2013/5/3 Quanah Gibson-Mount quanah@zimbra.com
--On Friday, May 03, 2013 6:24 PM +0200 Erwann Abalea eabalea@gmail.com wrote:
Can't you use the postalAddress attribute?
With your examples, it should be something like: postalAddress: 123 1st av$Montreal$QC$GGG RT3$CA
postalAddress: 321 42nd st$Montreal$QC$GGG RT1$CA
This is almost the correct way to format it... it should be:
postalAddress: 123 1st av $ Montreal $ QC $ GGG RT3 $ CA
If I correctly read RFC2252, the space character around the "$" isn't required:
postal-address = dstring *( "$" dstring ) dstring = 1*utf8
And the provided examples don't include such spaces.
I would also note that there is no guaranteed return order for values
unless you use weighted attributes.
Is the weighted attribute standardized LDAP, or specific to OpenLDAP? I can't find supportive definition in RFC45* documents.
Generally the best thing to do if you are going to have multiple addresses (say home, work, business, mailing, etc) is to have custom attributes specifically for those addresses
Or maybe a subordinate leaf for each address (with address elements splitted in several attributes), to be able to use search filters.
--On Friday, May 03, 2013 7:01 PM +0200 Erwann Abalea eabalea@gmail.com wrote:
2013/5/3 Quanah Gibson-Mount quanah@zimbra.com
--On Friday, May 03, 2013 6:24 PM +0200 Erwann Abalea eabalea@gmail.com wrote:
Can't you use the postalAddress attribute? With your examples, it should be something like: postalAddress: 123 1st av$Montreal$QC$GGG RT3$CA
postalAddress: 321 42nd st$Montreal$QC$GGG RT1$CA
This is almost the correct way to format it... it should be:
postalAddress: 123 1st av $ Montreal $ QC $ GGG RT3 $ CA
If I correctly read RFC2252, the space character around the "$" isn't required:
postal-address = dstring *( "$" dstring )
dstring = 1*utf8
And the provided examples don't include such spaces.
Please fix your email client to quote replies properly. ;)
Interesting, I've only ever seen it with spaces around the $'s.
I would also note that there is no guaranteed return order for values unless you use weighted attributes.
Is the weighted attribute standardized LDAP, or specific to OpenLDAP? I can't find supportive definition in RFC45* documents.
This is an OpenLDAP specific overlay (valsort).
Generally the best thing to do if you are going to have multiple addresses (say home, work, business, mailing, etc) is to have custom attributes specifically for those addresses
Or maybe a subordinate leaf for each address (with address elements splitted in several attributes), to be able to use search filters.
Personally, I would avoid subtrees for this. I prefer to see all my data for a given user stored with the user entry. But that's me. ;) I've used custom AUX objectClasses for this in the past to attach to the person entry if they had a specific type of addr.
--Quanah
--
Quanah Gibson-Mount Sr. Member of Technical Staff Zimbra, Inc A Division of VMware, Inc. -------------------- Zimbra :: the leader in open source messaging and collaboration
Thank for your answers,
Given the context, I would go for the postaladdress field if the entries are in RW mode, but if I can't have empty fields, is there a non ambiguous way to determine the different parts of the field when I use $ as a separator.
The VCard FRC describes an address structure as different components separated with a semicolon, and the components are : the post office box; the extended address (e.g., apartment or suite number); the street address; the locality (e.g., city); the region (e.g., state or province); the postal code; the country name (full name in the language specified in Section 5.1)
So translating an vcard address into ldap postaladdress is easy, but is the opposite possible (if not easy) ?
Le 03/05/2013 13:10, Quanah Gibson-Mount a écrit :
--On Friday, May 03, 2013 7:01 PM +0200 Erwann Abalea eabalea@gmail.com wrote:
2013/5/3 Quanah Gibson-Mount quanah@zimbra.com
--On Friday, May 03, 2013 6:24 PM +0200 Erwann Abalea eabalea@gmail.com wrote:
Can't you use the postalAddress attribute? With your examples, it should be something like: postalAddress: 123 1st av$Montreal$QC$GGG RT3$CA
postalAddress: 321 42nd st$Montreal$QC$GGG RT1$CA
This is almost the correct way to format it... it should be:
postalAddress: 123 1st av $ Montreal $ QC $ GGG RT3 $ CA
If I correctly read RFC2252, the space character around the "$" isn't required:
postal-address = dstring *( "$" dstring )
dstring = 1*utf8
And the provided examples don't include such spaces.
Please fix your email client to quote replies properly. ;)
Interesting, I've only ever seen it with spaces around the $'s.
I would also note that there is no guaranteed return order for values unless you use weighted attributes.
Is the weighted attribute standardized LDAP, or specific to OpenLDAP? I can't find supportive definition in RFC45* documents.
This is an OpenLDAP specific overlay (valsort).
Generally the best thing to do if you are going to have multiple addresses (say home, work, business, mailing, etc) is to have custom attributes specifically for those addresses
Or maybe a subordinate leaf for each address (with address elements splitted in several attributes), to be able to use search filters.
Personally, I would avoid subtrees for this. I prefer to see all my data for a given user stored with the user entry. But that's me. ;) I've used custom AUX objectClasses for this in the past to attach to the person entry if they had a specific type of addr.
--Quanah
--
Quanah Gibson-Mount Sr. Member of Technical Staff Zimbra, Inc A Division of VMware, Inc.
Zimbra :: the leader in open source messaging and collaboration
2013/5/3 Nicolas Mora nicolas@babelouest.org:
Thank for your answers,
Given the context, I would go for the postaladdress field if the entries are in RW mode, but if I can't have empty fields, is there a non ambiguous way to determine the different parts of the field when I use $ as a separator.
No. In the format described below, I understand that field elements are position related (i.e. the first field is always the "post office box", the second is always the "extended address", etc). Since a dstring (element stored in an LDAP postalAddress) can't be empty, then you can't reproduce the same mechanism. In other words, you can have ";;" in your VCard address, you can't have "$$" in an LDAP postalAddress, so you will necessarily lose information in your transformation.
The VCard FRC describes an address structure as different components separated with a semicolon, and the components are : the post office box; the extended address (e.g., apartment or suite number); the street address; the locality (e.g., city); the region (e.g., state or province); the postal code; the country name (full name in the language specified in Section 5.1)
So translating an vcard address into ldap postaladdress is easy, but is the opposite possible (if not easy) ?
Le 03/05/2013 14:26, Erwann Abalea a écrit :
In other words, you can have ";;" in your VCard address, you can't have "$$" in an LDAP postalAddress, so you will necessarily lose information in your transformation.
That's what I was afraid of...
I will find a workaround then.
Thanks a lot
Erwann Abalea wrote:
Is the weighted attribute standardized LDAP, or specific to OpenLDAP? I can't find supportive definition in RFC45* documents.
AFAICS draft-chu-ldap-xordered is only implemented in OpenLDAP.
http://tools.ietf.org/html/draft-chu-ldap-xordered-00
Other LDAP servers have similar approach but limited to certain attribute types, e.g. 'aci' in OpenDJ and others.
Ciao, Michael.
openldap-technical@openldap.org
-
Erwann Abalea -
Michael Ströder -
Nicolas Mora -
Quanah Gibson-Mount