Re: website cleanup: JIGS / RIGS

[Richard Frith-Macdonald]

On 21 Oct 2014, at 17:55, Gregory Casamento <greg.casamento@gmail.com> wrote:
> 
> Riccardo,
> 
> On Tue, Oct 21, 2014 at 6:14 AM, Riccardo Mottola
> <riccardo.mottola@libero.it> wrote:
>> Hi,
>> 
>> our website still needs improvement, remember that I only did a quick
>> release of our new look in 80% of the pages to ease Gregory's presentation.
>> 
>> If we check in our Developer Tools page, I think we have some sort of
>> inconsistency:
>> 
>> JIGS is hosted on "gnustep.it"
>> RIGS is on gnustep.org with the old look
>> 
>> I think they should be on the same place and follow the same "look".
> 
> They should be combined.
> 
>> I suppose Nicola knows best.
>> 
>> Are JIGS /RIGS maintained at all?
> 
> No.

In the case of JIGS it's more accurate to say it's stable ... there's no work 
required as a rule.  Nicola and I both make occasional (really very rare) 
updates.

>> do they work? should we even display them?
> 
> They work, sort of.   We should display them, but note that they are
> both experimental.

We have used JIGS in production systems in the tomcat servlet engine for over a 
decade.
That's quite limited use (we use it to call objc code from java), but it works 
reliably.
I've also made extensive use of it to call java code from objc (in order to use 
a jdbc database client library from objc code).

It does have a few disadvantages:
If you have bad objc code and cause a crash in the objc world, of course it 
kills the process, which is pretty bad if that process is the servlet engine 
dealing with a large web site.
When you have objc code that leaks memory (or just uses a lot) its invisible to 
the java garbage collector and you can carelessly run the process out of memory.
When those problems occur in the ObjC world, it's very hard to debug them 
inside a massively multi-threaded Java process.

None of those are really issues with JIGS though ... as a bridging technology 
it works well.