Re: 1.6.0 actions from RELEASE nearly finished.

[Dirk Herrmann]

On 28 Apr 2001, Rob Browning wrote:

> I've nearly finished taking the actions called for in RELEASE for the
> 1.6.0 release.  However there are a few things that I wasn't sure
> about:
> 
>   * Deprecate `read-only-string?'.
> 
> Is this something that should be done now, or for 1.8, or is it
> something that has prerequisites that haven't been satisfied yet?

I hope that all prerequisites are met, but:  Although I have changed a lot
of guile to get rid of dependencies on read only strings, I was not the
one who decided for their deprecation.  Thus, for the (somewhat political)
question if we should remove them for 1.6 (the version formerly called
1.5, right?) you better contact Mikael.

>   After signal handling and threading have been fixed:
>   - remove the code corresponding to GUILE_OLD_ASYNC_CLICK and the 
> corresponding
>     GUILE_OLD_ASYNC_CLICK macro.
> 
> Has any of this been done yet?  If not, is it a 1.6.0 goal, or is it
> fine to push it off for later?

Again, I think, this is better answered by Mikael.

>   - Q: Was SCM_FLOBUFLEN only deprecated publically, or was it supposed
>     to be removed from numbers.c as well?
> 
> I removed all the deprecated macros and functions from numbers.h, but
> that left numbers.c broken.  To fix it, I re-enabled the definition of
> SCM_FLOBUFLEN in numbers.c, but is that the correct long-term fix?

Yes, it is.  Maybe SCM_FLOBUFLEN should in the same step be renamed to
FLOBUFLEN, since I seem to remember that we are planning to use the SCM_
prefix for externally visible stuff only.

>   - remove code related to the name property of hooks.  Also, check init.c,
>     since the dependency between hooks and objprop will then be eliminated.
> 
> Has this been done?  If not, is anyone planning to do this soon, or
> should these modifications just become a 1.8.0 release action?

If you have removed SCM_HOOK_NAME and scm_make_hook_with_name from
hooks.[ch], everything should be done.  What's left to do in init.c is to
remove the comment that talks of the dependency between scm_init_hooks and
objprop.

Best regards,
Dirk Herrmann