Changes between Initial Version and Version 1 of TracEnvironment

May 7, 2010, 10:04:08 PM (12 years ago)



  • TracEnvironment

    v1 v1  
     1= The Trac Environment =
     3Trac uses a directory structure and a database for storing project data. The directory is referred to as the “environment”.
     5== Creating an Environment ==
     7A new Trac environment is created using  [wiki:TracAdmin trac-admin]:
     9$ trac-admin /path/to/myproject initenv
     12[wiki:TracAdmin trac-admin] will ask you for the name of the project and the
     13database connection string (explained below).
     15''Note: The user under which the web server runs will require file system write permission to
     16the environment directory and all the files inside. Please remember to set
     17the appropriate permissions. The same applies to the source code repository,
     18although the user under which Trac runs will only require write access to a Subversion repository created with the BDB file system
     19; for other repository types, check the corresponding plugin's documentation.
     20Also, it seems that project names with spaces can be problematic for authentication (see [trac:#7163]).''
     22''Note: "initenv" when using svn repository does not imply that trac-admin will perform "svnadmin create" for the specified repository path. You need to perform the "svnadmin create" prior to trac-admin initenv if you're creating a new svn repository altogether with a new trac environment, otherwise you will see a message "Warning: couldn't index the repository" when initializing the environment.''
     24== Database Connection Strings ==
     26Since version 0.9, Trac supports both [ SQLite] and
     27[ PostgreSQL] database backends.  Preliminary
     28support for [ MySQL] was added in 0.10.  The default is
     29to use SQLite, which is probably sufficient for most projects. The database
     30file is then stored in the environment directory, and can easily be
     31[wiki:TracBackup backed up] together with the rest of the environment.
     33=== SQLite Connection String ===
     34The connection string for an SQLite database is:
     38where `db/trac.db` is the path to the database file within the Trac environment.
     40=== PostgreSQL Connection String ===
     41If you want to use PostgreSQL or MySQL instead, you'll have to use a
     42different connection string. For example, to connect to a PostgreSQL
     43database on the same machine called `trac`, that allows access to the
     44user `johndoe` with the password `letmein`, use:
     48''Note that due to the way the above string is parsed, the "/" and "@" characters cannot be part of the password.''
     50If PostgreSQL is running on a non-standard port (for example 9342), use:
     55On UNIX, you might want to select a UNIX socket for the transport,
     56either the default socket as defined by the PGHOST environment variable:
     60or a specific one:
     65Note that with PostgreSQL you will have to create the database before running
     66`trac-admin initenv`.
     68See the [ PostgreSQL documentation] for detailed instructions on how to administer [ PostgreSQL].
     69Generally, the following is sufficient to create a database user named `tracuser`, and a database named `trac`.
     71createuser -U postgres -E -P tracuser
     72createdb -U postgres -O tracuser -E UTF8 trac
     74When running `createuser` you will be prompted for the password for the user 'tracuser'. This new user will not be a superuser, will not be allowed to create other databases and will not be allowed to create other roles. These privileges are not needed to run a trac instance. If no password is desired for the user, simply remove the `-P` and `-E` options from the `createuser` command.  Also note that the database should be created as UTF8. LATIN1 encoding causes errors trac's use of unicode in trac.  SQL_ASCII also seems to work.
     76Under some default configurations (debian) one will have run the `createuser` and `createdb` scripts as the `postgres` user.  For example:
     78sudo su - postgres -c 'createuser -U postgres -S -D -R -E -P tracuser'
     79sudo su - postgres -c 'createdb -U postgres -O tracuser -E UTF8 trac'
     82Trac uses the `public` schema by default but you can specify a different schema in the connection string:
     87=== MySQL Connection String ===
     89If you want to use MySQL instead, you'll have to use a
     90different connection string. For example, to connect to a MySQL
     91database on the same machine called `trac`, that allows access to the
     92user `johndoe` with the password `letmein`, the mysql connection string is:
     97== Source Code Repository ==
     99Since version 0.12, a single Trac environment can be connected to more than one repository. There are many different ways to connect repositories to an environment, see TracRepositoryAdmin. This page also details the various attributes that can be set for a repository (like `type`, `url`, `description`).
     101In Trac 0.12 `trac-admin` no longer asks questions related to repositories. Therefore, by default Trac is not connected to any source code repository, and the ''Browse Source'' toolbar item will not be displayed.
     102You can also explicitly disable the `trac.versioncontrol.*` components (which are otherwise still loaded)
     105trac.versioncontrol.* = disabled
     108For some version control systems, it is possible to specify not only the path to the repository,
     109but also a ''scope'' within the repository. Trac will then only show information
     110related to the files and changesets below that scope. The Subversion backend for
     111Trac supports this; for other types, check the corresponding plugin's documentation.
     113Example of a configuration for a Subversion repository used as the default repository:
     116repository_type = svn
     117repository_dir = /path/to/your/repository
     120The configuration for a scoped Subversion repository would be:
     123repository_type = svn
     124repository_dir = /path/to/your/repository/scope/within/repos
     127== Directory Structure ==
     129An environment directory will usually consist of the following files and directories:
     131 * `README` - Brief description of the environment.
     132 * `VERSION` - Contains the environment version identifier.
     133 * `attachments` - Attachments to wiki pages and tickets are stored here.
     134 * `conf`
     135   * `trac.ini` - Main configuration file. See TracIni.
     136 * `db`
     137   * `trac.db` - The SQLite database (if you're using SQLite).
     138 * `htdocs` - directory containing web resources, which can be referenced in Genshi templates. '''''(since 0.11)'''''
     139 * `log` - default directory for log files, if logging is turned on and a relative path is given.
     140 * `plugins` - Environment-specific [wiki:TracPlugins plugins] (Python eggs or single file plugins, since [trac:milestone:0.10 0.10])
     141 * `templates` - Custom Genshi environment-specific templates. '''''(since 0.11)'''''
     142   * `site.html` - method to customize header, footer, and style, described in TracInterfaceCustomization#SiteAppearance
     144'''Caveat:''' ''don't confuse a Trac environment directory with the source code repository directory.''
     146This is a common beginners' mistake.
     147It happens that the structure for a Trac environment is loosely modelled after the Subversion repository directory
     148structure, but those are two disjoint entities and they are not and ''must not'' be located at the same place.
     151See also: TracAdmin, TracBackup, TracIni, TracGuide