[1091] in Coldmud discussion meeting
[COLD] Genesis patch 14 and ColdCore 3.0 ALPHA3
daemon@ATHENA.MIT.EDU (Sun Sep 22 12:07:11 1996
)
Date: Sun, 22 Sep 1996 09:57:34 -0600 (MDT)
From: Brandon Gillespie <brandon@glacier.cold.org>
To: coldstuff@cold.org
Both are available now. Genesis Changes (Note the function name changes):
* at one point I had an odd idea to allow people to submit bytecode
as well as lists of strings. I adjusted the method handling
functions in anticipation of this, and later realized what a
nightmare this would be. To cleanup the changes I have made the
following alterations:
compile() => deleted--nobody uses this anyway
decompile() => list_method()
add_method()
del_method()
get_method() => method_bytecode()
The only things this will effect are methods which call decompile(),
which will need to be changed to call list_method() instead. If
you used get_method() change it as well (most people did not know
it existed).
* Changed configure to read the module configuration from
src/modules/mod.conf, rather than having it hardcoded (Brandon)
* coldcc now decompiles with object numbers by default, not mapping
everything to a name (it still binds the name). You use '-#' to
turn off objnum printing over objname (i.e. -# to get the older
behaviour). Objnums are much faster than names--this is noticed
most on larger db's (10MB+ text) (Brandon)
* after much trevail a small but impacting db corruption bug was
found (in regard to floats). (Miro)
* Fixed a ~fpe bug reported by Brad
* Added ~inf which can be thrown if the resulting value is Inf or NaN
* removed descendants() -- this is better implemented in ColdC
due to its intensive and laggy nature (you can pause()) (Brandon)
* added task_info() -- make sure to bind it to something (Brandon)
* new function dict_values() (Brandon)
* SHA-1 encryption module provided by Jon Callas <jon@worldbenders.com>,
written by Paul C. Kocher (read the copyright in shs.c for full
information). Changed crypt() to use SHA-1 encryption based off
FreeBSD's MD5 crypt(). Older DES encrypted passwords can still
be matched with 'match_crypted()' (it hooks into the system's
crypt() if it exists). (Brandon)
* fixed a leak in method_info() (Miro)
* optimized fgetstring() (Brandon)
* allow "\t" like "\n" etc, when converting to buffers, need to
eventually change this to be the actual characters. (Brandon)
* changed fread() on text to eat all meta characters (it converts
\t to a space). (Brandon)
Enjoy;
-Brandon Gillespie