Evening Ralf, > In the meantime, we are not QL-compatible in some cases, there are > programs, which just run under WMAN2, and that is ok. So why can't we > have real subdirectories for the harddisk side? A QXL file isn't > interchangable.
To be honest, I can't find a reason myself why we are still limited to 36 characters after the device name. Obviously, the directory entry holding a 36 byte space for the file name is partly to blame as the directory name is duplicated within the directory entry. Ok, to be really honest, I can think of one reason - someone will have to write it as part of SMSQ. > On the harddisk side, the way of the subdirectories could be better. > Maybe, all was designed for compatibility, understandable for disks but > not for harddisks. I would like to see something similar to the Linux/Unix manner of operating where the directory name is separate from the file name and thus, we could "easily" (I use the term, loosely - I have no idea how "easy" it would be) have just the filename as the 36 byte entry and the previous paths within the full (possibly longer than 36 character name) typed in filename would each be removed as the directory structure was 'walked'. For example: win1_sources_c68_MyProject_MySubProject_MySubSubProject_main_c is a long file name. On a hard disk/QXL file, the root of the Win1_ drive's directory would have the details of a 'file' named 'sources' and it would have the file_type of directory. Within that directory on disc, the 'c68' entry would also indicate a directory. It would not be stored as 'sources_c68' but simply as 'c68' and so on down the path. At the bottom of the path, when we eventually get there, the 'main_c' entry would have the file type of data file and we are done. Ok, a small problem raers it's ugly head - how to tell that main_c is a single file and not a further directory 'main' with a file called 'c' in it? Maybe we need a new file separator (equally, directory separator) like we have to differentiate between the directory names and the file extensions. I still think the extension part of the name of the file (or folder) should count in the number of characters allowed - so 'main.c' is 6 characters' I would love to see the directory system changed to allow either underscores as at present - for backward compatibility, plus, accept a Linux/Unix style '/' as a directory separator. So, using that option, I could open the above file using: win1/sources/c68/MyProject/MySubProject/MySubSubProject/main.c I can but dream. The advantage of not having the parent directory name encoded as part of the file name itself is obvious, we get a much deeper path in the file system and still have a maximum of 36 characters for the terminal filename. > Does anyone know the way, how the Rebel way or the Thor way was? Were > that real subdirectories or just the same as FS2? I'm not 100% sure, but I think the Thor was the first to offer sub-directories as we currently know them. They had a different file type to normal data files - but for the life of me, I can't remember what they were. The file_type of -1 (255) we have at present for directories was not the one used on the Thor. Cheers, Norman. _______________________________________________ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
