Hi, Since someone is probably going to undertake redeveloping the virtual file system heres some suggestions I have:
1) support for other resources ftp://www.blaaa.org/dir file:///home/genjix/pic.png Having just a path the / should default to mounted filesystems /home/ 2) Autodetecting of archives (detect magic bytes) and mounting /some/path/archive.zip/folderinside/file.dat 3) Abstraction of archive interface so that other archive formats could be added at a later date possibly. 4) Not such a monolithic config system needed (maybe VFS can check for hidden config.vfs files when directories are mounted?) 5) inode's or some kind of metadata `could' be useful depending on how much you would like the VFS to be utilised. At the least I think leaving the option open should be done if VFS is to change. 6) Some sort of persistent data _storage_ would be nice although I think this is unrelated to VFS (most likely) and would be an internal program registry/database. Plus VFS would then need write functions in its api, and I think most of iVFS beauty atm is in its ease and simplicity. On another note an internal registry could be a useful place to store VFS mounted paths. 7) Formalised root folders (like the /mnt suggested on crystalspace website) The support for resource prefixes should only exist where there is a single one of that resource (there is only ftp, filesystem whereas there can be multiple archives on a filesys). /mnt/cdrom rather than cdrom://path/ Would all superblocks go under a certain location like /mnt? This imho will allow a certain flexibility for developers using the VFS to try out novel concepts like a content wiki for a game... or something :P (this is just an example of an adverse way in which it could be (mis)used) http://www.cs.bell-labs.com/magic/man2html/4/wikifs Definetly some things in VFS would be nice like some automation, as an example being able to write a program to download a file straight onto the VFS and it will be installed correctly. Being able to automate VFS would be the most exciting feature. Things like users installing patch files and VFS being able to integrate the archive someway into the program by *overwriting* certain files (not permanemtly - just the mount points and mounted files) so that no changes to the program are needed would be great. I really see no point in mounting user paths unless VFS has some sort of write capacity in its api which I'm not sure whether I consider a good idea. Heres a link you may find interesting, eventhough I don't agree wholeheartedly with his design principles it makes for a fun read http://www.flipcode.com/articles/article_vfs01.shtml Why doesn't CS maintain a list every so often of tasks that need doing? With a bit more details than the 1.0 bullet list? Amir Taaki ------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid0709&bid&3057&dat1642 _______________________________________________ Crystal-main mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/crystal-main Unsubscribe: mailto:[EMAIL PROTECTED]
