Univention Bugzilla – Bug 30533
Limit substring search in UMC
Last modified: 2016-10-07 18:11:33 CEST
With UCS 3.1 search queries are automatically matching for substring results. Before UCS 3.1 there was the need of using wildcards for achieving this behaviour. In bigger environments, this change could be counterproductive. This was also requested at ticket #2013022021002929 It would be nice to have the opportunity to configure the old search behaviour via ucr.
Two possible implementations: On startup of module: univention/management/console/modules/sanitizers.py +407 self.use_asterisks = kwargs.is_true('use_asterisks', ucr.is_true('directory/manager/web/modules/substringsearch')) On each request, something like: univention/management/console/modules/sanitizers.py +423 if self.use_asterisks is None: use_asterisks = ucr.is_true('directory/manager/web/modules/substringsearch')) "On each request" needs to reload ucr for every PatternSanitizer but it does not require a reload of UMC. As I suppose this is changed at most once in the lifetime of a UCS server, I guess it would be sufficient to check on startup. If we make that configurable in UDM, should we also include UCR, Extended Software Management, Services modules?
Let's wait if we get more feedback like this.
I think it would be better if the search can be limited with ", for example: Search string: test → finds test, test1, test2, ... Search string: "test" → finds only test
(In reply to comment #3) > I think it would be better if the search can be limited with ", for example: > > Search string: test → finds test, test1, test2, ... > > Search string: "test" → finds only test Another option would be to use < and > from regular expressions (word boundaries).
(In reply to comment #4) > (In reply to comment #3) > > I think it would be better if the search can be limited with ", for example: > > > > Search string: test → finds test, test1, test2, ... > > > > Search string: "test" → finds only test > > Another option would be to use < and > from regular expressions (word > boundaries). Yes, but our * does not work like the regular expression * either. I guess this would be counter intuitive. "test" works more like google and more or less every admin uses it like this.
*** Bug 30190 has been marked as a duplicate of this bug. ***
(In reply to comment #6) > *** Bug 30190 has been marked as a duplicate of this bug. *** Also consider the information from Bug 30190: > Attributes that were not defined in the schema-file as: > SUBSTR caseIgnoreSubstringsMatch > > cannot be searched: Since UCS 3.1 asterisks are automatically added to the > search filter: > > Search for "user" -> Search for "*user*" > > This means that attributes without the SUBSTR definition will not be found even > if one searches for exactly the name as in LDAP. > > This also holds for third-party-extensions.
I improved the substring search in some UCS 4.1-0-errata and implemented the solution from comment #1 in Bug #42181. *** This bug has been marked as a duplicate of bug 42181 ***