emacs.
• A whitespace separator—one or more spaces and/or tabs.
• A file name or command-line template.

Blank lines are ignored. Lines that start with the character `#' are treated as comments. Long lines unfortunately can not be broken in two parts in any way.

The first regular expression that matches the current directory name in the repository is used. The rest of the line is used as a file name or command-line as appropriate.

C.3.2 Security and the Trigger Scripts

Security is a huge subject, and implementing a secure system is a non-trivial task. This section will barely touch on all the issues involved, but it is well to note that, as with any script you will be allowing an untrusted user to run on your server, there are measures you can take to help prevent your trigger scripts from being abused.

For instance, since the CVS trigger scripts all run in a copy of the user's sandbox on the server, a naively coded Perl trigger script which attempts to use a Perl module that is not installed on the system can be hijacked by any user with commit access who is checking in a file with the correct name. Other scripting languages may be vulnerable to similar hacks.

One way to make a script more secure, at least with Perl, is to use scripts which invoke the -T, or "taint-check" switch on their #! line. In the most basic terms, this causes Perl to avoid running code that may have come from an external source. Please run the perldoc perlsec command for more on Perl security. Again, other languages may implement other security verification hooks which look more or less like Perl's "taint-check" mechanism.

C.4 The commit support files

There are three kinds of trigger scripts (see section The Trigger Scripts) that can be run at various times during a commit. They are specified in files in the repository, as described below. The following table summarizes the file names and the purpose of the corresponding programs.

`commitinfo'

The program is responsible for checking that the commit is allowed. If it exits with a non-zero exit status the commit will be aborted.

`verifymsg'

The specified program is used to evaluate the log message, and possibly verify that it contains all required fields. This is most useful in combination with the `rcsinfo' file, which can hold a log message template (see section Rcsinfo).

`editinfo'

The specified program is used to edit the log message, and possibly verify that it contains all required fields. This is most useful in combination with the `rcsinfo' file, which can hold a log message template (see section Rcsinfo). (obsolete)

`loginfo'

The specified program is called when the commit is complete. It receives the log message and some additional information and can store the log message in a file, or mail it to appropriate persons, or maybe post it to a local newsgroup, or… Your imagination is the limit!

C.4.1 Commitinfo

The `commitinfo' file defines programs to execute whenever `cvs commit' is about to execute. These programs are used for pre-commit checking to verify that the modified, added and removed files are really ready to be committed. This could be used, for instance, to verify that the changed files conform to to your site's standards for coding practice.

As mentioned earlier, each line in the `commitinfo' file consists of a regular expression and a command-line template. The template can include a program name and any number of arguments you wish to supply to it. The full path to the current source repository is appended to the template, followed by the file names of any files involved in the commit (added, removed, and modified files).

The first line with a regular expression matching the directory within the repository will be used. If the command returns a non-zero exit status the commit will be aborted.

If the repository name does not match any of the regular expressions in this file, the `DEFAULT' line is used, if it is specified.

All occurrences of the name `ALL' appearing as a regular expression are used in addition to the first matching regular expression or the name `DEFAULT'.

The command will be run in the root of the workspace containing the new versions of any files the user would like to modify (commit), or in a copy of the workspace on the server (see section Remote repositories). If a file is being removed, there will be no copy of the file under the current directory. If a file is being added, there will be no corresponding archive file in the repository unless the file is being resurrected.

Note that both the repository directory and the corresponding Attic (see section The attic) directory may need to be checked to locate the archive file corresponding to any given file being committed. Much of the information about the specific commit request being made, including the destination branch, commit message, and command line options specified, is not available to the command.

C.4.2 Verifying log messages

Once you have entered a log message, you can evaluate that message to check for specific content, such as a bug ID. Use the `verifymsg' file to specify a program that is used to verify the log message. This program could be a simple script that checks that the entered message contains the required fields.

The `verifymsg' file is often most useful together with the `rcsinfo' file, which can be used to specify a log message template.

Each line in the `verifymsg' file consists of a regular expression and a command-line template. The template must include a program name, and can include any number of arguments. The full path to the current log message template file is appended to the template.

One thing that should be noted is that the `ALL' keyword is not supported. If more than one matching line is found, the first one is used. This can be useful for specifying a default verification script in a directory, and then overriding it in a subdirectory.

If the repository name does not match any of the regular expressions in this file, the `DEFAULT' line is used, if it is specified.

If the verification script exits with a non-zero exit status, the commit is aborted.

In the default configuration, CVS allows the verification script to change the log message. This is controlled via the RereadLogAfterVerify CVSROOT/config option.

When `RereadLogAfterVerify=always' or `RereadLogAfterVerify=stat', the log message will either always be reread after the verification script is run or reread only if the log message file status has changed.

See section The CVSROOT/config configuration file, for more on CVSROOT/config options.

It is NOT a good idea for a `verifymsg' script to interact directly with the user in the various client/server methods. For the pserver method, there is no protocol support for communicating between `verifymsg' and the client on the remote end. For the ext and server methods, it is possible for CVS to become confused by the characters going along the same channel as the CVS protocol messages. See Remote repositories, for more information on client/server setups. In addition, at the time the `verifymsg' script runs, the CVS server has locks in place in the repository. If control is returned to the user here then other users may be stuck waiting for access to the repository.

This option can be useful if you find yourself using an rcstemplate that needs to be modified to remove empty elements or to fill in default values. It can also be useful if the rcstemplate has changed in the repository and the CVS/Template was not updated, but is able to be adapted to the new format by the verification script that is run by `verifymsg'.

An example of an update might be to change all occurrences of 'BugId:' to be 'DefectId:' (which can be useful if the rcstemplate has recently been changed and there are still checked-out user trees with cached copies in the CVS/Template file of the older version).

Another example of an update might be to delete a line that contains 'BugID: none' from the log message after validation of that value as being allowed is made.

The following is a little silly example of a `verifymsg' file, together with the corresponding `rcsinfo' file, the log message template and an verification script. We begin with the log message template. We want to always record a bug-id number on the first line of the log message. The rest of log message is free text. The following template is found in the file `/usr/cvssupport/tc.template'.

BugId:

The script `/usr/cvssupport/bugid.verify' is used to evaluate the log message.

#!/bin/sh
#
#       bugid.verify filename
#
#  Verify that the log message contains a valid bugid
#  on the first line.
#
if sed 1q < $1 | grep '^BugId:[ ]*[0-9][0-9]*$' > /dev/null; then
    exit 0
elif sed 1q < $1 | grep '^BugId:[ ]*none$' > /dev/null; then
    # It is okay to allow commits with 'BugId: none',
    # but do not put that text into the real log message.
    grep -v '^BugId:[ ]*none$' > $1.rewrite
    mv $1.rewrite $1
    exit 0
else
    echo "No BugId found."
    exit 1
fi

The `verifymsg' file contains this line:

^tc     /usr/cvssupport/bugid.verify

The `rcsinfo' file contains this line:

^tc     /usr/cvssupport/tc.template

The `config' file contains this line:

RereadLogAfterVerify=always

C.4.3 Editinfo

The `editinfo' feature has been rendered obsolete. To set a default editor for log messages use the CVSEDITOR, EDITOR environment variables (see section All environment variables which affect CVS) or the `-e' global option (see section Global options). See Verifying log messages, for information on the use of the `verifymsg' feature for evaluating log messages.

If you want to make sure that all log messages look the same way, you can use the `editinfo' file to specify a program that is used to edit the log message. This program could be a custom-made editor that always enforces a certain style of the log message, or maybe a simple shell script that calls an editor, and checks that the entered message contains the required fields.

If no matching line is found in the `editinfo' file, the editor specified in the environment variable $CVSEDITOR is used instead. If that variable is not set, then the environment variable $EDITOR is used instead. If that variable is not set a default will be used. See Committing your changes.

The `editinfo' file is often most useful together with the `rcsinfo' file, which can be used to specify a log message template.

Each line in the `editinfo' file consists of a regular expression and a command-line template. The template must include a program name, and can include any number of arguments. The full path to the current log message template file is appended to the template.

One thing that should be noted is that the `ALL' keyword is not supported. If more than one matching line is found, the first one is used. This can be useful for specifying a default edit script in a module, and then overriding it in a subdirectory.

If the repository name does not match any of the regular expressions in this file, the `DEFAULT' line is used, if it is specified.

If the edit script exits with a non-zero exit status, the commit is aborted.

Note: when CVS is accessing a remote repository, or when the `-m' or `-F' options to cvs commit are used, `editinfo' will not be consulted. There is no good workaround for this; use `verifymsg' instead.

C.4.3.1 Editinfo example

The following is a little silly example of a `editinfo' file, together with the corresponding `rcsinfo' file, the log message template and an editor script. We begin with the log message template. We want to always record a bug-id number on the first line of the log message. The rest of log message is free text. The following template is found in the file `/usr/cvssupport/tc.template'.

BugId:

The script `/usr/cvssupport/bugid.edit' is used to edit the log message.

#!/bin/sh
#
#       bugid.edit filename
#
#  Call $EDITOR on FILENAME, and verify that the
#  resulting file contains a valid bugid on the first
#  line.
if [ "x$EDITOR" = "x" ]; then EDITOR=vi; fi
if [ "x$CVSEDITOR" = "x" ]; then CVSEDITOR=$EDITOR; fi
$CVSEDITOR $1
until head -1|grep '^BugId:[ ]*[0-9][0-9]*$' < $1
do  echo -n  "No BugId found.  Edit again? ([y]/n)"
    read ans
    case ${ans} in
        n*) exit 1;;
    esac
    $CVSEDITOR $1
done

The `editinfo' file contains this line:

^tc     /usr/cvssupport/bugid.edit

The `rcsinfo' file contains this line:

^tc     /usr/cvssupport/tc.template

C.4.4 Loginfo

The `loginfo' file is used to control where `cvs commit' log information is sent. The first entry on a line is a regular expression which is tested against the directory that the change is being made to, relative to the $CVSROOT. If a match is found, then the remainder of the line is a filter program that should expect log information on its standard input. Note that the filter program must read all of the log information or CVS may fail with a broken pipe signal.

If the repository name does not match any of the regular expressions in this file, the `DEFAULT' line is used, if it is specified.

All occurrences of the name `ALL' appearing as a regular expression are used in addition to the first matching regular expression or `DEFAULT'.

The first matching regular expression is used.

See section The commit support files, for a description of the syntax of the `loginfo' file.

The user may specify a format string as part of the filter. The string is composed of a `%' followed by a space, or followed by a single format character, or followed by a set of format characters surrounded by `{' and `}' as separators. The format characters are:

s

file name

V

old version number (pre-checkin)

v

new version number (post-checkin)

All other characters that appear in a format string expand to an empty field (commas separating fields are still provided).

For example, some valid format strings are `%', `%s', `%{s}', and `%{sVv}'.

The output will be a space separated string of tokens enclosed in quotation marks ("). Any embedded dollar signs ($), backticks (`), backslashes (\), or quotation marks will be preceded by a backslash (this allows the shell to correctly parse it as a single string, reguardless of the characters it contains). For backwards compatibility, the first token will be the repository subdirectory. The rest of the tokens will be comma-delimited lists of the information requested in the format string. For example, if `/u/src/master/yoyodyne/tc' is the repository, `%{sVv}' is the format string, and three files (ChangeLog, Makefile, foo.c) were modified, the output might be:

"yoyodyne/tc ChangeLog,1.1,1.2 Makefile,1.3,1.4 foo.c,1.12,1.13"

As another example, `%{}' means that only the name of the repository will be generated.

Note: when CVS is accessing a remote repository, `loginfo' will be run on the remote (i.e., server) side, not the client side (see section Remote repositories).

C.4.4.1 Loginfo example

The following `loginfo' file, together with the tiny shell-script below, appends all log messages to the file `$CVSROOT/CVSROOT/commitlog', and any commits to the administrative files (inside the `CVSROOT' directory) are also logged in `/usr/adm/cvsroot-log'. Commits to the `prog1' directory are mailed to ceder.

ALL             /usr/local/bin/cvs-log $CVSROOT/CVSROOT/commitlog $USER
^CVSROOT        /usr/local/bin/cvs-log /usr/adm/cvsroot-log
^prog1          Mail -s %s ceder

The shell-script `/usr/local/bin/cvs-log' looks like this:

#!/bin/sh
(echo "------------------------------------------------------";
 echo -n $2"  ";
 date;
 echo;
 cat) >> $1

C.4.4.2 Keeping a checked out copy

It is often useful to maintain a directory tree which contains files which correspond to the latest version in the repository. For example, other developers might want to refer to the latest sources without having to check them out, or you might be maintaining a web site with CVS and want every checkin to cause the files used by the web server to be updated.

The way to do this is by having loginfo invoke cvs update. Doing so in the naive way will cause a problem with locks, so the cvs update must be run in the background. Here is an example for unix (this should all be on one line):

^cyclic-pages		(date; cat; (sleep 2; cd /u/www/local-docs;
 cvs -q update -d) &) >> $CVSROOT/CVSROOT/updatelog 2>&1

This will cause checkins to repository directories starting with cyclic-pages to update the checked out tree in `/u/www/local-docs'.

C.5 Rcsinfo

The `rcsinfo' file can be used to specify a form to edit when filling out the commit log. The `rcsinfo' file has a syntax similar to the `verifymsg', `commitinfo' and `loginfo' files. See section The common syntax. Unlike the other files the second part is not a command-line template. Instead, the part after the regular expression should be a full pathname to a file containing the log message template.

If the repository name does not match any of the regular expressions in this file, the `DEFAULT' line is used, if it is specified.

All occurrences of the name `ALL' appearing as a regular expression are used in addition to the first matching regular expression or `DEFAULT'.

The log message template will be used as a default log message. If you specify a log message with `cvs commit -m message' or `cvs commit -f file' that log message will override the template.

See section Verifying log messages, for an example `rcsinfo' file.

When CVS is accessing a remote repository, the contents of `rcsinfo' at the time a directory is first checked out will specify a template which does not then change. If you edit `rcsinfo' or its templates, you may need to check out a new working directory.

C.6 Taginfo

The `taginfo' file defines programs to execute when someone executes a tag or rtag command. The `taginfo' file has the standard form for trigger scripts (see section The Trigger Scripts), where each line is a regular expression followed by a command to execute (see section The common syntax). The arguments passed to the command are, in order, the tagname, operation (add for tag, mov for tag -F, and del for tag -d), repository, and any remaining are pairs of filename revision. A non-zero exit of the filter program will cause the tag to be aborted.

Here is an example of using the `taginfo' file to log tag and rtag commands. In the `taginfo' file put:

ALL /usr/local/cvsroot/CVSROOT/loggit

Where `/usr/local/cvsroot/CVSROOT/loggit' contains the following script:

#!/bin/sh
echo "$@" >>/home/kingdon/cvsroot/CVSROOT/taglog

C.7 Ignoring files via cvsignore

There are certain file names that frequently occur inside your working copy, but that you don't want to put under CVS control. Examples are all the object files that you get while you compile your sources. Normally, when you run `cvs update', it prints a line for each file it encounters that it doesn't know about (see section update output).

CVS has a list of files (or sh(1) file name patterns) that it should ignore while running update, import and release. This list is constructed in the following way.

• The list is initialized to include certain file name patterns: names associated with CVS administration, or with other common source control systems; common names for patch files, object files, archive files, and editor backup files; and other names that are usually artifacts of assorted utilities. Currently, the default list of ignored file name patterns is:

  RCS SCCS CVS CVS.adm RCSLOG cvslog.* tags TAGS .make.state .nse_depinfo *~ #* .#* ,* _$* *$ *.old *.bak *.BAK *.orig *.rej .del-* *.a *.olb *.o *.obj *.so *.exe *.Z *.elc *.ln core

• The per-repository list in `$CVSROOT/CVSROOT/cvsignore' is appended to the list, if that file exists.
• The per-user list in `.cvsignore' in your home directory is appended to the list, if it exists.
• Any entries in the environment variable $CVSIGNORE is appended to the list.
• Any `-I' options given to CVS is appended.
• As CVS traverses through your directories, the contents of any `.cvsignore' will be appended to the list. The patterns found in `.cvsignore' are only valid for the directory that contains them, not for any sub-directories.

In any of the 5 places listed above, a single exclamation mark (`!') clears the ignore list. This can be used if you want to store any file which normally is ignored by CVS.

Specifying `-I !' to cvs import will import everything, which is generally what you want to do if you are importing files from a pristine distribution or any other source which is known to not contain any extraneous files. However, looking at the rules above you will see there is a fly in the ointment; if the distribution contains any `.cvsignore' files, then the patterns from those files will be processed even if `-I !' is specified. The only workaround is to remove the `.cvsignore' files in order to do the import. Because this is awkward, in the future `-I !' might be modified to override `.cvsignore' files in each directory.

Note that the syntax of the ignore files consists of a series of lines, each of which contains a space separated list of filenames. This offers no clean way to specify filenames which contain spaces, but you can use a workaround like `foo?bar' to match a file named `foo bar' (it also matches `fooxbar' and the like). Also note that there is currently no way to specify comments.

C.8 The checkoutlist file

It may be helpful to use CVS to maintain your own files in the `CVSROOT' directory. For example, suppose that you have a script `logcommit.pl' which you run by including the following line in the `commitinfo' administrative file:

ALL   $CVSROOT/CVSROOT/logcommit.pl

To maintain `logcommit.pl' with CVS you would add the following line to the `checkoutlist' administrative file:

logcommit.pl

The format of `checkoutlist' is one line for each file that you want to maintain using CVS, giving the name of the file, followed optionally by more whitespace and any error message that should print if the file cannot be checked out into CVSROOT after a commit:

logcommit.pl	Could not update CVSROOT/logcommit.pl.

After setting up `checkoutlist' in this fashion, the files listed there will function just like CVS's built-in administrative files. For example, when checking in one of the files you should get a message such as:

cvs commit: Rebuilding administrative file database

and the checked out copy in the `CVSROOT' directory should be updated.

Note that listing `passwd' (see section Setting up the server for password authentication) in `checkoutlist' is not recommended for security reasons.

For information about keeping a checkout out copy in a more general context than the one provided by `checkoutlist', see Keeping a checked out copy.

C.9 The history file

The file `$CVSROOT/CVSROOT/history' is used to log information for the history command (see section history—Show status of files and users). This file must be created to turn on logging. This is done automatically if the cvs init command is used to set up the repository (see section Creating a repository).

The file format of the `history' file is documented only in comments in the CVS source code, but generally programs should use the cvs history command to access it anyway, in case the format changes with future releases of CVS.

C.10 Expansions in administrative files

Sometimes in writing an administrative file, you might want the file to be able to know various things based on environment CVS is running in. There are several mechanisms to do that.

To find the home directory of the user running CVS (from the HOME environment variable), use `~' followed by `/' or the end of the line. Likewise for the home directory of user, use `~user'. These variables are expanded on the server machine, and don't get any reasonable expansion if pserver (see section Direct connection with password authentication) is in use; therefore user variables (see below) may be a better choice to customize behavior based on the user running CVS.

One may want to know about various pieces of information internal to CVS. A CVS internal variable has the syntax ${variable}, where variable starts with a letter and consists of alphanumeric characters and `_'. If the character following variable is a non-alphanumeric character other than `_', the `{' and `}' can be omitted. The CVS internal variables are:

CVSROOT

This is the absolute path to the current CVS root directory. See section The Repository, for a description of the various ways to specify this, but note that the internal variable contains just the directory and not any of the access method information.

RCSBIN

In CVS 1.9.18 and older, this specified the directory where CVS was looking for RCS programs. Because CVS no longer runs RCS programs, specifying this internal variable is now an error.

CVSEDITOR

EDITOR

VISUAL

These all expand to the same value, which is the editor that CVS is using. See section Global options, for how to specify this.

USER

Username of the user running CVS (on the CVS server machine). When using pserver, this is the user specified in the repository specification which need not be the same as the username the server is running as (see section Setting up the server for password authentication). Do not confuse this with the environment variable of the same name.

If you want to pass a value to the administrative files which the user who is running CVS can specify, use a user variable. To expand a user variable, the administrative file contains ${=variable}. To set a user variable, specify the global option `-s' to CVS, with argument variable=value. It may be particularly useful to specify this option via `.cvsrc' (see section Default options and the ~/.cvsrc file).

For example, if you want the administrative file to refer to a test directory you might create a user variable TESTDIR. Then if CVS is invoked as

cvs -s TESTDIR=/work/local/tests

and the administrative file contains sh ${=TESTDIR}/runtests, then that string is expanded to sh /work/local/tests/runtests.

All other strings containing `$' are reserved; there is no way to quote a `$' character so that `$' represents itself.

Environment variables passed to administrative files are:

CVS_USER

The CVS-specific username provided by the user, if it can be provided (currently just for the pserver access method), and to the empty string otherwise. (CVS_USER and USER may differ when `$CVSROOT/CVSROOT/passwd' is used to map CVS usernames to system usernames.)

LOGNAME

The username of the system user.

USER

Same as LOGNAME. Do not confuse this with the internal variable of the same name.

C.11 The CVSROOT/config configuration file

The administrative file `config' contains various miscellaneous settings which affect the behavior of CVS. The syntax is slightly different from the other administrative files. Variables are not expanded. Lines which start with `#' are considered comments. Other lines consist of a keyword, `=', and a value. Note that this syntax is very strict. Extraneous spaces or tabs are not permitted.

Currently defined keywords are:

RCSBIN=bindir

For CVS 1.9.12 through 1.9.18, this setting told CVS to look for RCS programs in the bindir directory. Current versions of CVS do not run RCS programs; for compatibility this setting is accepted, but it does nothing.

SystemAuth=value

If value is `yes', then pserver should check for users in the system's user database if not found in `CVSROOT/passwd'. If it is `no', then all pserver users must exist in `CVSROOT/passwd'. The default is `yes'. For more on pserver, see Direct connection with password authentication.

TopLevelAdmin=value

Modify the `checkout' command to create a `CVS' directory at the top level of the new working directory, in addition to `CVS' directories created within checked-out directories. The default value is `no'.

This option is useful if you find yourself performing many commands at the top level of your working directory, rather than in one of the checked out subdirectories. The `CVS' directory created there will mean you don't have to specify CVSROOT for each command. It also provides a place for the `CVS/Template' file (see section How data is stored in the working directory).

LockDir=directory

Put CVS lock files in directory rather than directly in the repository. This is useful if you want to let users read from the repository while giving them write access only to directory, not to the repository. It can also be used to put the locks on a very fast in-memory file system to speed up locking and unlocking the repository. You need to create directory, but CVS will create subdirectories of directory as it needs them. For information on CVS locks, see Several developers simultaneously attempting to run CVS.

Before enabling the LockDir option, make sure that you have tracked down and removed any copies of CVS 1.9 or older. Such versions neither support LockDir, nor will give an error indicating that they don't support it. The result, if this is allowed to happen, is that some CVS users will put the locks one place, and others will put them another place, and therefore the repository could become corrupted. CVS 1.10 does not support LockDir but it will print a warning if run on a repository with LockDir enabled.

LogHistory=value

Control what is logged to the `CVSROOT/history' file (see section history—Show status of files and users). Default of `TOEFWUPCGMAR' (or simply `all') will log all transactions. Any subset of the default is legal. (For example, to only log transactions that modify the `*,v' files, use `LogHistory=TMAR'.)

RereadLogAfterVerify=value

Modify the `commit' command such that CVS will reread the log message after running the program specified by `verifymsg'. value may be one of `yes' or `always', indicating that the log message should always be reread; `no' or `never', indicating that it should never be reread; or value may be `stat', indicating that the file should be checked with the file system `stat()' function to see if it has changed (see warning below) before rereading. The default value is `always'.

The `stat' mode can cause CVS to pause for up to one extra second per directory committed. This can be less IO and CPU intensive but is not recommended for use with large repositories

See section Verifying log messages, for more information on how verifymsg may be used.

This document was generated by Derek R. Price on June, 9 2006 using texi2html 1.77.