cvs update - not (always) pruning empty directories...
I've been noticing something odd over the past few EAP releases,
occasionally, when doing a CVS update, I get ancient old empty
directories turning up in my workspace.
i.e. ages ago their used to be a ./source directory, but the code is now
in ./src/java
When I update ( with prune on ), this directory, and many many others
magically appear ( and are empty ), if I do an update again, they
dissapear and are pruned.
Has anyone else noticed behaviour similiar to this at all?
请先登录再写评论。
Mark Derricutt wrote:
never happened here - is it possible you accidently unchecked the prune
checkbox?
It's a known bug http://www.jetbrains.net/jira/browse/IDEADEV-664
I have seen it too. There is even a workaround: remove filewatcher.dll.
Bas
Bas Leijdekkers wrote:
Hmmm - I'm seeing this on linux where there is no filewatcher.dll....
Now that's really strange...
Problem I'm aware of is due async nature of the filesystem synchronization
with filewatcher.
-
Maxim Shafirov
http://www.jetbrains.com
"Develop with pleasure!"
Maxim Shafirov (JetBrains) wrote:
I'm seeing it occur often, but not consistently. I'll pay more
attention to it and see if I can spot any consistentcy with it
Mark Derricutt wrote:
>> Now that's really strange...
>> Problem I'm aware of is due async nature of the filesystem
>> synchronization with filewatcher.
It's simple. It happens every other time, at least for me.
Like this:
Update -> get the directories
Update -> prune the directories
Update -> get the directories
Update -> prune the directories
And so on...
I, of course, have it set to "Create new directories (-d)" and "Prune
empty directories (-P)".
/Kreiger
Attachment(s):
signature.asc
You're on Linux?
-
Maxim Shafirov
http://www.jetbrains.com
"Develop with pleasure!"
>> Maxim Shafirov (JetBrains) wrote:
>>
>>> Now that's really strange...
>>> Problem I'm aware of is due async nature of the filesystem
>>> synchronization with filewatcher.
>> I'm seeing it occur often, but not consistently. I'll pay more
>> attention to it and see if I can spot any consistentcy with it
>>
Maxim Shafirov (JetBrains) wrote:
Quite right.
/Kreiger
Attachment(s):
signature.asc