Reply To: "File changed on disk" but should be ignored?

Product Compare Forums Multi-Edit Support "File changed on disk" but should be ignored? Reply To: "File changed on disk" but should be ignored?

#5343
samej71
Participant

So it seems MEW is right and there is a reason to reload file. Now the question is why times differ. Which file has older time? I’d bet it is the file loaded in MEW and server version is a bit newer. Could you verify it for several cases?[/quote:1g1ut11b]

Certainly. The “original” file is (about) two seconds older than the “new” file as reported by the file changed dialog.

BTW, does it depend on the time when you’re out of MEW? I mean, does problem occur when you switch our just after file save and back or is it necessary to have some pauses between? If so, how long?[/quote:1g1ut11b]

It seems like a very short switch out doesn’t invoke the problem, but if it’s over 10 seconds or so then it seems to happen. I haven’t noticed 100% consistent but it’s generally what happens.

It shouldn’t be so hard but I’m not sure if it is a good idea. You described how you lost data when somebody else edited file. If you have bad luck, you can encouter it, too, if somebody saves the file within “tolerated” interval.[/quote:1g1ut11b]

It still checks the filesize, which when someone edits the file (in my enviornment at least) it’s to add or remove something; very, very, rarely would the byte size remain the same. So to have someone make an edit on a file that doesn’t change the bytesize within 3 seconds of me saving the same file is an acceptable risk. It’s only a four second window more than currently present (a single second window is currently possible, if I understand correctly) :)

It would be better to change file use and always work with local file copy which can’t be rewritten by somebody else.[/quote:1g1ut11b]

In most environments I 100% agree, however we work with files that are operated on by a server, so every edit needs to be placed on the server so it can be tested with; and at this time it’s not feasible to set each developer up with a personal server that has the same access (with databases, configurations, mounts to other file servers or SAN) as the primary server.

That unfortunately limits my options to trying to troubleshoot samba (or it might even be brought about by a SMB protocol itself) to identify and rectify the time difference or make MEW slightly forgiving when two machines don’t appear to have a perfect time sync.

I greatly appreciate all the insight and help everyone has provided; and I’m open to any additional suggestions to better work around this.

–James