From: Anton Bassov on
Corinna,

> You're mixing DOS device names with NT device names. There's no
> restriction for native NT symlinks to be in a certain directory.

I am afraid we just misunderstood one another.

Please look at the original question - the OP asks us why his CreateFile()
call fails.
Therefore, I was speaking only about those links that interface devices to
applications. These links have to be specified in the form
"\\DosDevices\\xxxx". This is what I was speaking about in my reply to the OP
- you just don't specify "\\Device\\xxxx" in CreateFile() calls.

However, you disagreed with me - you said that
"\Device\HarddiskX\PartitionY" is not invalid. In context of this discussion,
this statement is plainly wrong, unless you make it clear that you speak
about native NT symlinks (which you did not do). The way I understood it, you
were arguing with the fact that apps have access only to those links that are
specified in "\\DosDevices\\xxxx" form. Now it turns out that you were
speaking about native NT symlinks.....

I just wonder how native NT symlinks are related to this discussion.....

Anton Bassov


"Corinna Vinschen" wrote:

> Anton Bassov wrote:
> >> > Symbolic link just cannot be in the form
> >> > "\\Device\\HarddiskX\PartitionY" - it has to be in the form
> >> > "\\DosDevices\\xxxx" and not "\\Device\\xxxx".....
> >
> >> This is simply wrong.....
> >> Did you have a look into the native NT name space lately?
> >
> > Everything seems to be clear with you - once you don't seem to see the
> > difference between native NT names and symbolic links, apparently, you
> > just have never written drivers. If you have written a *SINGLE*
> > driver, you would not have questioned the fact that native NT device
> > name that you specify in IoCreateDevice() call must be in the form
> > "\\Device\\xxxx", and that symbolic link that you associate with this
> > native name in IoCreateSymbolicLink() call has to be in the form
> > "\\DosDevices\\xxxx". [BLAH]
>
> You're mixing DOS device names with NT device names. There's no
> restriction for native NT symlinks to be in a certain directory.
>
> >> Sysinternal's winobj.exe is your friend (as far as it still exists).
> >
> > This is a good tool. Run it, look at contents of 'Global??' folder,
> > and, at this point, you will see that there are no symbolic links
> > like"HarddiskX\PartitionY", i.e. something that you mentioned in your
> > previous post (I skipped this piece of nonsense in my reply to it).
> > In fact, > [drop "fact"]
>
> You apparently didn't read my post closly. I never wrote about symlinks
> of the form \Global??\HarddiskX\PartitionY, I wrote about
> \Device\HarddiskX\PartitionY.
>
> The funny thing here is that I know that these symlinks exist and (this
> is an incredible new concept) even *tested* and *use* them. Guess how
> raw disk access is implemented in Cygwin.
>
> Instead of telling me where to look, maybe you should better take a
> look yourself.
>
>
> Corinna
>
> --
> Antworten an o.g. (existierende) Adresse werden ungelesen verworfen.
> Private Mails bitte an corinnaPLOPvinschenPINGde.
>
From: Corinna Vinschen on
Anton Bassov wrote:
> I just wonder how native NT symlinks are related to this
> discussion.....

I'm sorry to read this. I suggest to read the whole thread again
to find out.


Corinna

--
Antworten an o.g. (existierende) Adresse werden ungelesen verworfen.
Private Mails bitte an corinnaPLOPvinschenPINGde.
From: Corinna Vinschen on
Anton Bassov wrote:
> Maxim,
>
>> No, Corinna is correct.
>
> She just should have made it clear that she speaks about native links and

It's funny how you shift the blame for this "misunderstanding" to me.

EOD,
Corinna

--
Antworten an o.g. (existierende) Adresse werden ungelesen verworfen.
Private Mails bitte an corinnaPLOPvinschenPINGde.
From: Anton Bassov on
Corinna,

> It's funny how you shift the blame for this "misunderstanding" to me.

"Shift the blame" is a bit to the extreme - after all, this is not a
situation when someone has to be blamed.....

Have a nice day


Anton Bassov

"Corinna Vinschen" wrote:

> Anton Bassov wrote:
> > Maxim,
> >
> >> No, Corinna is correct.
> >
> > She just should have made it clear that she speaks about native links and
>
> It's funny how you shift the blame for this "misunderstanding" to me.
>
> EOD,
> Corinna
>
> --
> Antworten an o.g. (existierende) Adresse werden ungelesen verworfen.
> Private Mails bitte an corinnaPLOPvinschenPINGde.
>