Fissle

About Fissle

Project Goals

When completed, fissle will be easily configured to use one or more local or network-mounted directories. When a user submits a request for a file, fissle will look through the directory tree for files with that name, and then provide a URL so that the user can obtain the file via HTTP (using Apache). For example, a user could request an RPM from a fissle service without knowing the exact path or file name of the RPM; if a match is found, a URL is returned. This simplifies the process of locating a file within a filesystem.

Fissle will apply appropriate access control to files so the user is not aware of files that cannot be accessed by that user. Fissle will also support chaining to other known fissle servers (provided via configuration).

Process

This list of tasks generally follows the tasks that I am going through to create fissle. The tasks jump back and forth between sections of the articles listed below; this list more or less identifies the sections in chronological order.

• Creating A New Project
• Changing Build Options
• Building Your Project
• Handling Command Line Parameters With CommonC++
• Reading Directory Contents
• File Attribute Tests

Articles

The list of articles below are being generated as I work on fissle.

• Using KDevelop
• Using CommonC++
• Programming to the Linux Filesystem

Next Steps

1. Create fissled to run as a daemon application. (Developer content: daemon applications, init scripts, signal handling, syslog)
2. Repurpose fissle as a command-line client to connect to a local fissled. (Developer content: unix domain sockets)
3. Implement a CIM Provider to remotely manage fissled. (Developer content: CIM Providers, enhanced CIM SDK, YaST Modules)
4. Implement caching to speed up searches.
5. Implement multithreading and/or multiprocessing in the daemon to distribute tasks. (Developer content: Multiprocessing and Multithreading, Thread safety, pthreads, IPC, memory-sharing mechanisms)
6. Develop an Apache module to handle remote queries via a web service. (Developer content: Apache modules as web services)
7. Add advanced identity management to more tightly control access to files. (Developer content: identity management, identity-based web services, LDAP)
8. Implement chaining to known alternative Fissle sources. (Developer content: sockets)
9. Create source and binary RPMs. (Developer content: RPM, FHS, LSB, Build.RPM)

Possible Additional Steps

1. Implement listen socket for direct requests. (Developer content: sockets, security)
2. Implement persistent caching. (Developer content: filesystem, searching, memory-mapped files?)
3. Tie system into xinetd. (Developer content: xinetd)

Project Information

News

A preliminary version of Fissle is available in Subversion. Using this version you can execute fissle from the command line to locate a file within a directory. You can search for complete paths or specific file names. It also supports case insensitive searching.

Okay, so far it is not nearly so powerful as find. It isn't done yet.  :)

--Matt ryan 17:21, 16 Sep 2005 (MDT)

[ News Archive ]

Activity

Project Members