Roundup Tracker - Issues

Message7127

Author schlatterbeck
Recipients ber, schlatterbeck
Date 2021-03-18.19:30:32
Message-id <20210318193028.xmh7sdznobmihh7g@runtux.com>
In-reply-to <1616087556.9.0.562823454978.issue2551119@roundup.psfhosted.org> <1616079586.89.0.165307937175.issue2551119@roundup.psfhosted.org>
Hi Bernhard, thanks for the quick reaction.

On Thu, Mar 18, 2021 at 02:59:46PM +0000, Bernhard Reiter wrote:
> 
> Hi Ralf,
> you seem to use the postgresl backen.

Yes. But I guess the problem is also reproduceable with mysql, the
problem is in the query generator not backend specific.

> AFAIK we still use this feature (though I don't know how many searches
> are done with it). We haven't looked at the code for a long while, the
> idea is to be able to compose complex searches.

Cool.

> What is the fastest way to get a reproducable problem setup?

It would be nice if you could come up with examples of all the usual
queries involving AND, OR, NOT and send me
- The parameters to the filter function for a specific query
- Optionally the generated sql (I can infer this myself).

So this probable involves specifying some common queries via the web
interface and looking what the filter method in backends/rdbms_common.py
is called with. Unfortunately there is no logging for this, so you have
to either instrument filter or set a breakpoint there in the python
debugger.
Accompanying the examples with the expression specified in the web
interface would be nice, so I won't have to reverse-engineer that :-)

The whole idea of all this is to come up with a test that covers
functionality people (your customers among them :-) rely on before
proceeding.

On Thu, Mar 18, 2021 at 05:12:36PM +0000, Bernhard Reiter wrote:
> 
> Ralf,
> here is an into-the-blue-patch by one of our devs,
> more like a guess, but maybe it helps you anyway.
> (Sorry for not being more solid here, this certainly
> could have been documented better.)

Hmm, such a patch without a proper test is just what I want to avoid at
this time :-) Once the behaviour is nailed down it should be easy to
come up with a proper patch. Note that I have other systems where this
misbehaves. I ideally would find out what goes wrong there and come up
with additional tests (which would fail at this time) and proceed from
there. I already have some analyses on what I *think* is wrong. But
first I have to find out what it does when it's right :-)

Thanks
Ralf
-- 
Dr. Ralf Schlatterbeck                  Tel:   +43/2243/26465-16
Open Source Consulting                  www:   www.runtux.com
Reichergasse 131, A-3411 Weidling       email: office@runtux.com
History
Date User Action Args
2021-03-18 19:30:33schlatterbecksetrecipients: + schlatterbeck, ber
2021-03-18 19:30:33schlatterbecklinkissue2551119 messages
2021-03-18 19:30:32schlatterbeckcreate