Hi Neil,

>If write caching is disabled the edits should be written to disk immediately.

If you're talking about the buffer on the disklevel, I don't think that is true also, but I'm not sure though. Write Buffers can be held on the OS level whether or not using write buffers on the disk. I've been playing with this a few years ago when doing FPW applications on the WAN. If you could use opportunic locking the all changes where made to a OS buffer on the workstation. only when closing the files or doing a flush would flush the changes to the storage medium. Since these buffers are held on the OS level, I'm not sure what happens when a C5 error of the application occurs. My guess is that as long as the workstation hold references to the filebuffer it is not flushed yet, and waiting for it to be closed or a message from the server to break to the next oplock mode.

Walter,


>>David,
>>
>>>>Especially while dealing with a beta release you need to save early and save often. Anytime a C05 hits you will lose anything that hasn't been committed to disk.
>>
>>I wonder if this is always true. File write-buffers are held on the OS level rather than on the application level. If an application crashes, the filebuffers are either discarded or flushed to disk.
>>
>>Walter,