Re: [pdbsql-devel] 3_2 version

[Collen B. <co...@he...>]

Hi, i see what you mean.

i fixed a little part for mysql (see svn)
the "net sam list user" works again, but "net user" doesn't (but the 
core dumps are gone)

still to struggle with:

net user -> still no good.
listing users also spit out the workstations accounts ?!?!?
listing worksttions also spit out users (makes sense)

so far so good....
little steps but where getting there.

hope you'll get more out of it this weekend..

Cheers, Collen


Herwin Weststrate wrote:
> I guess the bugs are caused by the lazy filling of the entry struct: 
> when using the command `pdbedit -L' only the sid is read, probably the 
> other parts of the structure are used in this command (see my comments 
> on line 508-511 in the pdb_pgsql.c file). Currently these are not 
> initialized whatsoever, so that might cause the segfaults.
> I won't have time until the weekend to have a look at it, if someone 
> would be able to check this (just filling the struct with some phony 
> values would be enough), I'd like to hear the results of it.
>
> Herwin
>
> Collen Blijenberg wrote:
>   
>> Well, here are the test results of the dutch jury (-:
>>
>> looks good (also cleaned up the svn code)
>>
>> bugs found:
>>
>> - net user -> seg. fault + core dump
>> - user manager (windows) -> can't exec RPC + core dump
>> - net sam list users -> strange sighs like (like -?_)
>> - net sam list workstations -> dido strange signs.
>>
>> pdbedit -L works just fine...
>> so we're getting there... thx.
>>
>> dunno if these faults are related to our module or to samba, but my 
>> guess is the module..
>> have fun...
>>
>> Collen.
>>
>>
>> Collen Blijenberg wrote:
>>     
>>> Oeh, excellent Herwin,
>>>
>>> i saw you did mysql also... ! thx
>>> we were busy with it last Friday, but did come to a sollution on the 
>>> search_users thing.
>>> also good notice of the init_module. now both will work.
>>>
>>> i'll down the svn to see the thing in action (-:
>>>
>>> - the table name, don't think it's much of a problem, 2 different names.
>>>   so I'll suggest to leave it as it is. but must say users has my pref.
>>>
>>> - sure upcase the true/false statements.. no worry here. (and looks nicer)
>>> - there hasn't been done mutch work on the pgsql module, maybe we 
>>> simply forgot to alter it ?!
>>>
>>> Thx, i'll get back on it as soon as i test all kind of things....
>>>
>>> Cheers, nice work!!
>>>
>>> Collen.
>>>
>>> Herwin Weststrate wrote:
>>>       
>>>> On Fri, Jun 27, 2008 at 02:03:41PM +0200, Collen Blijenberg wrote:
>>>>   
>>>>         
>>>>> [...]
>>>>> still need to figure out:
>>>>>
>>>>> - enumeration on users (seems that it's no longer there with samba 3.2) 
>>>>>     
>>>>>           
>>>> I fixed that one in pdb_pgsql.c yesterday: samba 3.2 uses a kind of
>>>> iteration-mechanism to retreive the userlist, see the
>>>> pgsqlsam_search_user function to see how it works.
>>>>
>>>> While working on it I noticed a number of things that were a little bit
>>>> unlogical:
>>>> - The mysql table defaults to `user', the postgres table to `users'.
>>>>   Probably because `user' is a keyword in postgres. This kind of
>>>>   inconsistency makes it hell to migrate to another database. I'd
>>>>   propose to use the names that work on both databases as a default, so
>>>>   the usertable should be called `users', the new grouptable should be
>>>>   called `groups' and the column with the name of the group should be
>>>>   called `groupname'. (Groups are added in Samba 3.2.0)
>>>> - In the pdb_pgsql.c file I've counted three occurences of the `execute
>>>>   query, check results, print warnings if erorr'-session (line 358-369,
>>>>   470-479, 524-533) and without looking at the source I guess the mysql
>>>>   version is not much different. A simple function that accepts a
>>>>   string, tries to run the query, DEBUG and return NULL if error, return
>>>>   result pointer would be nicer, it saves some code-copy-paste.
>>>> - The names True/true and False/false are both valid and both in use and
>>>>   sometimes mixed, but this problem also exists in the original source
>>>>   (take a look at pdb_smbpasswd). As long as there's no ``official''
>>>>   Samba standard for this I'd propose to use the uppercased names (True
>>>>   and False).
>>>>
>>>> Any comments on these proposals?
>>>>
>>>> Regards, Herwin
>>>>
>>>> -------------------------------------------------------------------------
>>>> Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
>>>> Studies have shown that voting for your favorite open source project,
>>>> along with a healthy diet, reduces your potential for chronic lameness
>>>> and boredom. Vote Now at http://www.sourceforge.net/community/cca08
>>>>   
>>>>         
>>> ------------------------------------------------------------------------
>>>
>>> -------------------------------------------------------------------------
>>> Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
>>> Studies have shown that voting for your favorite open source project,
>>> along with a healthy diet, reduces your potential for chronic lameness
>>> and boredom. Vote Now at http://www.sourceforge.net/community/cca08
>>>       
>> ------------------------------------------------------------------------
>>
>> -------------------------------------------------------------------------
>> Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
>> Studies have shown that voting for your favorite open source project,
>> along with a healthy diet, reduces your potential for chronic lameness
>> and boredom. Vote Now at http://www.sourceforge.net/community/cca08
>>     
>
>
> -------------------------------------------------------------------------
> Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
> Studies have shown that voting for your favorite open source project,
> along with a healthy diet, reduces your potential for chronic lameness
> and boredom. Vote Now at http://www.sourceforge.net/community/cca08
>