Re: [Sbcl-devel] freeze for sbcl-1.2.10

[Zach B. <xa...@xa...>]

Zach Beane <xa...@xa...> writes:

> Christophe Rhodes <cs...@ca...> writes:
>
>> Hi,
>>
>> It's time to freeze development, aiming to release sbcl-1.2.10 towards
>> the end of next week.  We need to sort out the deprecation fallout
>> regarding QUIT; what's the current status on the list of quicklisp
>> failures?  Testing and reports of other regressions are also very
>> welcome.
>
> There are many failures in current git.
>
> I personally would prefer that the symbol QUIT remain indefinitely to
> avoid breaking old code. If it can be made to do something useful (or at
> least non-harmful) in the process, that would be even better. If there
> is any extra research I can do to help make a decision, I can try to do
> it. For example, if recklessly-p is the main problem with the interface,
> I could check to see which projects actually use that option.
>
> If keeping QUIT is out of the question, I'd prefer that it go back to
> being a style-warning for some time. I'd like an opportunity to
> publicize the issue more widely (via Planet Lisp, reddit, and twitter).
> I'd also like to work directly with the authors of projects that use
> QUIT, so projects can be updated by their authors, or, in the case of
> abandoned software, adopted and updated by the Sharplispers.
>
> If this route is chosen, is there an easy way to ask the SBCL
> environment to upgrade deprecation style-warnings to full warnings so I
> can measure the progress of project updates?
>
> If QUIT remains deprecated with a full warning for 1.2.9, I think it
> will require a similar outreach to explain to people why their programs
> don't work with 1.2.9. I'd rather see that effort spent on helping
> people understand and adapt to the change in advance, instead of dealing
> with breakage after the fact.

Sorry, I wrote 1.2.9 where I meant 1.2.10.

http://report.quicklisp.org/2015-03-22/failure-report.html hes the
current Quicklisp failure report.

Zach