The Rational ClearCase family consists of several software tools for supporting software configuration management (SCM) of source code and other software development assets. It is developed by the Rational Software division of IBM. ClearCase forms the base for configuration management for many large and medium sized businesses and can handle projects with hundreds or thousands of developers.
A part of Rational ClearCase is revision control system, which is a feature for end users.
ClearCase supports two kinds of use models, UCM (Unified Change Management), and base ClearCase. UCM provides an out-of-the-box model while base ClearCase provides a basic infrastructure (upon which UCM is built). Both can be configured to support a wide variety of needs. UCM is part of RUP and therefore all process templates and roles can be used from RUP.
ClearCase can run on a number of platforms including AIX, z/OS, Linux, HP-UX, Solaris, and Windows.[1] It can handle large binary files, large numbers of files, and large repository sizes. It handles branching, labeling, and versioning of directories.
Contents |
ClearCase was developed by Atria Software and first released in 1992[2] on Unix and later on Windows. Some of the Atria developers had worked on an earlier system: DSEE (Domain Software Engineering Environment) from Apollo Computer. After Hewlett-Packard bought Apollo Computer in 1989, they left to form Atria.[3][4][5] Atria later merged with Pure Software to form PureAtria.[6] That firm merged with Rational Software, which was purchased by IBM in 2003.[7] IBM continues to develop and market ClearCase.
DSEE introduced many concepts now used in ClearCase. The Apollo Domain file system allowed special handler programs to intervene during file access, and DSEE made use of this to invisibly substitute a versioned copy when a particular file was opened.[8] With the versioning specification resident in the user environment, all accesses to versioned files were redirected, including such mundane accesses as printing, viewing in a generic text editor etc.
DSEE relied heavily on a file which described all the software modules and their dependencies. This file had to be generated manually, which was a major impediment to its use in large systems. However, once generated it allowed DSEE to calculate the optimum way to perform a build, re-using all modules that had previously been processed and whose version specifications matched the specifications for the current build.
DSEE also introduced the "version spec" which was then called a "thread". This was a list of possible versions that could be in the user environment, or in a build. The major innovation was using build signatures and software release signatures in the thread. The items in a thread might thus be:
Threads were processed from top to bottom for each file. A developer thread might have "reserved" at the top, followed by a labeled version. For a quick fix to an existing release, the thread would be "reserved", then the release signature.
In the absence of the invisible file redirection of the Apollo Domain file system, Rational ClearCase uses the virtual file system provided by the MVFS feature described below. The "thread" concept corresponds to the dynamic view. Support for derived objects in a view is similar to DSEE's concept.
The database system that ClearCase uses is RDM Embedded from Raima. In ClearCase terminology, an individual database is called a VOB (Versioned Object Base). On this layer, maintenance takes place using Raima tooling. Around this layer, a set of interfaces with accompanying tools is available to manage the physical database system. This requires specific Database administrator skills.
Before version 7 every system, both clients and servers, would run the same ClearCase services or daemons. The most important service is called the Atria location Broker Daemon (ALBD) which takes care of all the communication between stations. After version 7 the server platform runs Websphere Application Server with a server application called the "CM Server" which at the same time serves Rational ClearQuest clients and Rational ClearCase Clients via the HTTP Protocol. Before version 7 there was a webservice by which users could access ClearCase via their browser but this has been discontinued. The older client service architecture is still supported. Before version 7 load balancing would be more difficult (specific services would have to run on specific servers) to implement after version 7 it is just a matter of putting multiple servers in sequence.
In a multi-site scenario every server runs ClearCase MultiSite Server and sends packages back and forth (via any protocol available) to synchronize.
Traditionally ClearCase supported full/fat clients. Since version 7 the CCRC 'ClearCase Remote Client' is available which is based on Eclipse (software). It is supplied in both fully packaged Eclipse versions as well as a plugin-in for Eclipse and other environment such as Visual Studio. Communication takes place via the HTTP Client protocol.
The CCRC client is now the preferred client.
A distinguishing feature of ClearCase is a proprietary networked filesystem (MVFS: MultiVersion File System), which can be used to mount VOBs as a virtual file system through a dynamic view, selecting a consistent set of versions and allowing for the production of derived objects. The dynamic view allows this to map to a Software Configuration. This was a departure from the repository/sandbox model, allowing for the early management of artifacts (before they are checked in, and not limited to these first order configuration items).
Alternatively, ClearCase supports snapshot views which are just copies of repository data specified by a config-spec which spans one or several VOBs. As opposed to dynamic views, snapshot views are maintained on a local (OS-specific) filesystem and do not require network access. Instead, a snapshot view stores a copy of the VOB data locally on the user's computer. Snapshot views can be used while disconnected from the network and later resynchronized to the VOB when a connection is reestablished. This mode of operation is similar to how the widely-used CVS (Concurrent Versions System) software works.
From the perspective of software on the client computer, a view appears as just another file system. If new data is created in a ClearCase view then ClearCase will refer to the new data (files, directories, ...) as view-private to indicate it is specific to the view being used and not being version controlled (private) by ClearCase. This allows build systems to operate on the same file system structure as the source code, and ensures that each developer can build independently of each other. A view-private object can be added to source control and become a versioned element at any time, making it visible to other users.
Each developer typically has one or more views at their disposal. It is sometimes practical to share views between developers, but sharing branches is usually used instead. Having a branch hierarchy is often useful, so an entire development project shares a common development branch, while a smaller team shares a sub-branch, and each developer has his or her own private branch. Whenever some change is deemed stable enough for a larger group, it can be merged to the parent branch.
Under base ClearCase, each view is controlled by its associated configuration specification, commonly referred to as a config spec. This is a collection of rules (stored internally in a text file, but compiled to a more efficient form before use) that specifies what elements (files or directories) should be visible in a view, and which versions of these elements. When deciding which version, if any, of an element should be visible, ClearCase traverses the configuration specification line-by-line from top to bottom, stopping when a match is found. Earlier rules always take priority over later ones.
A typical configuration specification could look like this:
# Show all elements that are checked out to this view, regardless any other rules. element * CHECKEDOUT # If an element has a version on the 'module2_dev_branch', then the latest # version of this branch shall be the visible version in this view. element * .../module2_dev_branch/LATEST # For all files named 'somefile', regardless of location, always show the latest version # on the main branch. element .../somefile /main/LATEST # Use a specific version of a specific file. Note: This rule must appear before # the next rule to have any effect! element /vobs/project1/module1/a_header.h /main/proj_dev_branch/my_dev_branch1/14 # For other files in the 'project1/module1' directory, show versions # labeled 'PROJ1_MOD2_LABEL_1'. Furthermore, don't allow any checkouts in this path. element /vobs/project1/module1/... PROJ1_MOD2_LABEL_1 -nocheckout # Show the 'ANOTHER_LABEL' version of all elements under the 'project1/module2' path. # If an element is checked out, then branch that element from the currently # visible version, and add it to the 'module2_dev_branch' branch. element /vobs/project1/module2/... ANOTHER_LABEL -mkbranch module2_dev_branch
A configuration specification can also reference other configuration specifications using the 'include' statement.
Under ClearCase UCM these config specs still exist but do not need to be maintained or edited anymore by end-users or administrators.
The dependency information is stored in a hidden configuration record which can be shown for each derived object. The configuration record can be used to set up another view showing all files that have been read before during the build time. Alternatively, you can use the configuration record to apply a label to the files (and versions) that were read during the build.
|
|||||||||||||||||||||||||||||||||||||||