urgent bug report/help requested (resent)

[Derek Atkins]

Quoting Thomas Bushnell BSG <tb at becket.net>:

> This is extremely odd.  Adding (require 'hash-table) in
> src/scm/main.scm does not change anything.  Adding it into the guile
> slib glue code for slib does: and then we get random reports of
> "sprintf" being unbound.  (Which is defined by slib's printf feature.)
> Surely I could then define that.  We could load *all* of slib on every
> inclusion of it.  But surely that is not the correct solution.

What do you mean by "adding it into the guile slib glue code for slib?"

> A guile session using the new things does not report any problems:
>
> $ guile
> guile> (use-modules (ice-9 slib))
> guile> (require 'hash-table)
> guile> hash-for-each
> #<procedure hash-for-each (proc ht)>
> guile>

This is certainly very odd..  I wish I understood gnucash's guile code
better than I do.  This seems to be a pretty esoteric guile environment
issue, I think.  :(

> The relevant Debian changes provoking this are:
>
>  Switch from slib-3a1 to slib-3a2.
>  Have guile load the slib guile.init file instead of its own badly
>    hacked version.

Hmm..  Now that you mention it, I think I was having trouble on Athena
with 3a2 as well when I was trying to build 1.8.12 there -- I had to
back out to 2d6 to get it working.   *ponders*

There's clearly a bad slib interaction going on..    :(

-derek

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available