hackathon Sunday 24 Jan

Jaap Eldering jaap at jaapeldering.nl
Sat Jan 23 21:02:50 CET 2016

On 23-01-16 18:21, Marc Furon wrote:
> Jaap Eldering wrote:
>> This is to confirm that we will do the hackathon tomorrow. At least
>> Thijs, Nicky, Jan and I will organize ourselves at Thijs' place
>> tomorrow, starting around 11am CET (UTC+1). We will certainly be
>> available via the IRC channel and if people are interested we could have
>> a skype call to discuss things to work on.
> Unfortunately, this is at 2 AM in the greater Los Angeles area. 
> There once was a time when I would have been programming at that 
> hour, but no longer.
> :-)

We'll work throughout the day so you're welcome to join at a more
reasonable local time ;-) but don't feel any obligation.

>> Some issues that could be worked on:
>> - Improve the UI for adding a combined user+team, both manually in the
>> webinterface and via import.
>> - Allow filtering the submissions list on teams,problems,languages,etc.
> I would really like to see the ability to import users and teams
> (with passwords), so if you are taking priority requests, that is one
> of mine.

Ok, we'll take it into account.

>> Some issues for discussion:
>> - (How to) run the compilation phase in a chroot? This requires more
>> dependencies in the chroot environment, conflict of /bin mountpoint,
>> /tmp, ...
>> - Integrate any features from online-judge branch?
> I would also like to see the ability for teams to download any test
> cases (input and output) that are marked as "sample".  I believe that
> is offered in the online-judge branch, although I have not looked at it.

That sounds like a reasonable feature to port.

>> Suggestions for other features welcome (also from people who are not
>> going to join tomorrow)!
> One thing we would like to be able to do is to include a message with
> an execution result as part of the review process.  We normally do
> this to comply with the (perhaps older) rule that a run-time error
> result should include the nature of the error, but not any
> information that would disclose information about the judges' secret
> data.  For example, a contestant can be told "Segmentation Fault", or
> even "Index out of range", but not a line number or call trace.  This
> would be done as part of the verification process.

Ok, so this would essentially be a message manually entered by the
judges at verification, but shown to the team, right?


More information about the DOMjudge-devel mailing list