The sdap_save_* refactoring brought one bug where instead of saving into SYSDB_ORIG_MEMBEROF we would save into SYSDB_MEMBEROF. I rechecked all other attribute names and this was the only one that was changed.
The sdap_save_* refactoring brought one bug where instead of saving into SYSDB_ORIG_MEMBEROF we would save into SYSDB_MEMBEROF. I rechecked all other attribute names and this was the only one that was changed.
Nack,
the issue is deeper than it looks. In the original code, there was conversion from SYSDB_MEMBEROF to SYSDB_ORIG_MEMBEROF, hence this fix should be done not only here, but also in user map. Also please review if the memberof attribute is mapped correctly in other maps, I think I saw some similar code on other places.
Thanks Jan
On Tue, Dec 20, 2011 at 01:45:59PM +0100, Jan Zelený wrote:
The sdap_save_* refactoring brought one bug where instead of saving into SYSDB_ORIG_MEMBEROF we would save into SYSDB_MEMBEROF. I rechecked all other attribute names and this was the only one that was changed.
Nack,
the issue is deeper than it looks. In the original code, there was conversion from SYSDB_MEMBEROF to SYSDB_ORIG_MEMBEROF, hence this fix should be done not only here, but also in user map. Also please review if the memberof attribute is mapped correctly in other maps, I think I saw some similar code on other places.
Thanks Jan
Thanks, you're right. New patch attached.
On Tue, 2011-12-20 at 16:16 +0100, Jakub Hrozek wrote:
On Tue, Dec 20, 2011 at 01:45:59PM +0100, Jan Zelený wrote:
The sdap_save_* refactoring brought one bug where instead of saving into SYSDB_ORIG_MEMBEROF we would save into SYSDB_MEMBEROF. I rechecked all other attribute names and this was the only one that was changed.
Nack,
the issue is deeper than it looks. In the original code, there was conversion from SYSDB_MEMBEROF to SYSDB_ORIG_MEMBEROF, hence this fix should be done not only here, but also in user map. Also please review if the memberof attribute is mapped correctly in other maps, I think I saw some similar code on other places.
Thanks Jan
Thanks, you're right. New patch attached.
Ack
On Tue, 2011-12-20 at 13:18 -0500, Stephen Gallagher wrote:
On Tue, 2011-12-20 at 16:16 +0100, Jakub Hrozek wrote:
On Tue, Dec 20, 2011 at 01:45:59PM +0100, Jan Zelený wrote:
The sdap_save_* refactoring brought one bug where instead of saving into SYSDB_ORIG_MEMBEROF we would save into SYSDB_MEMBEROF. I rechecked all other attribute names and this was the only one that was changed.
Nack,
the issue is deeper than it looks. In the original code, there was conversion from SYSDB_MEMBEROF to SYSDB_ORIG_MEMBEROF, hence this fix should be done not only here, but also in user map. Also please review if the memberof attribute is mapped correctly in other maps, I think I saw some similar code on other places.
Thanks Jan
Thanks, you're right. New patch attached.
Ack
Pushed to master.
On Tue, 2011-12-20 at 16:16 +0100, Jakub Hrozek wrote:
On Tue, Dec 20, 2011 at 01:45:59PM +0100, Jan Zelený wrote:
The sdap_save_* refactoring brought one bug where instead of saving into SYSDB_ORIG_MEMBEROF we would save into SYSDB_MEMBEROF. I rechecked all other attribute names and this was the only one that was changed.
Nack,
the issue is deeper than it looks. In the original code, there was conversion from SYSDB_MEMBEROF to SYSDB_ORIG_MEMBEROF, hence this fix should be done not only here, but also in user map. Also please review if the memberof attribute is mapped correctly in other maps, I think I saw some similar code on other places.
Thanks Jan
Thanks, you're right. New patch attached.
Ack
Although this was the original approach for the conversion, I'd like to discuss the possibility of doing it directly in the map (i.e. changing sysdb attribute for ldap_user_member_of to SYSDB_ORIG_MEMBEROF). That way the sdap_attrs_add_list() could be used again.
Jan
On Wed, 2011-12-21 at 08:48 +0100, Jan Zelený wrote:
On Tue, 2011-12-20 at 16:16 +0100, Jakub Hrozek wrote:
On Tue, Dec 20, 2011 at 01:45:59PM +0100, Jan Zelený wrote:
The sdap_save_* refactoring brought one bug where instead of saving into SYSDB_ORIG_MEMBEROF we would save into SYSDB_MEMBEROF. I rechecked all other attribute names and this was the only one that was changed.
Nack,
the issue is deeper than it looks. In the original code, there was conversion from SYSDB_MEMBEROF to SYSDB_ORIG_MEMBEROF, hence this fix should be done not only here, but also in user map. Also please review if the memberof attribute is mapped correctly in other maps, I think I saw some similar code on other places.
Thanks Jan
Thanks, you're right. New patch attached.
Ack
Although this was the original approach for the conversion, I'd like to discuss the possibility of doing it directly in the map (i.e. changing sysdb attribute for ldap_user_member_of to SYSDB_ORIG_MEMBEROF). That way the sdap_attrs_add_list() could be used again.
Jan
Ok, I misunderstood your proposal originally. I don't want to do that right this moment, though. I'm absolutely certain that this will have further-reaching impact. Please open a bug on this and we'll look into it in 1.8 or 1.9.
On Wed, 2011-12-21 at 08:48 +0100, Jan Zelený wrote:
On Tue, 2011-12-20 at 16:16 +0100, Jakub Hrozek wrote:
On Tue, Dec 20, 2011 at 01:45:59PM +0100, Jan Zelený wrote:
The sdap_save_* refactoring brought one bug where instead of saving into SYSDB_ORIG_MEMBEROF we would save into SYSDB_MEMBEROF. I rechecked all other attribute names and this was the only one that was changed.
Nack,
the issue is deeper than it looks. In the original code, there was conversion from SYSDB_MEMBEROF to SYSDB_ORIG_MEMBEROF, hence this fix should be done not only here, but also in user map. Also please review if the memberof attribute is mapped correctly in other maps, I think I saw some similar code on other places.
Thanks Jan
Thanks, you're right. New patch attached.
Ack
Although this was the original approach for the conversion, I'd like to discuss the possibility of doing it directly in the map (i.e. changing sysdb attribute for ldap_user_member_of to SYSDB_ORIG_MEMBEROF). That way the sdap_attrs_add_list() could be used again.
Jan
Ok, I misunderstood your proposal originally. I don't want to do that right this moment, though. I'm absolutely certain that this will have further-reaching impact. Please open a bug on this and we'll look into it in 1.8 or 1.9.
I already filed related ticket: https://fedorahosted.org/sssd/ticket/1120
This particular issue is tracked in: https://fedorahosted.org/sssd/ticket/1122
Jan
sssd-devel@lists.fedorahosted.org