[ << ] | [ >> ] | [Top] | [Contents] | [Index] | [ ? ] |
One of the central organizing concepts of ArX
is the project
tree. For the most part, a project tree is an ordinary tree of
files, directories and symbolic links -- all files that you create and
maintain as part of your project.
In addition to the tree content you create, ArX
itself maintains
some control files in a project tree. You should regard these extra
control files as part of your project's content. For example, if you
distribute source for a program managed by ArX
, you will
ordinarilly want to include the control files: they are useful to
other people.
To initialize a project tree for the first time, from the root of the tree, use the command:
% arx create-version (project name) |
That will create a subdirectory, at the root of the tree, called
{arch}
. It will also open an editor for entering your
initial log file. Most ArX
control files are stored in the
tree rooted at {arch}
. You should never create, remove, or
modify files there by hand.
When you have multiple project trees for related projects it is good
practice to make them sibling directories. This is because ArX
sometimes caches information in project trees and those caches can
speed up some operations. When looking for information in a cache,
ArX
looks not only in the current project tree, but in any sibling
project trees. For example, while working on ArX
, I might have
several copies of ArX
, each for working on a different set of
features:
% cd ~/wd % ls ArX ArX-branches ArX-inventory ArX-reporting |
In this manual, ~/wd
always refers to a "directory of project
trees". (The convention means the same thing as, but is less
cumbersome than ${PROJECT_TREES}
-- there is no requirement that
your directory of project trees be called ~/wd
, or that you only
have one such directory.)
[ << ] | [ >> ] | [Top] | [Contents] | [Index] | [ ? ] |