Re: [Chicken-hackers] CR #1142 and upcoming changes

[Jörg F. Wittenberger]

Am 18.08.2014 17:06, schrieb Peter Bex:
> Yeah, we're pretty thinly spread right now.
>
> I think calling this CHICKEN 5 may be a good idea.  I don't know for
> sure though: adding backwards compatibility may actually be easier
> in this situation.  Ripping out the SRFIs from core should be pretty
> simple, but it does require updating all the eggs.  And of course
> there's shitloads of eggs that are unmaintained, which will break
> and never get fixed.  So all in all, starting over might be easiest.
>
> I'd love to hear from some of the people using CHICKEN in their business
> or for other Serious Projects (Kristian? Ivan? Andy?) how painful this
> would be for them.

I rely on chicken every day.  Web pages, file system (at least one 
directory) with content synchronization, calendar etc. Customers depend 
on it too.

Our code is fairly demanding (after all: p2p consensus is demanding) and 
moreover required to produce the same results (down to crypto hashs) 
using alternative implementations. (Which is not always easy, since 
things like print representations of number pose easily a problem. 
Another is that SQLite might change the memory layout it produces 
without notice - which is why we now include a reference to a specific 
version.)

Also: we often use "strange" hosts: no root, badly administrated (for 
whatever reasons).

This lead to the decision that we don't use anything eggs infrastructure 
at all.  Instead we always test the static compiled build too.

I'll have to see how to deal with the planned changes.  Currently I'm a 
bit busy writing a little application.

Best

/Jörg