RE: unlock question

I believe one problem is that the 412 status is
basically a grab-bag of error conditions, that varies
from method to method.
The "pre-condition failed" definition is an admission of the
grab-bag nature of a 412, since any error condition can be
interpreted as some kind of pre-condition failure.

Restricting to If header failure would
have been good, but is not what 2518 says (it throws in
keepalive and Overwrite:false behavior).  But I believe
we can minimize the confusion if we rigidly restrict
412 to be the minimum required by 2518 (i.e. If header,
keepalive failure, and Overwrite:false failure).
For everything else that you might be tempted to use
412 for, use either 403 (Forbidden) if the client
cannot change the resource to make the request succeed,
and 409 (Conflict) if it can change the resource to
make the request succeed.

Greg:
2616 states that 400 (Bad Request) means "malformed syntax".
I believe that it is pretty clear that submitting
the wrong lock token with an UNLOCK is not a "malformed
syntax", and therefore mod_dav should not be returning
a 400 in this case.  I'd use 403 (Forbidden).

Cheers,
Geoff

-----Original Message-----
From: Jason Crawford [mailto:ccjason@us.ibm.com]

<<
The interesting thing here is that this is completely symmetric with
LOCK, which *does* specify the token in an IF (for renewal).  I think
there's another issue here entirely:

ISSUE - Lock Renewal should specify the lock to renew with a
LOCK-TOKEN header, not an IF header.  Reasoning: If you have multiple
locks on a resource, then you might specify more than one of them in
the IF.  How does the server know then which one to renew?
>>
I agree.  I think that is a very poor design decision in the spec.  It's
going on the list.


<<
So consider the request

UNLOCK /url ...
LOCK-TOKEN: <some-lock-token>
IF: <some-lock-token>

If I have the resource locked with <some-lock-token>, this will
succeed.  If I've lost the lock and noone else has relocked, then I
get a 400.  But if I've lost the lock and someone else has locked,
then I get a 412.  This seems like an unnecessary assymetry, and
confusing for a client.
>>
First let me say that I intuitively prefer that the IF header checking
should occur before the method specific checks.   I don't think rfc2518
says what it requires, but my preference would be in conflict with the
IF-MATCH header spec after which we say the IF header is modeled.  Although
it's not my preference, below I'll assume the IF header is not checked
first because that's what you've assumed.   But we should decide what
semantics we want and update the spec, right?

Anyway, onward...

If you didn't include the IF: header, then I think 400 would be what
returned as you indicated.  Although returning 400 seems to bother you,  I
think returning something other than 412  is potentially valuable since the
possibility of three well defined error statuses allow the client to easily
distinguish between the three situations you mentioned.

<<
Putting this another way, I'm arguing that I should *always* get a
412 in this situation, whether or not I specify the IF header.
Because ultimately what's going on is exactly that the LOCK-TOKEN
header is claiming the same thing that an IF header would: that I
have this lock!  If I don't, then I should get a 412.
>>
Whether or not you have the IF header?  Hmmm.  This argument could be
extended to many things I think since almost every 4xx error code is a
conceptual precondition fault.

I don't think either option is much better than the other. 400 is a pretty
generic error so unless the text after the status code clarifies, it
doesn't provide a lot of information about the error and could be ambiguous
with other types of errors.  Saying an error is a pre-req error is saying
more about it, but not much more.  And returning 412 if the implicit lock
check failed would create an amgibuity with the situation where the IF
header was actually doing an etag check and the client would like to know
if it's etag check failed or the implicit lock check failed.

Neither options seems great, but I'd think returning 400 is better of the
two.  At least we know with 412 (I hope) that the status indicates that the
client driven check failed and not something else.

<<
To see that UNLOCK changes the resource, consider a lock-null
resource.  Unlocking it makes it go away.  (I can come up with other
examples, but that's the one that requires the least context.)
>>
I agree with Greg.  We've had long discussions about locking that we
suspended in Dec 99 so that we could devote time to other issues.  Despite
the wording of things like the IF header documentation, in those
discussions we all agreed that the lock really isn't meant to change the
resource.  It just acts on the resource and/or the URI.

And the case of null lock resources is odd ball.  It's very controversial
and likely is going to need conceptual revision even if the actual outward
behavior doesn't change.

<<
And the spec mandates that a 412 should be returned in exactly that
case.  The spec is saying that if a resource is locked then you
*must* present the token that locks that resource in order to modify
it.  It's not saying you get a 412 because the IF fails, it's saying
you get a 412 because the precondition that you specify the lock
token has failed.  To see this, suppose I've locked a resource /url
and gotten <some-token> back.  If I execute

PUT /url ...
IF: (NOT <some-other-lock-token>)

then I should get a 412 back.  Not because the IF failed (it hasn't),
but because I haven't presented <some-token> in an IF.
>>
Assuming the IF header is checked second...

You are right is so far as that the implicit check for the locks before
doing a PUT would fail.  But I think you'd get "423 Locked" response though
not a "412 Precondition Failed".

I'm not sure I'm getting your point unless it is that everything is a
pre-condition check and should return "412 Precondition Failed".  I'm
sceptical of that.  I feel that 412 Precondition Failed is largely intended
for optional client driven checks like IF:, IF-MATCH, etc.



<<
I believe the function of the IF header is for *clients* to specify
*additional* preconditions on the execution of a method *other* than
those mandated by the spec.  In this reading, the use of the IF
header for presenting a lock token in order to *satisfy* a spec
precondition is a little confusing, but since such presentation is
also useful as a client-generated precondition (only do this if I
still have this lock) it makes sense.
>>
I agree.  The IF header should be reserved for client-directed checks.
There should be a seperate header for presenting our tokens for the sake of
generic lock checking.  Perhaps call it "Locks-Held:".  No double or triple
duty for the IF header.

J.

Received on Sunday, 4 March 2001 22:32:48 UTC