Alexander Shishkin schrieb: > 2009/6/11 walter harms <[email protected]>: >> [email protected] schrieb: >>> From: Pekka Pessi <[email protected]> >>> >>> Issue fsync() system call on a file to ensure its buffers are synchronized >>> with the backing storage. >>> >>> >> According to the man page fsync is not in any standard, just to reduce the >> number >> of applets maybe we can add this as option to the "sync" command ? > Well, in my opinion, introducing options that don't exist in any > standard nor are implemented in existing utilities (which is 'sync' in > our case) is not better than adding new applets. And you can always > compile it out if you don't want it which is one of my personal > favorite features of busybox. And in my opinion, reducing the number > of applets is not a worthy goal in itself for busybox these days. > >> I am not sure what use this applet may have in an embedded system but >> what is about fdatasync(2) ? perhaps this can be added as command line >> option too ? > fdatasync is an optional behavior of this fsync applet. > The usecase is "nosync" filesystems like ubifs where you'd want to > make sure a particular file has been written to the medium. I'm sure > there are other use cases as well, though. >
I have never used/needed fsync (i did know there is a shell cmd before this patch). but most people will be aware there is a sync what makes it more likely to be "discovered". IMHO reducing the number of applets is preferable but this is clearly my personal taste. just my 2cents, wh _______________________________________________ busybox mailing list [email protected] http://lists.busybox.net/mailman/listinfo/busybox
