not create a dot file in same folder as document when opening it for viewing

users
#21

question: doesn't the Linux file system (ext2; ext3) register access time to a directory? in that case, wouldn't writing a file to a directory register in the directory's time-stamp? if the OP's back-up methodology is sensitive to that, the problem would remain.

am I off the track?

F.

#22

Yes, they does.

On filesystem-level, directory is (roughly speaking) special file with names of
other files. If you create file, then this list gets updated (directory
modification time changes). If you remove that file, then it gets updated again.

I don't know about OP backup tools (did not read whole thread), but rsync has
--ignore-times and --size-only command-line options which may come handy in
such situation.

#23

wouldn't really help her. I was pointing out that it is not enough to worry about the file's time-stamp; the directory's time-stamp is also altered. if her backup software is sensitive to the latter, then she's no further along. if the backup software isn't sensitive to directory time-stamps, then fine, my point is moot.

still, I'm surprised if there is no way to tell LO not to make a lock file. googling suggests there might be, see 'libreoffice disable file locking' or similar.

F.

#24

And these options of rsync are supposed to ignore files and directories whose
only change is timestamp. Or, in other words, they change rsync algorithm for
finding new files from "if size or timestamp is different, then file has changed"
to "if size is different, then file has changed".

If you have directory of size X, and then add file and remove it, then you
still have directory of size X. And since rsync compares only size, it will
think that directory has not been changed - and skip it.

At least theoretically, because I did not check it.
And, after all, I don't know if Enda even uses rsync for her backups, so my
point may be irrelevant.

#25

you're absolutely right and I was unclear. when I said, 'it wouldn't help her', I meant she doesn't want to ignore directories that have had changes, she wants to ignore directories that have changed specifically because of a lock file.

I'm thinking the best way is, as she proposed originally, stop creating lock files in the first place. the first link I came to googling suggested a way of doing that.

F.