... returns success with the text "value does not conform to assertion syntax"; is this the intended behavior, or should the filter error go completely unnoticed, and the text rather be logged, or what?
The current behavior seems to be a consequence of
servers/slapd/ava.c 1.45 -> 1.46 (by hyc)
returning LDAP_SUCCESS instead of rc, without intentionally?) clearing text.
p.
Ing. Pierangelo Masarati OpenLDAP Core Team
SysNet s.r.l. via Dossi, 8 - 27100 Pavia - ITALIA http://www.sys-net.it ----------------------------------- Office: +39 02 23998309 Mobile: +39 333 4963172 Fax: +39 0382 476497 Email: ando@sys-net.it -----------------------------------
Pierangelo Masarati wrote:
... returns success with the text "value does not conform to assertion syntax"; is this the intended behavior, or should the filter error go completely unnoticed, and the text rather be logged, or what?
The current behavior seems to be a consequence of
servers/slapd/ava.c 1.45 -> 1.46 (by hyc)
returning LDAP_SUCCESS instead of rc, without intentionally?) clearing text.
Probably the text should be NULLed out. Illegal filters are supposed to be ignored; certainly no diagnostic should be sent back to any client.
Howard Chu wrote:
Illegal filters are supposed to be ignored; certainly no diagnostic should be sent back to any client.
Well... I think _illegal_ filters should get a protocolError.
OTOH filtering for unrecognized schema should get no error or diagnostic, the filter component just evaluates to Undefined (like False, except Not(Undefined) = Undefined).
Hallvard B Furuseth wrote:
Howard Chu wrote:
Illegal filters are supposed to be ignored; certainly no diagnostic should be sent back to any client.
Well... I think _illegal_ filters should get a protocolError.
OTOH filtering for unrecognized schema should get no error or diagnostic, the filter component just evaluates to Undefined (like False, except Not(Undefined) = Undefined).
OK right, that's what we were really talking about here anyway.
Howard Chu wrote:
Hallvard B Furuseth wrote:
Howard Chu wrote:
Illegal filters are supposed to be ignored; certainly no diagnostic should be sent back to any client.
Well... I think _illegal_ filters should get a protocolError.
OTOH filtering for unrecognized schema should get no error or diagnostic, the filter component just evaluates to Undefined (like False, except Not(Undefined) = Undefined).
OK right, that's what we were really talking about here anyway.
I'd also appreciate if (objectClass=UNDEFINED) does not lead to protocolError or another LDAP error being returned instead of success.
Ciao, Michael.
Michael Ströder wrote:
I'd also appreciate if (objectClass=UNDEFINED) does not lead to protocolError or another LDAP error being returned instead of success.
Don't worry, it won't.
Hallvard B Furuseth wrote:
Howard Chu wrote:
Illegal filters are supposed to be ignored; certainly no diagnostic should be sent back to any client.
Well... I think _illegal_ filters should get a protocolError.
:)
OTOH filtering for unrecognized schema should get no error or diagnostic, the filter component just evaluates to Undefined (like False, except Not(Undefined) = Undefined).
Yes. But nothing prevents success from being accompanied by further informative info in the diagnosticMessage field. That's why I asked first. I concur the filter above qualifies as undefined, and should be treated much like other undefined filters. Already fixed in HEAD (ITS#5688).
p.
Ing. Pierangelo Masarati OpenLDAP Core Team
SysNet s.r.l. via Dossi, 8 - 27100 Pavia - ITALIA http://www.sys-net.it ----------------------------------- Office: +39 02 23998309 Mobile: +39 333 4963172 Fax: +39 0382 476497 Email: ando@sys-net.it -----------------------------------
On Sep 5, 2008, at 4:05 PM, Pierangelo Masarati wrote:
OTOH filtering for unrecognized schema should get no error or diagnostic, the filter component just evaluates to Undefined (like False, except Not(Undefined) = Undefined).
Yes. But nothing prevents success from being accompanied by further informative info in the diagnosticMessage field.
Correct. IIRC, there being at least a few of cases where we intentionally provide a diagnosticMessage with success. Whether this is one of those intentional cases or not, well, I'm not sure but I vaguely recall something about this one.
That's why I asked first. I concur the filter above qualifies as undefined, and should be treated much like other undefined filters. Already fixed in HEAD (ITS#5688).
Kurt Zeilenga wrote:
On Sep 5, 2008, at 4:05 PM, Pierangelo Masarati wrote:
OTOH filtering for unrecognized schema should get no error or diagnostic, the filter component just evaluates to Undefined (like False, except Not(Undefined) = Undefined).
Yes. But nothing prevents success from being accompanied by further informative info in the diagnosticMessage field.
Correct. IIRC, there being at least a few of cases where we intentionally provide a diagnosticMessage with success. Whether this is one of those intentional cases or not, well, I'm not sure but I vaguely recall something about this one.
If that was the desire, it wasn't very effective. E.g., if there were multiple problems with a filter only the last message would have been returned.
-
Hallvard B Furuseth -
Howard Chu -
Kurt Zeilenga -
Michael Ströder -
Pierangelo Masarati