[Ecls-list] Megapatches ahead

[Gabriel Dos Reis]

On Sun, May 9, 2010 at 4:13 PM, Juan Jose Garcia-Ripoll
<juanjose.garciaripoll at googlemail.com> wrote:
> On Sun, May 9, 2010 at 8:35 PM, Gabriel Dos Reis
> <gdr at integrable-solutions.net> wrote:
>>
>> Another early feedback: ECL HEAD can no longer compile OpenAxiom.
>> More details to follow.
>
> Yes, sorry for the inconvenience. You just have found that ECL is undergoing
> a rather dramatic change. I will try to sketch the problems I am causing and
> solving, but first of all let me apologize for making the CVS/git unstable
> -- though in full rigor nobody should expect it to be other thing :-)
> What I am doing is re-engineering ECL's database of proclamations,
> optimizations and inline forms. There are several delicate changes behind
> - Changes in how the compiler environment is structured
> - Changes in how we store toplevel macros
> - Changes in the types that ECL's compiler accepts
> - Changes in the allowed proclamations and declarations
> - Removal of some compilation-time side effects
> ...
> The list is quite long, but right now the most dramatic change was the third
> one. ECL's compiler used to have a type filter that converted all types in
> declarations into a very small set. That was too small for practical type
> inference, but in removing this type filter a few rough corners were also
> uncovered: functions in the compiler that assumed that reduced set of types.
> I am slowly but quite productively updating things, and uploading them in
> batches so that the test farm can work on the changes and produce some
> feedback.
> Sorry for the inconvenience

First of all, thanks for doing this job.
I was expecting some instabilities -- I just thought you might be interested
in some feedbacks.  I'll hold my "fire" till things get more stable, though
I will be running ECL HEAD from time to time to see what is still working
and what is apparently broken.  Again, thanks!

-- Gaby