Re: [Chicken-users] vector-copy! compatibility and handling + what else changed around -extend?

[Peter Bex]

On Mon, Mar 04, 2013 at 10:58:05AM +0100, Jörg F. Wittenberger wrote:
> Hi all,
> 
> as usual it's challenging to compile an application using a new
> version of chicken.  (In this case I try to upgrade from 5.7.5 to
> current git master.)
> 
> Somehow the initialization has changed.  Now it breaks upon the
> use or slightly complex irregexp's.
> 
> I figured out that this is due to the use of srfi-43.
>
> The problem boils down to chicken having a vector-copy! procedure,
> which takes an incompatible argument order wrt. srfi-43.

I don't see what this has to do with irregex; it uses the "incompatible"
definition from core.  Importing the srfi-43 definitions from the
vector-lib egg shouldn't cause that to break.

> Since chicken claims (or claimed?) to support srfi-43, how
> should I do so now?

I did run into this incompatibility yesterday, as well.  I think
the SRFI-43 definition is superior, as it allows you to cleanly indicate
copying a subsection of one vector to a subsection of another, and
it's more intuitive for people used to srfi-13's string-copy! procedure.

I think it would be nice if we could support both signatures for a while,
deprecating the old one.  If others agree I'll make a CR for this.

> This is not the only change I noticed.  Before I ran into
> the problem that I used to feed "define-type"s into the compilation
> via the use of -extend.  Beware: this doesn't work anymore.
> One has to put them into a separate file and (include "typedefs.scm")
> everywhere they are use.

Can you give an example?  Maybe file a proper bugreport in Trac, so we
can keep track of this issue.

> While I can live with the latter option, frankly I liked the former
> better.  Why may I not pass them via -extend anymore?  Or, for that
> matter more worrying the question is: what else from my -extend
> stuff is no longer visible as it was before?

I don't remember this being a change we made.  Perhaps it's a regression.

Cheers,
Peter
-- 
http://www.more-magic.net