[SCM] Samba Shared Repository - branch master updated
Matthias Dieter Wallnöfer
mdw at samba.org
Sun Nov 14 14:45:16 MST 2010
ekacent,
please double-check this with metze. And if the reason is really
serious, then feel free to revert my commit and remove the parsing code
from "ldb_controls.c".
Greets,
Matthias
Matthieu Patou wrote:
> Hello Matthias,
>
>
>>
>> # We cannot remove all object classes by an empty replace
>> m = Message()
>> diff --git a/source4/scripting/bin/upgradeprovision
>> b/source4/scripting/bin/upgradeprovision
>> index 5d217ab..27e142b 100755
>> --- a/source4/scripting/bin/upgradeprovision
>> +++ b/source4/scripting/bin/upgradeprovision
>> @@ -490,7 +490,7 @@ def handle_special_add(samdb, dn, names):
>> if len(res)> 0 and len(res2) == 0:
>> message(CHANGE, "Existing object %s must be replaced by
>> %s,"
>> "Renaming old object" % (str(oldDn),
>> str(dn)))
>> - samdb.rename(oldDn, objDn, ["relax:0",
>> "local_oid:1.3.6.1.4.1.7165.4.3.16:0"])
>> + samdb.rename(oldDn, objDn, ["relax:0", "provision:0"])
>>
>> return 0
>>
>> @@ -602,7 +602,7 @@ def add_missing_object(ref_samdb, samdb, dn,
>> names, basedn, hash, index):
>> delta.dn = dn
>> if not skip:
>> message(CHANGE,"Object %s will be added" % dn)
>> - samdb.add(delta, ["relax:0",
>> "local_oid:1.3.6.1.4.1.7165.4.3.16:0"])
>> + samdb.add(delta, ["relax:0", "provision:0"])
>> else:
>> message(CHANGE,"Object %s was skipped" % dn)
>>
>> @@ -655,7 +655,7 @@ def add_deletedobj_containers(ref_samdb, samdb,
>> names):
>> for att in hashAttrNotCopied.keys():
>> delta.remove(att)
>>
>> - modcontrols = ["relax:0",
>> "local_oid:1.3.6.1.4.1.7165.4.3.16:0"]
>> + modcontrols = ["relax:0", "provision:0"]
>> samdb.add(delta, modcontrols)
>>
>> listwko = []
>> @@ -992,7 +992,7 @@ def update_present(ref_samdb, samdb, basedn,
>> listPresent, usns, invocationid):
>> #for checkedatt in relaxedatt:
>> for attr in delta.keys():
>> if attr.lower() in relaxedatt:
>> - modcontrols = ["relax:0",
>> "local_oid:1.3.6.1.4.1.7165.4.3.16:0"]
>> + modcontrols = ["relax:0", "provision:0"]
>> message(CHANGE, "%s is different from the reference
>> one, changed"
>
> In my memory when first talked about this control it seems that metze
> told me not to make it public (ie. that can be called with
> --controls="name:0").
>
> That's why I used this format because I was quite sure that you didn't
> make it public.
>
> Metze: should this "relax tests for provision" control be public or
> shoud we used the oid notation ?
>
>
More information about the samba-technical
mailing list