oqgraph-dev team mailing list archive

[Andrew McDonnell <bugs@xxxxxxxxxxxxxxxxxxx>]

Hi Arjen,

what external constraints if any are imposed on us by the project?

I am assuming but dont know enough to be certain:

* reasonably comprehensive regresison test suite
* Builds on Linux (I have only tested on debian-squeeze-amd64), what other
platforms do we need to rerun the tests on?
* Builds on common Windows (work in progress)
* Builds using various boost versions (myself have tested on
1.54.0(linux,Windows) and 1.42.0(Linux only))
* What about OS/X (which I dont have), FreeBSD, etc?
* Do we need to run a test regression _without_ boost to verify that when
oqgraph doesnt get included by cmake we havent broken anything else?
* any known segfaults are incidental to normal operation and documented in
release notes


--Andrew

On 06/08/13 23:15, Andrew McDonnell wrote:
> Hi Arjen
> 
> I just failed to build on Winxp/32bit/MSVCexpress8 looks like some unixy
> assumptions.  Takes a lot longer in virtualbox than building the native Linux
> 64... Havent setup Win7/64 yet
> 
> Apart from the bugs listed below, we need to go through the bugs added by
> people other than me.
> I assume I need to get it working on Win7/64/MSVCexpress10 and
> Winxp/32/MSVCexpress8(or10) at a minimum
> 
> Bugs I have been working on or I added:
> 
> #459714 Proper test suite for OQGRAPH
> ^^^ I have significantly extended the test suite.  I almost think it is at the
> point where we just add regression tests for specific bugs, YMMV?
> 
> 
> #1196036 V3 Djikstras algorithm Find reachable vertices (origid=X) returns
> weight=0 instead of weight=count of hops
> ^^^ I am unsure about the importance of this one - am I understanding the
> meaning of weights in this context correctly?
> 
> #1195735 V3 mysqld hangs on select from OQGRAPH engine table after deleting
> records from underlying table provided at least one select made and on record
> exists before the delete
> ^^^ I dont think this is a show stopper, can we put in the release notes?
> 
> #1195778 V3 What should be returned when using no_search with 'and origid=x
> and destid=y'
> ^^^ the last comment to this is the nub, i.e. this bug is a requirements
> clarification
> 
> #1134355 V3 Attempting to rename the ORIGID attribute on a OQGRAPH engine
> table to 'parent' will cause mysqld to segfault
> ^^^ I dont think this is a show stopped, can we put in the release notes?
> 
> there are a heap more in launchpad on https://bugs.launchpad.net/oqgraph, some
> seem quite old and/or not relevant to v3 so maybe some triaging and
> classification is needed
> 
> You can probably close  #645830 Insifficient test coverage for oqgraph  as a
> duplicate of #459714, I havent yet looked at the others in detail
> 
> Have you actually checked out my branch and tried the tests for yourself?  It
> would probably be good to get some peer review...
> 
> https://code.launchpad.net/~andymc73/oqgraph/10.0-oqgraph3-varchar
> 
> --Andrew
> 
> 
> On 06/08/13 08:10, Arjen Lentz wrote:
>> Hi Andrew
>>
>> Ok what steps still need to be done before we can submit this for merging into
>> MariaDB 10.0 ?
>> tnx
>>
>>
>> Cheers,
>> Arjen.
> 
>