Thanks for finding these, Zach.

multiple-value-variants is broken due to breakage in 'positional-lambda-2.0' which is apparently yet another system that defines a positional argument syntax. Not sure how I missed it, but yet again it's trying to grovel un-eval'd backquoted expressions.

The common theme among the implementations of these constructions is that they're all naive in their code walking, and incidentally in their treatment of literal symbols that happen to coincide with the chosen syntax for positional args, which is irrelevant but interesting in that all such constructs were apparently designed using worse-is-better style.
* (macroexpand-1 '(plambda `(funcall ,:1 ,:2)))
=> (LAMBDA () `(FUNCALL ,:|1| ,:|2|)) ; wrong

Having looked at the source code for plambda, I'd probably rather just suggest that 'multiple-value-variants' not use the broken plambda. It's awfully obfuscated imho and I don't see any immediate fix.

On Tue, Jul 22, 2014 at 12:10 PM, Zach Beane <> wrote:
Christophe Rhodes <> writes:

> Hi all,
> Could we please freeze development in favour of identifying and tracking
> down any regressions, for a release of sbcl-1.2.2 near the end of next
> week?  As ever, reports of testing on applications and libraries,
> whether failures or successes, are most welcome.

I just tried building the Quicklisp world with, and
there are a number of (related?) failures:

  - nuclblog:

  - multiple-value-variants:

  - lol-re:

  - let-over-lambda:

  - esrap-liquid:

  - cl-larval:

  - cl-secure-read:

  - cl-read-macro-tokens:

  - cl-yaclyaml:

I haven't done any analysis or bug reporting on these failures.


Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
Sbcl-devel mailing list