NAME

DESCRIPTION

The location of a module's files are known as its CVSROOT. The files within CVSROOT are as follows:

$CVSROOT/CVSROOT

Directory containing repository administrative files.

$CVSROOT/CVSROOT/config

File containing various configuration options. The format of each line is as follows:

keyword=value

Extraneous spaces or tabs are not permitted. A line beginning with a hash character (‘#’) is considered a comment and ignored.

The following options are available:

dlimit='value'

Set the process data size limit.

tag='value'

Set the name of a local tag to use in addition to Id.

umask='value'

Set the default umask to use when creating files and directories.

$CVSROOT/CVSROOT/history

File containing reports of checkout, commit, export, release, rtag, and update commands that have been issued. By default, the init command creates the history file. To disable history logging, the history file should be removed manually. The history command shows these reports according to several criteria.

$CVSROOT/CVSROOT/loginfo

File containing associations between modules and handlers for post-commit logging.

$CVSROOT/CVSROOT/modules

The format of each line is as follows:

module [option] directory [file ...] [&module ...]

Empty lines are ignored. A line beginning with a hash character (‘#’) is considered a comment and ignored. A long line can be continued on the next line by specifying a backslash (‘\’) character as last character.

The following options are available:

-d name

Allow check out of module under different name.

-i prog

Specify program prog to run whenever the commit command is issued.

-o prog

Specify program prog to run whenever the checkout command is issued.

-s status

Set status to module.

-t prog

Specify program prog to run whenever the rtag command is issued.

-u prog

Specify program prog to run whenever the update command is issued.

Each directory that is managed by cvs contains a CVS directory. This directory is used by various cvs commands to record the status of files in the current working directory. These files should not be modified manually by the user but by the cvs commands instead. Files that are always stored in every CVS directory are as follows:

CVS/Entries

File containing a list of files managed by cvs. Entries has one line per file or directory describing the state of it with regard to the source repository. These lines have the following format:

/name/revision/timestamp/options/tag

name

Name of the file in the directory.

revision

Revision of the file in the working directory. The revision is equal to ‘0’ if the file has been added with the add command or preceded by the ‘-’ character if the file has been removed with the remove command.

timestamp

Contains the timestamp of the file at the time cvs created it. If the timestamp is different from the actual modification time of the file, it means that the file has been modified.

options

Contains the keyword substitution mode used for this file.

tag

Contains ‘T’ followed by a tagname or ‘D’ followed by a date.

Entries corresponding to directories have the following format:

D/name////

The lines in the Entries file can be in any order.

CVS/Repository

File containing the path to the corresponding directory in the source repository.

CVS/Root

File containing the value of the CVSROOT variable at the time of checkout. This file is used by all cvs commands instead of CVSROOT once cvs has checked for its existence.

Depending on the circumstances, other files may exist in the CVS directory:

CVS/Base

If pseudo-lock mode is enabled in client-server mode, the Base directory contains a copy of the file on which the edit command has been issued. This allows the unedit command to operate even if the CVS server is unreachable.

CVS/Baserev

File listing the revisions for each file contained in the Base directory under the following format:

name/revision/

CVS/Baserev.tmp

Temporary file used to write CVS/Baserev content then atomically renamed to Baserev.

CVS/Checkin.prog

File containing the path to the command specified with the -i option in the CVSROOT/modules file.

CVS/Entries.Backup

Temporary file used to write CVS/Entries content then atomically renamed to Entries.

CVS/Entries.Static

Tells cvs to not add files to the directory unless the update command is issued with the -d option.

CVS/Notify

File containing the notifications that could not be sent to the CVS server but will be sent at the next successful connection.

CVS/Notify.tmp

Temporary file used to write CVS/Notify content then atomically renamed to Notify.

CVS/Tag

File containing the symbolic revision that was used at checkout. The first character of the line is a single letter indicating the type of tag: ‘D’, ‘N’, or ‘T’, for date, non-branch tag, or branch tag, respectively. The rest of the line is the tag or date itself. The Tag file should not be removed manually: instead, use “update -A”.

CVS/Template

File containing the template specified by the CVSROOT/rcsinfo file. It is only used in client-server mode. Locally, cvs uses the rcsinfo file directly.

CVS/Update.prog

File containing the path to the command specified with the -u option in the CVSROOT/modules file.

CVS/<name>,t

File containing the description of the file file that has been added with the add command. This description is used by cvs when archiving this file with the commit command.

$HOME/.cvsignore

This file provides a list of files (or sh(1) file name patterns) that should be ignored by cvs(1) during the import, release, and update commands.

The syntax of the cvs file consists of a series of lines, each of which contains a space-separated list of filenames. There is currently no way of using comments in this file.

Default patterns that are ignored by cvs(1) are as follows:

CVS, RCS, RCSLOG, SCCS, TAGS, tags, core, cvslog*, *.o, *.so, *.a, *.bak, *.orig, *.rej, *.old, *.exe, *.depend, *.obj, *.elc, *.ln, *.olb, *.core, .#*, *~, _$*, *$, #*, ,*.

The list can be modified using the methods described below:

1. The $CVSROOT/CVSROOT/cvsignore file appends patterns to ignore for the whole repository.
2. The $HOME/.cvsignore file appends patterns to ignore for user only.
3. The content of the CVSIGNORE variable is appended to the list.
4. Options to the command line -I flag, available for the import and update commands, appends patterns for the current command line action only.
5. .cvsignore files placed in the repository directories allow to add patterns to ignore locally. They do not take effect on sub-directories.

The ‘!’ character can be used to reset the list of patterns using any of these five methods.

For the update command, cvs(1) handles files that are already present in the repository even if they are in the list to ignore. Files that are in the list to ignore but not in the repository will not be listed. This way, files preceded by the ‘?’ character will not appear even if they should.

$HOME/.cvsrc

This file provides a way to give the cvs(1) program implicit global options and command-specific options. Unless the -f option is specified, cvs(1) reads its startup configuration file .cvsrc from the home directory of the user who invoked it.

The format of each line is as follows:

command [arg ...]

where command is either the ‘cvs’ keyword to specify global options, one of the supported cvs(1) commands or a command alias. Arguments following command will be added implicitly to the appropriate command's argument vector if it is run. Lines whose command argument is not a valid command will generate a warning when running with the -V flag.

Empty lines are ignored. A line beginning with a hash character (‘#’) is considered a comment and ignored.

For example, to specify that cvs(1) should always run in quiet mode and the diff internal command should always produce unified output:

cvs -q
diff -u
    

$HOME/.cvswrappers

This file, located in $CVSROOT/CVSROOT and/or $HOME/.cvswrappers, provides a way to configure filters for cvs(1) based on file type (name). This works by specifying a pattern to match for varying file types.

The format of each line is as follows:

pattern [option 'value'] [option 'value'] ...

The following options are supported:

-f filter

Execute filter when the file is extracted from the repository (for the checkout, export, and update commands).

-k mode

Specify the keyword substitution mode. See the KEYWORD SUBSTITUTION section of rcs(1) for more information.

-m method

Specify the merge methodology to be used when a file is updated.

The methods are as follows:

COPY

When the update command is used, cvs(1) will merely copy one version over another and let the user do the merge by himself. This method is used by default on binary files (see the -k b option).

MERGE

Try to merge the files. This method is the default.

-t filter

Execute filter before the file is archived in the repository (for the commit, and import commands).

$TMPDIR/cvs-serv⟨pid⟩

Temporary directory created by the server where pid is the process ID of the server. It is located in the directory specified by the TMPDIR environment variable or the -T global option. See above for more information.

SEE ALSO