David Cassidy wrote:
Guys,

We deploy tomcat in a multi-hosting environment. We make rather extensive use of 
symlinks so that we can
update the system. ie when the user wants to publish their content the symlink is 
changed and hey bingo ...

The problem we're finding is that tomcat is expanding the symlink to the real path. At 
midnight we kinda
clean the system so that old dead versions are removed to save disk space etc.

Of course then tomcat gets broken due the the file not being there.
I've checked the source of Tomcat 4.1.24 and it seems that getCanonicalPath and 
getAbsolutePath
have been used interchangably.

Is there a reason for this ?

If not taken as though the two methods return the same(ish *) result can we have 
getAbsolutePath used
exclusivly ?

No, sorry.


* they both return the absolute path, just is getCanonicalPath expands symlinks to the real dir name.

It does much more than that (ex: removing stuff like /./, etc): it is a minimal access path to the file (starting from the filesystem root).


Remy


--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]



Reply via email to