Re: INFORMIXTEMP directory

Jonathan Leffler wrote: > Helmut Leininger wrote: > > Vic Glass wrote: > > > INFORMIXTMP *must* exist under the root directory, but it can be a > > > symbolic link to a directory some place else. For example: > > > > > > ln -s /tb/INFORMIXTMP /INFORMIXTMP > > > > ??? I think you mean: > > ln -s INFORMIXTMP /tb/INFORMIXTMP > > No; the first version is correct. It took me a while (about 5 years, > probably) to get the hang of symbolic links. > > In the simple case where both names specify a file or directory and > arg2 does not yet exist, what happens with a symlink is actually the > same as with hard links: > arg1 specifies the 'existing' file > arg2 specifies the newly created file > So, in Vic's example, /tb/INFORMIXTMP is a pre-existing directory, > and /INFORMIXTMP does not exist prior to the command being executed. > Afterwards, the /INFORMIXTMP symlink contains the information > /tb/INFORMIXTMP, do that when a program tries to access /INFORMIXTMP, > the Unix kernel's name parsing code starts over on the string > /tb/INFORMIXTMP, ending up at the directory as specified. > > Handling relative symlink values is fraught with complications. > The effect of Helmut's example depends on the current directory > when the command is execute. > > -- > Jonathan Leffler (jleffler@informix.com, jleffler@earthlink.net) > Guardian of DBD::Informix v0.62 -- see http://www.perl.com/CPAN > #include <disclaimer.h> Hi Art, Jonathan, It was a silly mistake. I always think twice about the direction of symbolic links - but sometimes it happens, though. Regards -- Helmut Leininger Bull AG / Vienna Open Systems Support Email: h.leininger@bull.at helmut.leininger@bull.net This opinion is mine and not necessarily that of my employer. No guarantees whatsoever.