Re: Hobbit - some notes

[Bernard URBAN <Bernard.Urban@meteo.fr>]

>>>>> "Miroslav" == Miroslav Silovic <silovic@zesoi.fer.hr> writes:

    Miroslav> I just tested hobbit and I must say I'm quite impressed
    Miroslav> (after having to port nontrivial software from SCM to
    Miroslav> Guile, I know the feeling). Here is a short wishlist -
    Miroslav> things I'd really like to see; not that I'm in any
    Miroslav> hurry:

    Miroslav>   1) Support for uniform array literals. These look like
    Miroslav> #i(1.0 2.0) etc., and would make compilation of my
    Miroslav> software MUCH easier, as I used these a lot. Even
    Miroslav> scm_eval_string("#i(1.0 2.0 3.0)") with the result
    Miroslav> stored in a variable would be great, as it'd be done
    Miroslav> only once.

Hobbit priorities are currently to be R4RS compliant, not guile or SCM
specific. But your suggestion seems easy enough to be implemented to
have some chance to be available in a (near) future release.
So it is added to the TODO list.

    Miroslav>   2) not trying to compile ice-9 : ice-9 currently seems
    Miroslav> to be in state of flux and old version of ice-9 that
    Miroslav> arrives with hobbit breaks the new snapshots (perhaps
    Miroslav> configure --disable-ice-9 ?)

This option will certainly be available in next release.

    Miroslav> 1) is the only feature I /really/ miss (and I need it to
    Miroslav> use hobbit properly). Apart from this, this compiler
    Miroslav> really works for the things I tested it with (*non*
    Miroslav> trivial programs).

I still consider it as beta software.

--

B. Urban