This is the mail archive of the kawa@sourceware.org mailing list for the Kawa project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: LOOP


* Charles Turner [2012-06-06 17:39] writes:

> On 6 June 2012 10:37, Helmut Eller <eller.helmut@gmail.com> wrote:
>> Apropos LOOP: a while ago I wrote a loop macro in Kawa/Scheme:
>> https://github.com/ellerh/kloop
>
> Gah, I wish I could use this, but I've spent today just trying to
> figure out what the various
>
> "call to unnamed procedure has too many arguments (2; must be 0)"

That's the bug I was talking about.  Savannah bug #36556 should be a
smaller version of this problem.

> are actually telling me (sometimes the warning seems completely wrong,
> and I get differing results with and without -f it seems),

The above errors happen during macro expansion.  For some reason Kawa
continues after that when invoked -f.  That's probably also a bug or a
confusing feature.

> and the
> various "warning - no use of GS.XXXX" are referring to. IIUC, these
> are identifiers created during hygienic expansion, but I can't figure
> how why they're being generated (is this the bug you allude to?).  I
> tried to figure this out using EXPAND from 'syntax-utils; while I may
> have grown some extra hairs on my chest (and a few gray ones on my
> head), I still don't know why those warnings are being generated.

The warnings about unused gensyms are annoying but harmless.  The new
data-flow analysis takes its toll :-).  In CL we could perhaps add some
(declare (ignorable ...)).  Maybe something like 
(with-compile-options no-warn-unused: GS.XXXX body) would work in Kawa.

Helmut




Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]