home *** CD-ROM | disk | FTP | other *** search
- Okay, here's the updated verion of the PC Angband executable. It is
- 100% savefile compatible with the previous executable. This update
- corrects a number of problems, including:
-
- * A serious stack-overflow problem, which could potentially (but
- very unlikely) could mess up your CMOS settings (or worse) by
- currupting data. To insure that this problem doesn't occur again,
- this update has stack-checking code compiled in. This slightly
- increases the size of the executable, and maginally slows the
- performance of Angband, but it protects your system from any
- problems related to stack overflow. If Angband suddenly exits,
- with the message "stack overflow", please let me know ASAP so I
- can look into it. This fix alone is important enough for you to
- get the updated executable.
- * Monochrome monitors showed dark-gray monsters and items as black
- spaces. There is now an option (accessible by the '=' key) to
- turn off color.
- * Scrolls and Staves of Darkness will now be properly IDed when
- used in a room.
- * The Potion of Self Knowledge will now correctly indicate SLAY
- EVIL.
- * Minor change in potions code now means several potions will ID
- correctly when used.
- * HPs for some high-level unique monsters have been fixed. (It was a
- "wrap-around" problem with the c_list[] array of monster info).
- * Experience gained from high level spells has been fixed.
- * Monster pits are no longer always considered "special". As you
- get deeper and deeper, monster pits are less likely to give you
- the "special" message. This was how the Unix code worked, but a
- 16-bit wrap-around caused all monster pits to be special on the PC
- version.
- * This version seems to require slightly LESS free memory than the
- previous version... My estimate of memory requirements is now
- 500-505k, instead of 515-530k. I'm a bit curious why it needs
- less memory than before. Maybe my overhaul to fix the stack
- overflow made it more efficient, but more likely it's because
- I've found a more exact way to measure memory requirements. (I
- kept running "command" to enter more and more subshells. Each
- one ate about 3k. Then I ran 'mem' to see the current free memory,
- and tried running Angband. If it didn't run, I exit-ed one of the
- subshells and repeated from 'mem'. When it ran, I made the current
- free memory my estimate, give or take a couple of k).
-
-
-
-