[PATCH] ldb_debug: Fix for bug 13185

[Stefan Metzmacher]

Hi Björn,

can you test this patch and check if it fixes the bug for you?

If that's ok, I'd propose we take that for now unless we get a
test for it shortly.

Thanks!
metze

Am 22.02.2018 um 05:16 schrieb Douglas Bagnall via samba-technical:
> On 22/02/18 16:53, Douglas Bagnall via samba-technical wrote:
>> On 22/02/18 16:06, Gary Lockyer via samba-technical wrote:
>>> +	/*
>>> +	 * Check for duplicates in unique indexes
>>>  	 */
>>>  	if (list->count > 0 &&
>>>  	    ((a != NULL
>>>  	      && (a->flags & LDB_ATTR_FLAG_UNIQUE_INDEX ||
>>> -		 (el->flags & LDB_FLAG_INTERNAL_FORCE_UNIQUE_INDEX))) ||
>>> -	     ldb_attr_cmp(el->name, LTDB_IDXDN) == 0)) {
>>> +		  (el->flags & LDB_FLAG_INTERNAL_FORCE_UNIQUE_INDEX))))) {
>>
>> This looks a bit wrong. I don't think it actually is wrong, but it
>> just looks scary having "&" mixing with other operators without tight
>> parentheses. I had to read random web pages to be sure the precedence
>> of (a & b || c) works as intended.
>>
>> I think something like this is logically the same but a bit clearer:
>>
>>  	if (list->count > 0 &&
>>             a != NULL &&
>>  	    ((a->flags & LDB_ATTR_FLAG_UNIQUE_INDEX) ||
>>              (el->flags & LDB_FLAG_INTERNAL_FORCE_UNIQUE_INDEX))) {
>>
> 
> Well, I spoke with Andrew, and it turns out he meant this:
> 
> 	if (list->count > 0 &&
> 	    ((a != NULL && (a->flags & LDB_ATTR_FLAG_UNIQUE_INDEX)) ||
> 	     (el->flags & LDB_FLAG_INTERNAL_FORCE_UNIQUE_INDEX))) {
> 
> as attached.
> 
> Douglas
> 


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: OpenPGP digital signature
URL: <http://lists.samba.org/pipermail/samba-technical/attachments/20180222/1021c703/signature.sig>