Start to solve the problem by calling md_init() to set _fmode. Now
reading a savefile only triggers an error handler and causes an abort
instead of a crash.
This is done by make, in the 'docs' target, which is now part of the
normal build process.
Unfortunately, not all the games include troff sources. Getting decent
HTML output from groff is still a difficult process which will not be
attempted at this time. There are a few bugs in the 'install' and
'uninstall' rules.
Not to mention that the documentation is sometimes inaccurate.
A custom autoconf macro searched for the curses library, and provided an
option to use ncurses instead of a (presumably deficient) curses
implementation. Unfortunately, some of the Makefiles ignored the
search's results.
Now that this is fixed, building against pdcurses should be easier too.
In wanderer(), if rp == hr, the loop could test ch before its value has
been set. Initializing ch to '-' now causes the loop to repeat until
a location is found.
curses.h is already included in rogue.h and doesn't need to be included
twice more in mdport.c.
Also set flags for some features when they are present.
The implementation was copied from rogue4. Using fdopen() is necessary
because the scorefile needs both encread() and encwrite(). For some
reason I have failed to discover, one of them uses FILE *'s and the
other uses file descriptors.
If SCOREFILE is not defined, roguehome() is called to find a directory
for the score file. It copies up to PATH_MAX-20 bytes from an
environment variable to a static buffer. Later these are strcpy()'d to
scorefile, which is of size LINLEN. Unfortunately LINLEN is 80 and
PATH_MAX is at least 256. On Linux, it happens to be 4096.
I haven't yet managed to crash or exploit it, but there are surely no
beneficial consequences, so roguehome() has been modified to check the
length, and the string it returns is also checked in main().
md_shellescape() sets SIGINT and SIGQUIT to be ignored, storing the
previous handlers, and restores them after the shell exits. But it
mixed up the two handlers.
Since the signals were usually handled by the same function, this fix
doesn't have much effect, but anything that makes signal code less
confusing is a good thing.
The player name is stored in whoami[], which is length 80 in most games
(1024 in rogue5). Only the first 10 chars were used to create
file_name, because that buffer is the same length. Increasing the size
of file_name to 256 permits using all of whoami.
The name is also no longer truncated to 20 chars when writing the log.
All games should now be able to handle 79-character names without
collisions. Anything more would break save compatibility.
Some .o files need to be rebuilt if config.h changes. Adding it to the
list of headers may still fail to solve the problem, because some of
the Makefiles use implicit rules or do not list dependencies properly.
In all games, rs_write_room_reference() stored -1 for a nonexistent
room, but rs_read_room_reference() did not check for out-of-bounds
values, leading to pointers to rooms[-1], which sometimes caused
crashes. rs_read_room_reference() has now been modified to use NULL
instead.
Some of the games required further changes to replace NULL with the
pointer to the actual room. Others are capable of handling NULL for
objects not in any room.
Super-Rogue, like Rogue V4, stored data of machine-dependent length in
the savefile, to prevent cheating. This made saved games non-portable.
Also deleted was a check that used this data, and prevented restoring
savefiles from backup.
This change BREAKS SAVEFILE COMPATIBILITY, but old files can be
converted by removing the block at offset 0x1e with length
sizeof(ino_t) + sizeof(dev_t) + 2 * sizeof(time_t). That seems to be
0x14 on i686 and 0x20 on x86_64.
