[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Feature request

From: Stefan Küng <tortoisesvn_at_gmail.com>
Date: 2007-09-13 21:39:35 CEST

fj wrote:
> Stefan Küng wrote:
>>
>> Those links which work on NTFS5 are not symbolic links but hard links.
>> Those are completely different than symbolic links.
>>
>> Stefan
>>
> respectfully, Stefan, are these not symbolic links they're talking about?
> http://msdn2.microsoft.com/en-us/library/aa363878.aspx
>
> I googled the terms:
>
> symbolic links windows
>
> and got quite a few returns that discuss symbolic links. many talking
> about third party utilities that make the underlying capabilities easier
> to access (e.g. http://shell-shocked.org/article.php?id=284, and
> http://blog.taragana.com/index.php/archive/how-to-create-hardlinks-junctions-and-symboliclinks-on-windows/),
> perhaps most well known is sysinternal's junction.exe (free), and then
> there's microsoft's linkd.exe from the win2k resource kit (not free).

But these are *not* symbolic links. If you read the MSDN docs carefully,
you will see that these resemble much more the hard links in Linux.
Symbolic or soft links are actually files (in the filesystem), hard
links however are links in the underlying filesystem (and therefore much
more dangerous if not used carefully).
Windows itself doesn't call those symbolic or hard links anyway: they're
called "reparse points".
For example, the junction points for directories (which is what
resembles the linux hardlinks the most) are created by using reparse points.

Only Vista has now symbolic links. It still has the reparse points

> Others have pointed out potential operational problems arising out of
> having any kind of link in a repository -- I take no issue with that at
> all, but this thread seems to have gotten onto whether windows actually
> supports various types of links or not, irrespective of the advisability
> of having any links in a repository.
>
> I know hard links are different than symbolic links (multiple entries in
> a table in the fs, vs a construct that intercepts io), but based on the
> info, samples of which I've just provided, I believe I am correct wr2
> symbolic links on ntfs5+.
>
> do you still disagree?

Yes, I still disagree :)
There's a reason why MS implemented the symbolic links in Vista: they
didn't exist before :)

Yes, this discussion has turned a little bit into the wrong direction.
I would recommend that people who really want to have the linux symbolic
links turned into Windows shell links to just create *two* of them: one
symbolic link for Linux (which will turn into a text file on Windows)
and a *.lnk file (symbolic shell link) for windows. By doing that,
there's absolutely no problem with compatibility. The only "problem" is
that if one link is modified, the other must be too manually.

Stefan

-- 
        ___
   oo  // \\      "De Chelonian Mobile"
  (_,\/ \_/ \     TortoiseSVN
    \ \_/_\_/>    The coolest Interface to (Sub)Version Control
    /_/   \_\     http://tortoisesvn.net
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tortoisesvn.tigris.org
For additional commands, e-mail: users-help@tortoisesvn.tigris.org
Received on Thu Sep 13 21:36:14 2007

This is an archived mail posted to the TortoiseSVN Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.