c.lib 1.215 (26.6.2017)

- Added -fno-aggressive-loop-optimizations option when building crtbegin.c
  to work around constructor/destructor hack with GCC 5.4.0 on AmigaOS 4.

- Added -fno-builtin option to fix conflicts with builtin memset()
  with GCC 5.4.0 on AmigaOS 4.


c.lib 1.214 (27.4.2017)

- Added integer overflow test to calloc().

- Tiny change in getopt_long() so that the value pointed to by longindex
  is always initialized to an invalid index position (that being -1),
  instead of 0. The value of 0 can break some shell commands, most notably
  GNU wget.


c.lib 1.213 (4.12.2016)

- Added the __decay_unused_slabs() function which brings all currently
  empty slabs which are still protected from reuse closer to getting
  reused or released.

- The slab-test program now exercises the memory allocation functions
  to a greater degree. Memory is allocated in random chunk sizes,
  the allocations are resized (to other random chunk sizes),
  33% of all allocations are randomly freed, empty slabs readied for
  reuse then discarded. The output in JSON format now shows a bit
  more information as to what is being done.

- Rewrote __get_slab_stats() to use setjmp() and longjmp() in the
  print() callback invocation.
  
- __get_slab_stats() now reports how many times a slab was reused
  after having stuck around in the "empty slab" list.
  
- Changing the slab size through an environment variable is now
  a feature of the debug build.

- Small changes to allow the library to be built with SAS/C again.
  This includes adding code to disable/re-enable profiling,
  fixing "stdlib_profile.h" and updating the smakefiles.

- Still not sure what it does, but _CXV45 now sits along with _CX25
  and _CX35 in "sas_cxv.asm". "sas_cxv54.asm" is not needed any
  more.

- Found the last use of MEMF_PRIVATE which should have been compiled
  only for the OS4 version.


c.lib 1.212 (27.11.2016)

- Unused slabs which get recycled are no longer reinitialized from
  scratch if their chunk size matches what the allocator needed.
  If the chunk size matches, the list of available chunks is
  left unchanged, and just the various counters are reset.

- Added __get_slab_stats() function.

- Added support for global __slab_purge_threshold tuning variable.


c.lib 1.211 (23.11.2016)

- Added more consistency checking to the slab allocator, which is
  built if DEBUG is defined in "stdlib_slab.c".

- Memory allocations are no longer guaranteed to be aligned to
  64 bit word boundaries. In fact, this has not even worked
  reliably in the past 10 years.

- Memory allocation request sizes are now rounded to multiples of
  32 bit words (the size of an address pointer) instead to the
  size of a 64 bit word.

- Reduced the memory footprint of the memory allocation management
  data structures by reusing the most significant bit of the
  memory allocation size. This allows many more allocations to fit
  into the 32 byte chunk slabs, but limits the maximum memory
  allocation size to a little less than 2 GBytes.

- Added integer overflow checks to the memory management code.

- Reduced the memory management overhead further. This cuts an
  additional 8 bytes per allocation, unless neither the slab
  allocator nor memory pools are available. With this reduction
  the slab allocator is able to use 16 byte chunks, which cover
  memory allocation requests of 1..8 bytes.

- Fixed a bug caused by returning an allocation back to a slab
  which passed the wrong pointer.


c.lib 1.210 (22.11.2016)

- Added __get_slab_allocations() function which will report information
  about each memory allocation made by the slab allocator which does
  not come from a slab.

- If the first slab in the list of slabs which share the same chunk
  size has no more room, it means that all other slabs following
  it have no room either. This speeds up the test to find a slab with
  free space, which can now abort and directly proceed to allocate
  memory for a new slab.

- If an empty slab's decay count hits zero, it is moved to the front
  of the empty slab list to be reclaimed more quickly.

- Allocations made from the slab now carry a pointer back to the
  slab which they are a part of. This speeds up deallocation but
  has the downside of making the smallest usable slab chunk size
  64 bytes, which is double what used to be the minimum before.


c.lib 1.209 (21.11.2016)

- The maximum slab size is now 2^17 bytes (= 131072). If you request
  a slab size larger than this, you will get slab sizes of 131072
  bytes instead.

- Enabling the memory management debugging code no longer produces
  compiler errors.


c.lib 1.208 (19.11.2016)

- Updated <stdlib.h> with new functions and data structures for
  use with the slab allocator.

- Added __get_slab_usage() function which can be used to query
  the slab allocator memory usage at runtime.


c.lib 1.207 (18.11.2016)

- Added a slab allocator which replaces the use of memory pools or the
  plain AllocMem() operations, respectively. In order to activate the
  slab allocator, choose a slab size (e.g. 2048 bytes or 4096 bytes)
  and declare a global variable like this:

     ULONG __slab_max_size = 2048;

  Memory allocations smaller than the slab size will be made from
  "slabs", i.e. large chunks of memory of the given size. Larger
  allocations will be managed separately.


m.lib 1.206 (24.4.2015)

- The fscanf() family failed to parse and convert %f parameters correctly
  if the respective number did not begin with a digit, but a decimal
  point. Hence ".7" would not be processed, but "0.7" would.

c.lib 1.206 (24.4.2015)

- Reworked the __putc() and putc() macros to reference the 'c' input
  parameter only once, and to be free of side-effects when tinkering
  with the buffer position.

- isatty() had the __fd_lock() call in the wrong place, which could have
  led to cleanup problems later.

- The close action in the stdio, socket and termios hook code now
  also zaps the fd pointer itself after cleaning up the file descriptor
  table entry.

- Removed the remains of all the stack extension and stack overflow/underflow
  checking code. It never actually worked. The bit that does work is the stack
  usage measurement code, plus the bit that sets up the the custom stack
  according to local setting or by calling a query function.


c.lib 1.205 (21.8.2010)

- Added dlclose(), dlerror(), dlopen() and dlsym() functions, which are
  available only under OS4. There is a variant of dlopen() in libunix.a
  which will perform a path name conversion. Note that these functions
  will not work in the thread-safe variant of the library because it
  would be unwise to tinker with the currently running program's binary.

- Added support for ELF shared objects and libraries. This is implemented through
  constructor/destructor functions, which means that you can use this
  functionality even in Amiga Exec style shared libraries, with the proper
  library initialization code to invoke the constructor/destructor functions
  for you.

- Updated uname() to recognize AmigaOS 4.1.

- The translation from Unix to Amiga path names did not properly process
  multiple occurences of "/./" in the path name. Thanks go to Steven Solie
  for finding the issue.

- The detection of "/./" and "/../" patterns in Unix path names to be
  translated into Amiga path names did not test if it was overrunning
  the end of the string.

- If strcmp(), strncmp() and memcmp() detect a pair of different
  characters, then the function result must be calculated as if the
  characters were of type "unsigned char". This is a requirement
  according to the ISO 'C' (1994) standard. Thanks go to Georg Steger
  for finding the issue.

- The definitions for INT_MIN, INT_MAX and UINT_MAX in <limits.h> no
  longer use long integer types, as prompted by Steven Solie.


c.lib 1.204 (11.11.2008)

- The memory allocated by malloc() and friends is now of type MEMF_PRIVATE
  under OS4 and beyond. The AmigaOS 2.x/3.x compatible code will still
  use MEMF_ANY in the same situation, though. Other uses of MEMF_ANY have
  been replaced as well where MEMF_PRIVATE would have made better sense.

- I/O buffers allocated are now aligned according to the CPU cache line size,
  if the operating system can supply such detailed information.

- unsetenv() now returns a status value.

- Corrected the function prototype for wcspbrk().

- Added function prototypes for mbrtowc_l(), wcscoll_l(), wcscspn() and wcsrchr().


c.lib 1.203 (28.4.2008)

- Added 68k stubs to amiga_rexxvars.c for the OS4 build to use. The new code now
  works just about exactly like the amiga.lib RVI code used to do, which means
  that there is no artificial length limit for the number of characters a string
  retrieved may have, and the code is largely reentrant.

- To work around a bug in the Roadshow TCP/IP stack (since fixed), the waitselect()
  function now substitutes a 10 microsecond timeout for a zero length timeout.

- isatty() no longer crashes if passed a socket rather than a file.

- Rewrote the GetRexxVar()/SetRexxVar() functions (mostly) in 'C', using available
  code as a reference. The 68k stubs currently only work for the 68k version of
  the library, and a solution for the OS4 build still needs to be found. The code
  is currently untested, but it should be complete.

- The printf() family now ignores argument specifications, as in "%2$d %1$d",
  which are used for localization on some platforms. This is a non-standard
  feature and the way clib2 treats it for now is just intended to avoid
  trouble while performing the conversion.

- [tboeckel]: when compiling amiga_rexxvars.c for m68k/OS3 the GetRexxVarFromMsg()
  and SetRexxVarFromMsg() can no longer cause linkage errors if the header files
  predate the SDK for OS 3.5.

- [jlangner]: the log() and log10() functions always returned -inf() even if the
  arguments was within the valid range for a log() function. Using DBL_EPSILON as
  the threshold was not correct as EPSILON is 2.2204460492503131E-16 whereas
  values like 1E-200 are still valid double values for a log().


c.lib 1.202 (16.1.2007)

- Added llrint() function contributed by Henning Nielsen Lund. Thank you
  very much!

- <unistd.h> now also include <stdio.h>, so that the SEEK_SET, etc. macros
  are defined for lseek() to use.

- Added a wrapper function which handles the thread-safe stdio stream
  resolution.

- In tcflush() a break signal can no longer cause the read flush loop
  to be quit with two semaphores still locked.

- In __obtain_daemon_message() the test to verify if the bsdsocket.library API
  would support the server API functionality checked the wrong feature. Fixed.

- Switched over the fd->fd_DefaultFile references to fd->fd_Socket where
  sockets are used rather than file handles.

- Added functions which modify the callback function and the userdata pointer
  stored in a low level unbuffered file/socket data structure. These function
  perform the proper locking and are thus safe to use in a thread-safe environment.

- The low level unbuffered file/socket now has a public equivalent, which
  is defined (along with the typedefs and flags) in <sys/clib2_io.h>. Functions
  for tinkering with it are still to come.

- The math kernel code no longer uses its own private scalbn() function.

- Added a function prototype for the _exit() function. Note that _exit() is
  not an ISO 'C' function.

- Corrected the getopt() function prototype, as prompted by Henning Nielsen Lund.

- The printf() family no longer adds a 0 or 0x prefix if the alternate
  conversion modifier is present for the %o and %x conversions and the
  value to be converted is 0 already. Put another way, printf("%#x %#o",0,0);
  now prints "0 0". This required another change so that %p always includes
  the 0x prefix even if the pointer involved is a NULL pointer.

- readlink() no longer sort-of-works for files and directories. It now only
  works for soft linked objects and returns an error for everything else.
  This is based upon a fix by Peter Bengtsson. Thank you very much!

- Moved the lstat() local Lock() function into its own separate file.

- uname() now returns correct and robust information for OS version
  numbers > 36. This integrates a fix by Peter Bengtsson. Thank you
  very much!

- Moved the crtbegin.o/crtend.o files out of the link libraries. Moving
  them in was intended to work as a fix for the shared library build, but
  now it seems that this has to be done at the link stage through the
  GCC specs file...

- Integrated a fix for __rem_pio2() which affects sin(), tan() and cos(),
  contributed by Steven Solie. Thank you very much!

- The internal 'struct fd' file descriptor table entry data structure
  now has a user data field entry.

- Rearranged the contents of the 'struct fd' file descriptor table entry
  data structure in preparation for making it public. Also added a version
  field so that user code can handle changes to it gracefully. The default
  file is no longer a BCPL pointer to a file handle by default, but
  both a BPTR and a socket identifier, wrapped into a union.

- Added experimental tilde expansion in Unix path names. This still needs
  some more work.

- __get_default_file() called __fd_unlock() without having called
  __fd_lock() first. Ouch.

- Removed an unnecessary pair of __fd_lock()..__fd_unlock() calls from
  ttyname_r().

- The libunix.a unlink() function is now reentrant, or at least thread-safe.

- You can now make unlink() stop after a failed deletion attempt which
  failed because the object to be deleted was reported as being "in use".
  The libunix.a variant defaults to report the deletion to have succeeded
  under these circumstances and later tries to delete the files marked
  for deletion. See <dos.h> for a brief documentation of how to change
  the behaviour.

- basename() and dirname() can no longer return NULL. They truncate the
  resulting path name instead. This is done so because some code that
  calls basename() or dirname() does not check if the function's return
  value is NULL.

- The SetOwner() fall-back code for Kickstart 2.04 was passing the wrong
  parameters to the file system. The first (dp_Arg1) should have been
  zero. Ouch.

- basename() is not supposed to modify the string it is passed and should
  return a pointer to a string which can be modified. Now it does. Same
  thing for dirname().

- asctime_r() now returns NULL if the buffer is too short to hold even a single
  byte of data.

- ttyname() now calls ttyname_r(). Also, the libunix.a version of ttyname_r()
  will produce "/CONSOLE" rather than "CONSOLE:".


c.lib 1.201 (21.9.2006)

- If defined, the local environment variable "DISABLE_COMMANDLINE_WILDCARD_EXPANSION"
  will disable expansion of wildcard patterns passed on the command line.
  Note that if the variable is not set then the global variable
  '__expand_wildcard_args' will provide the defaults for the switch that
  controls whether the wildcard expansion takes place. And after the
  environment variable has been checked, the '__expand_wildcard_args_check'
  function pointer can still be used to override the switch.

- fstat() now works with "NIL:" and "/dev/null", respectively. Not that
  it returns much useful information, though.

- The _PC_MAX_INPUT query for file handles now returns the default
  buffer size.

- Integrated the new OS4 build makefile collection, as contributed by
  Steven Solie. Thank you very much!

- Added the missing strerror_r() function.

- fpathconf() should work with the stdio streams, even in the thread-safe
  library version, again.

- Updated m68k specs file in /documentation to contain an own __CLIB2__
  define so that existing m68k compilers also have this define. In addition,
  the common "-noixemul" option can now also be specified but will do a NOP
  so that Makefiles sharing multiple runtime lib setups work without
  a warning.

- Added a new callback function which can be used in programs which want
  to avoid that the command line wildcard expansion takes place.  

- __get_default_file() now dynamically fills in file handles for the
  stdin/stdout/stderr streams if it's part of the thread-safe library.

- fpathconf() now checks if the file descriptor is really referring to a file.

- The termios hook entry code could file descriptor's embedded file handle
  rather than what the thread safe library had dynamically bound to the
  stdin/stdout/stderr streams.

- execve() now finds commands in the current directory again, even if you
  omit the leading "./" path name.

- The execve() code that looks for the command/script file now begins by
  checking if the file name includes path separators. If it does not,
  then the search for the command begins with the resident command list;
  otherwise the local directories are checked. The new code also properly
  cleans up after itself (a FreeDeviceProc() was missing) and the "PROGDIR:"
  lock now always refers to the directory in which the command/script
  file is found.

- Added the missing fdim() and fdimf() functions to the build
  makefiles. The OS4 build makefile was missing lrint() and
  lrintf(). Ouch.

- Added the POSIX exec() family functions, based upon code contributed
  by Henning Nielsen Lund. Thank you very much!

- atoll() no longer sets 'errno' directly but now calls __set_errno()
  instead.

- Added pathconf(), fpathconf(), sysinfo(), ftime(), ulimit(), getrlimit()
  and setrlimit() as contributed by Peter Bengtsson. Thank you very much!

- Added the missing S_ISSOCKET() macro to <sys/stat.h>. Note that this
  is not actually a POSIX feature.

- Added fmin(), fminf(), fmax() and fmaxf() to the build makefiles.
  Somehow I must have forgotten about them :-(

- select() and waitselect() can now be called without any file descriptor
  sets to work with, as some software does which uses select() in place
  of sleep().

- The <sys/ioctl.h> header file now includes both <sys/select.h> and
  <sys/filio.h> for better compatibility with the TCP/IP stack header
  files.

- Added the global variable __expand_wildcard_args which can be used
  to disable wildcard pattern expansion of command line parameters when
  linked against "libunix.a". Note that this has no effect on the "regular"
  libc.a behaviour.


c.lib 1.200 (17.4.2006)

- The default break signal mask (SIGBREAKF_CTRL_C) is no longer
  hard-coded. You can override it at link time with a different
  variable value for __break_signal_mask. This may have to be
  augmented by an API for changing the value.

- Added a waitselect() function which works very much like the
  bsdsocket.library/WaitSelect() function.

- Added <sys/time.h> include to <sys/socket.h> due to latest "TimeVal" change
  in the OS4 SDK. Otherwise "struct timeval" will not be defined at the time
  the <net/if.h> of the netincludes will be included by <proto/bsdsocket.h>.

- The strftime() hook function had the locale and character parameters
  switched, which made 'setlocale(LC_ALL,""); strftime(..);' unusable.
  Fixed.

- fchown() and chown() now accept owner and group ID values of -1, which
  indicate that the respective information should not be changed.

- The OS4 library build now includes the crtbegin.o and crtend.o object
  files in the libc.a library, which solves a problem with the thread-safe
  shared library support code.


c.lib 1.199 (6.3.2006)

- In <stdio.h> MAXPATHLEN is now equivalent to PATH_MAX (from <limits.h>),
  as suggested by Henning Nielsen Lund.

- Added lockf() support, as supplied by Henning Nielsen Lund.
  Thank you very much!

- symlink() now also works for absolute and relative links, and is
  part of "libunix.a". This patch was supplied by Henning Nielsen Lund.
  Thank you very much!

- The __main() stub function expected by the 68k GCC build is no
  longer part of "stdlib_main.c". This helps the C++ support, since the
  __main symbol definition no longer clashes with the definition in
  "libgcc.a".

- The printf() family stripped trailing zeroes from the integer part
  of %g output. Fixed.

- Moved an allocation size roundup operation in realloc().


c.lib 1.198 (11.12.2005)

- Added <fenv.h> for C99, and the (yet unimplemented) functions
  feclearexcept(), fegetenv(), fegetexceptflag(), fegetround(),
  feholdexcept(), feraiseexcept(), fesetenv(), fesetexceptflag(),
  fetestexcept(), fetestround() and feupdateenv().

- Replaced the old pow() implementation. However, powf() may need to
  be changed to set a domain error.

- In libunix.a malloc(), calloc() and realloc() no longer treat a
  request to allocate 0 bytes as an error, returning NULL. They all
  return a pointer sized memory chunk (= four bytes) initialized to
  NULL (= 0) instead.

- The alloca() implementation which allocates memory from the system
  rather than the local stack frame is thread-safe now. It also
  interacts with the realloc(), calloc(), free() and malloc() functions
  in that the alloca() cleanup routine is called once alloca() has
  done its job. If all the memory allocated through alloca() has been
  released no further calls to the cleanup function will be made.

- In the thread-safe library, realloc() permitted two different overlapping
  calls to succeed in trying to reallocate the same chunk of memory due to
  a race condition. Fixed.

- Added a new function __get_mem_stats() (the prototype is in <dos.h>) which
  can be used to query the current and maximum memory usage, counting the
  allocations performed through malloc(), free() and all other functions which
  use them.

- Added another function called __reset_max_mem_stats() which will reset the
  counters for "maximum amount of memory used" and "maximum number of chunks
  allocated" to the current figures for these values.

- Fixed the alloca() declaration in <stdlib.h> so that software which keys
  off the fact whether or not the alloca preprocessor symbol is defined
  will do the right thing.

- Added an optional call-back function which can be called if alloca()
  is about to return NULL, which some software does not consider.
  Rather than letting such software drop into an illegal memory access
  or worse, that call-back function will be invoked instead, which can
  print an error message and eventually call abort().


c.lib 1.197 (4.11.2005)

- Updated math_hypot.c to a newer (e_hypot.c 1.3 95/01/18) version from
  fdlibm which uses macros for manipulating the high and low words of a
  double, like the rest of fdlibm functions in clib2. The previous version
  would give bogus results when compiled with -O3 in clib2 which lead to
  "interesting" results (and lots of fun while searching for the problem)
  in Ghostscript. Wish I managed to track this down yesterday for 1.196
  release... <aantonijevic>

- Implemented atanh() and atanhf() which were not listed in the TODO
  file but were still unimplemented up until now.

- Replaced ldexp() and modf().

- Added __set_h_errno() and __get_h_errno() functions.

- The thread-safe library now tries to enable bsdsocket.library base
  sharing and attempts to make the TCP/IP stack call the local
  __set_errno() and __set_h_errno(), so that any Process calling
  the library functions will get a chance to see proper error
  codes. Whether any of these features could be enabled can be tested
  through two global variables __can_share_socket_library_base and
  __thread_safe_errno_h_errno which are defined in <dos.h>. Note that
  for both features to work you will need the Roadshow bsdsocket.library
  version 4.275 or higher.

- In <stdbool.h> the preprocessor definition should read
  "__bool_true_false_are_defined" rather than "__bool_true_and_false_are_defined".
  Thanks go to Peter Bengtsson for finding and fixing it.

- The network startup code no longer checks for the presence of a possible
  daemon startup message by default. You now have to enable this feature
  by declaring a global variable called "__check_daemon_startup" which is
  described in <dos.h>.

- Moved the code which rebinds the standard I/O streams to the server
  socket into a separate function which can be overridden by user code.

- Updated the new __obtain_daemon_message() function to call a
  bsdsocket.library API function to determine if what appears to be
  a valid daemon startup message is sound.

- Replaced the OS4 specific build makefile with an updated version
  prepared by Steven Solie. Thank you very much!


c.lib 1.196 (11.10.2005)

- Removed the various workarounds associated with <unistd.h>, required
  for building code that references the networking API. It is now possible,
  provided you build your code with an existing TCP/IP API header file
  set (such as from the "Roadshow" SDK) without running into data type
  or function prototype declaration conflicts. I have added local versions
  of <arpa/inet.h>, <netinet/in.h>, <netdb.h>, <sys/filio.h>, <sys/ioccom.h>,
  <sys/ioctl.h>, <sys/select.h> and <sys/socket.h> and updated <unistd.h>,
  <pwd.h> and <grp.h> accordingly. Note that some of the function prototypes
  will conflict with the bsdsocket.library/usergroup.library API, which is
  why you can disable the declarations through the __NO_NET_API preprocessor
  symbol: if defined, none of the function prototypes will be defined.

- Implemented lrintf(), lrint(), lroundf(), lround(), nearbyintf(),
  nearbyint(), remquof(), remquo(), roundf(), round(), tgammaf(),
  tgamma(), truncf(), trunc(). Sort of implemented fmaf() and fma(),
  which really ought to be done in "SIMD" fashion.

  This completes the "real" floating point math library (ignoring
  for a moment that the floating point environment code is still
  not implemented).

- accept() now calls the bsdsocket.library accept() function first
  and then hooks up the socket with the clib2 data structures. This
  makes it possible to have several Processes calling the accept()
  stub at the same time.

- When the main() function returns directly or indirectly by way of
  calling exit(), the IoErr() value set at this time will now be
  preserved until the program really returns control to the shell.
  This turned out to be easier to implement than I had originally
  thought, thanks to the fact that the atexit() functions are all
  invoked straight after the exit() call.

- strtoimax(), strtoumax(), strtod(), strtof(), strtol(), strtoll(),
  strtoul() and strtoull() now count how many digits were converted.
  If that number is zero, then the conversion is considered to have
  failed: 0 is returned and the 'ptr' parameter reference is initialized
  with a pointer to the string to be converted. This has two consequences:

     1) A floating point number that has no digits in the significand
        is no longer converted to zero. For example ".e10" can no longer
        result in a successful conversion.

     2) You can detect if a conversion is successful by comparing pointers
        after the conversion function has returned:

           char * str;
           char * ptr;
           long n;

           n = strtol(str,&ptr,10);
           if(n == 0 && str == ptr)
              printf("conversion has failed");

- Following detection of a stack overflow, the startup code eventually
  called _exit(). It should have called exit() instead.

- tmpnam() wrote more than L_tmpnam bytes to the name buffer. Also, the
  TMP_MAX value was off by one.


c.lib 1.195 (3.9.2005)

- In __time_delay() the CheckIO() test was wrong and should have
  tested for a request still in progress. Fixed.

- Modified select() to support plain files, too. The new code now
  compares the current file position against the current file size,
  if that file is on a file system, or simply checks if there is
  any data in the file (which works for the default PIPE: device).

- The raw file descriptor structure member fd_Position is now
  always updated, regardless of whether the file supports random
  access or not.

- The wildcard escape character used in __wildcard_expand_init() was
  wrong. It should have been "'" but it was "`". Fixed.

- The startup code now references the linker symbol generated for
  the user-supplied main() function.

- log(+/-0), logb(+/-0), log10(+/-0) now return -infinity.

- getcwd() now considers a buffer size of 0 an error, and it sets
  the errno code to ERANGE if the buffer is too small.

- With Unix path semantics enabled, rename() now fails if either
  the old or the new name are empty strings. Same thing for
  opendir(), utime(), unlink(), chown(), link(), readlink(),
  realpath(), chdir(), access(), stat(), rmdir(), lstat(),
  chmod(), statfs() and open()/fopen().

- Fixed several issues with the formatted output of strftime()
  for the "C" locale: "%p" could return " PM", "%c" is now
  equivalent to "%a %b %e %T %Y", "%x" is equivalent to "%m/%d/%y"
  and "%X" is now equivalent to "%T".

- The 68020/030/040/060 versions of the library now include the
  64 bit integer conversion code. The plain 68000 version does not.

- The 64 bit integer conversion code used in by the scanf() family
  was broken due to a misplaced overflow check, which had the cruel
  effect of triggering for 32 bit integers as well. Fixed.

- DoTimer() did not allocate the local MsgPort correctly. Fixed.


c.lib 1.194 (15.7.2005)

- Added getopt_long to libunix.a <tfrieden>

- Added new libprofile.a which implements gprof gmon.out output <tfrieden>

- Added the function profil() to libc.a <tfrieden>

- Added ftello() and fseeko() which both merely call the equivalent (at
  least for clib2) ftell() and fseek() functions, respectively.

- statfs() never stated that a disk was mounted for read-only access
  because the wrong InfoData structure member was tested. Fixed.

- lstat would overwrite the name parameter via ReadLink(..., name, ...)
  instead of ReadLink(..., new_name, ...) <tfrieden>

- Implemented exp2()/exp2f() and log2()/log2f() as suggested by
  Henning Nielsen Lund. Thank you very much!

- The pattern matching code which expands command line arguments, as part
  of libunix.a, now translates the "*" wildcard into "#?" unless you prefix
  it with a backtick ("'"), which is the wildcard pattern escape character
  used on AmigaOS.

- Repaired the pattern matching code which expands command line arguments:
  it no longer requires that any parameters are quoted and it is automatically
  enabled if you link against libunix.a, without libc.a standing a chance to
  accidentally override it.

- Fixed a bug in vfprintf that would surface when the buffer was enlarged
  if the pattern was too large to fit the static internal buffer <tfrieden>

- Added file system names to statfs <tfrieden>

- Moved all the constructor code out of "stdlib_init_exit.c" and into the
  files which initialize global data, such as the new "stdlib_program_name.c"
  and "stdlib_malloc.c".

- Until I can find a way to invoke it from within the library, the
  __machine_test() function is no longer invoked.

- Tagged global data with NOCOMMON attributes.

- Added the __lib_init() and __lib_exit() functions, which are part of the
  thread-safe library and which can be used to hook up clib2 with standard
  Amiga shared library/device code. Some documentation on how to use them
  can be found in the <dos.h> header file.

- Added a complete shared library skeleton example to demonstrate how the
  clib2 shared library initialization/cleanup functions should be used.
  This is a dual-build library which will both work with the plain 68k
  AmigaOS and the PowerPC native AmigaOS4. The example source code can be
  found in the "skeleton_library" subdirectory.

- The thread-safe version of isatty() should now work for stdio
  file descriptors, too.

- Retrofitted thread-safety into the termios code.

- The socket/usergroup API function and data type definitions that used to
  be defined in <unistd.h> have been moved into a separate header file
  <sys/clib2_net.h> which, for backwards compatibility, can be include
  from within <unistd.h> if the preprocessor symbol __USE_CLIB2_NETLIB
  is defined early on. Eventually, <sys/clib2_net.h> should be removed
  from the general header files and become a local library build header
  file altogether. Note that in this case you would have to build your
  own software using common TCP/IP SDK header files.

- <sys/types.h> now also includes <time.h> and <stddef.h>.


c.lib 1.193 (4.6.2005)

- fpclassify() now returns FP_ZERO both for 0 and -0.

- nan() and nanf() now return quiet NaNs.

- Added internal __inf() and __inff() functions.

- strtof() now calls nanf() and __inff(), respectively, to produce
  the special floating point values for nan/inf/infinity.

- strtod() now calls nan() and __inf(), respectively, to produce
  the special floating point values for nan/inf/infinity.

- The scanf() family now calls nan() and __inf(), respectively, to
  produce the special floating point values for nan/inf/infinity.

- Added %D, %e, %F, %g, %G and %h for strftime(); repaired %I.

- Documented __strip_double_slash() and plugged in a memmove()
  in place of the copying loop.

- Modified __translate_unix_to_amiga_path_name() and
  __translate_amiga_to_unix_path_name() to not to call strlen()
  on the results of __strip_double_slash().

- For %C strtime() should return the century number, not the
  year number mod 100.

- The record locking code in fcntl() is no longer part of the
  regular libc.a, but only of libunix.a.

- Changed the definition of the D_S() macro to cast the pointer
  address to an unsigned 32 bit integer.

- Modified the locale-aware isalnum(), isalpha(), iscntrl(),
  isdigit(), isgraph(), islower(), isprint(), ispunct(), isspace(),
  isupper(), isxdigit(), tolower() and toupper() functions
  to clamp the input parameter to the "unsigned char" range before
  it is submitted to the respective locale.library function. This
  should be in sync with what the specs demand, which state that
  if the input parameter is neither EOF nor in the range of an
  "unsigned char" variable, then the results are undefined.

- ungetc() now returns the input character, clamped to an unsigned
  char value, upon success. Previously, if the input parameter was
  negative and not EOF, the result was identical to the input, which
  could have had negative side-effects.

-Added NAN, INFINITY, FLT_EVAL_METHOD, float_t and double_t to
  <math.h> (C99).

- Added skeleton code for acosf(), acosh(), acoshf(), asinf(), asinh(),
  asinhf(), atan2f(), atanf(), atanh(), atanhf(), cbrt(), cbrtf(),
  ceilf(), cosf(), coshf(), erf(), erfc(), erfcf(), erff(),
  exp2(), exp2f(), expf(), expm1(), expm1f(), fdim(), fdimf(),
  floorf(), fma(), fmaf(), fmax(), fmaxf(), fmin(), fminf(),
  fmodf(), frexpf(), hypotf(), ilogb(), ilogbf(), ldexpf(), lgamma(),
  lgammaf(), log10f(), log1p(), log1pf(), log2(), log2f(), logbf(),
  logf(), lrint(), lrintf(), lround(), lroundf(), modff(), nearbyint(),
  nearbyintf(), powf(), remainder(), remainderf(), remquo(), remquof(), round(),
  roundf(), scalbn(), scalbnf(), sinf(), sinhf(), sqrtf(), tanf(),
  tanhf(), tgamma(), tgammaf(), trunc() and truncf(), to be filled in
  later...

- Implemented fmin()/fminf(), fmax()/fmaxf(), fdim()/fdimf() for C99.

- Ported acosf(), asinf(), atan2f(), atanf(), ceilf(), expf(), floorf(),
  fmodf(), frexpf(), ldexpf(), log10f(), logbf(), logf(), modff(), powf(), sqrtf(),
  scalbn() and scalbnf() for C99.

- Ported cbrt(), cbrtf(), erf(), erff(), erfc(), erfcf(), expm1(),
  expm1f(), ilogb(), ilogbf(), log1p() and log1pf() for C99.

- Ported cosf(), coshf(), sinf(), sinhf(), tanf(), tanhf()
  and hypotf() for C99.

- Ported acosh(), acoshf(), asinh(), asinhf(), lgamma(), lgammaf(),
  remainder() and remainderf() for C99.

- The scanf() family now supports character ranges for the %[
  conversion. Note that this is a non-standard feature!

- Integrated Peter Bengtsson's termios code. Thank you very much! I chose
  to add it to libunix.a rather than keeping it in a separate libtermios.a
  library.


c.lib 1.192 (12.5.2005)

- Changed how errors are detected, as returned by Write(), Read() and
  Seek(). Seek() is particularly challenging because the value it
  returns might be a valid file position and not an error.

- Replaced numeric function return codes of 0 and -1 with macros OK,
  SEEK_ERROR/ERROR to clarify the respective purposes.

- Changed how ftell() and fseek() are used, double-checking the return
  value and the errno code.

- The record locking semaphore could wind up getting added to the public
  list twice. Fixed.

- Fixed two linker errors which were caused by duplicate symbol definitions.

- Added code to the startup routine which allows you to monitor where a command
  was started from and which parameters it was invoked with.

- If fread()/fwrite() fail to read/write any data because either the number
  of records or the size of each record is zero, both now call clearerr() to
  avoid giving the caller the wrong impression that an EOF or error occured.

- The libunix.a flavour of system() no longer attempts to translate the name
  of a command unless it contains path separator characters.

- Added strtof(), llabs(), lldiv(), vsscanf() and vscanf() for C99.

- strftime() now supports %C, %n, %r, %R, %t, %T, %u, %V, and %z for C99.
  And it ignores the E and O modifiers.

- The printf() family now supports the %hh, %j, %t and %z modifiers and the
  %a/%A conversions for C99. The %j is treated like %ll; %t and %z are treated
  like %l. Also, the "infinity"/"not a number" signals now come out as the
  strings "inf" and "nan".

- Added HUGE_VALF to <math.h>.

- For the printf() "%a" conversion the exponent now comes out as a binary
  number rather than a decimal one. Now how odd is that?

- strtod() and strtof() now support "inf"/"infinity"/"nan"/"nan(..)" and
  hexadecimal floating point numbers, for C99.

- Added the fpclassify(), isfinite(), isnormal() and signbit() macros for C99.

- Reimplemented isnan() and isinf() as macros for C99. The corresponding
  functions will be dropped from the library. Note that the isinf() macro
  does not return -1, 0 or 1 like the old function did, but only 0 or 1
  depending upon whether the parameter represents +/- infinity or not.

- Added fabsf() for C99.

- The scanf() family now supports the %hh, %j, %t and %z modifiers and the
  %a/%A conversions for C99. The %j is treated like %ll; %t and %z are treated
  like %l. Also, the "inf"/"infinity"/"nan"/"nan()" keywords are processed.

- The strftime() %z conversion now prints the time zone difference as a
  "decimal" number. That is, if the difference is 5 hours and 30 minutes,
  then %z will now print "530" rather than "330".

- mktime() now handles one leap second gracefully.

- Added isblank().

- Added isunordered(), isgreater(), isgreaterequal(), isless(),
  islessequal() and islessgreater() to <math.h> for C99.

- The wchar_t type is now an 'unsigned short' integer (16 bits wide).

- Added PTRDIFF_MIN/PTRDIFF_MAX, WCHAR_MIN/WCHAR_MAX and
  WINT_MIN/WINT_MAX to <stdint.h> for C99.

- Added imaxdiv() and imaxabs() for C99.

- Added strtoimax() and strtoumax() for C99.

- Added nextafter() and nextafterf() for C99.

- Added copysign() and copysignf() for C99.

- Unless I missed something, clib2 should now be functionally complete
  with regard to C99, except for the floating point operations covered.
  These are a major challenge all by themselves, and I wonder both
  whether they are worth the effort and how one could implement them
  correctly.

- fflush() now consistently supports a NULL parameter, causing all
  streams to be flushed for which this behaviour is defined.

- The printf() family can now produce output for floating point
  numbers with more than about 77 characters, provided sufficient
  memory is available. C99 calls for a minimum of 4095 characters,
  but we're trying to allocate the space required dynamically.


c.lib 1.191 (9.4.2005)

- The name of the public record locking semaphore has to be preallocated
  for OS4 if the AllocSysObject() function is used to create it.

- Moved the signal block mask and the signal function table out of
  signal_data.c and into signal_raise.c where they are actually used.
  This makes it possible to override the default definition of the
  __check_abort_enabled variable in your own programs.

- raise() no longer resets the signal handler before it invokes the one
  currently configured. It merely blocks the delivery of the respective
  signal to prevent recursion.

- raise() now drops into abort() if a signal handler is set to
  SIG_DFL. The exception is in SIGINT delivery, which has the
  effect of printing a different termination message but otherwise
  program flow takes the same path as abort().

- Moved the __UtilityBase/__IUtility variable declarations into
  a separate file.

- Simplified the library/open close code in "stdlib_main.c".

- File descriptors produced by dup() or dup2() now work exactly like
  the original file descriptors they are duplicates of. I modified the
  function which maps file descriptor numbers to file descriptor
  table entries to return the table entries of the original files.

- In the thread-safe library, duplicated stdin/stdout/stderr
  descriptors now work like the "real" ones. This is true even if
  the "real" ones were closed and only their former aliases remain.

- Invoking fstat() on what maps to a con-handler stream now produces
  information identifying it as a character special file.

- Added more code and changes contributed by Peter Bengtsson, thank you
  very much! This includes the following:

  - Added SSIZE_MAX to <limits.h>.

  - Added <sys/uio.h>, readv() and writev().

  - Cut back the soft link resolution code in lstat().

  - In <fcntl.h> O_NDELAY is now an alias for O_NONBLOCK.

  - Added <complex.h>, carg(), cargf(), cargl(), cimag(), cimagf(),
    cimagl(), conj(), conjf(), conjl(), creal(), crealf() and creall().

  Note that the C99 support for the complex floating point data
  types is limited to GCC 3.x for now.

- Added va_copy() to <stdarg.h>.

- Added _Exit() to <stdlib.h>.

- Added <stdbool.h>.

- Added vfscanf() to <stdio.h>.

- The stdio locking in fcntl() wasn't working correctly. Fixed.

- Made the clearerr(), feof() and ferror() macros thread-safe.


c.lib 1.190 (25.3.2005)

- DoTimer() now calls AllocSysObject() rather than making up
  a MsgPort locally.

- The record locking semaphore code now builds a semaphore to add
  before it tries to find the public one in memory. That way, the
  code can spend less time in Forbid() state and, heaven forbid,
  refrain from allocating memory while in that state.

- Split the general stdio initialization/cleanup code from the
  initialization of the stdin/stdout/stderr streams.

- Moved the Workbench console stream initialization into the
  initialization code for the stdin/stdout/stderr streams and
  out of the program parameter setup.

- The current program name is now set up in the stdlib
  initialization function.

- Simplified the machine test code; moved the FPU check into
  the math initialization code.

- Added more safety checks to verify that file descriptor
  file handles are valid.

- Made the file descriptor checks in the fsync() and fdatasync()
  functions more robust.

- Cleaned up the 68k build makefile, so that the CPU and FPU
  tests and the error message display can run safely even
  on plain 68000 machines. This won't work for the 32 bit small
  data model, which implies 68020 code, but so there...

- Moved the CPU/FPU type tests into the respective linker
  libraries.

- Moved the data declarations out of math_data.c, stat_data.c,
  socket_data.c, dirent_data.c and stdio_data.c and into the
  code that initializes them.

- Moved a few __delete_semaphore() calls into conditional compilation
  sections where they should have been in the first place.

- Thanks to Jrg Strohmayer, the GCC library build now manages to
  invoke the library's constructor/destructor functions in a
  very particular order. That way, you can use constructor/destructor
  functions in your own code and not have them clash with the library's
  own functions.

- Reimplemented the constructor/destructor invocation code for GCC.
  Both the 68k and PowerPC platform now invoke them in the same order
  and the 68k code uses the designated invocation priorities. The
  PowerPC destructor function now sets up the exit() jmp_buf before
  the destructor functions are called.

- Added S_IREAD, S_IWRITE and S_IEXEC aliases to <sys/stat.h>.

- Moved data out of stdlib_data.c and into the code that references
  or initializes it.

- The stdlib constructor now performs the CPU/FPU compatibility test.

- Introduced new constructor types and changed the overall priority
  order.

- Switched over the startup code and the library itself to use
  constructor/destructor functions for initialization and cleanup
  purposes.

- The destructor function invocation code no longer calls
  setjmp(). This is now done within stdlib_main.c prior to
  calling the destructor function invocation code.

- Fixed the SAS/C destructor function prototypes and verified
  that the constructors/destructors are called in the proper
  order, too.

- Aliases of file descriptors are now using the signal semaphore
  of the original file descriptor.

- close() did not return 0 if the file descriptor in question
  was really just an alias. Fixed.

- Added a feature which makes it possible to have several clients
  use the standard I/O streams (stdin/stdout/stderr) and have these
  referring to their process' Input()/Output()/ErrorOutput()
  streams. This is intended to support the upcoming shared
  library feature.

- Turns out that the 68k GCC port does not sort constructor and
  destructor functions in any way at all. I reimplemented the
  entire library constructor/destructor functionality to use the
  same approach as libnix.

- ftruncate() ended up changing the current file position, contrary
  to what it is supposed to do. Fixed.

- fcntl() did not work for sockets. Fixed.

- Reorganized the local header files, removing redundancies.

- Dropped unused stub code which is now redundant because of
  the constructor/destructor mechanism.

- Moved the call chain printing out of stdlib_main.c and into
  separate files.

- Removed some more redundant data from stdlib_main.c.

- Added the first "real" C99 function: _Exit() ;-)

- assertion failures early on during program initialization
  should no longer spell big trouble on account of the stdio
  data structures possibly not being in a well-defined and
  initialized state.

- Turns out that the constructor/destructor calling sequence
  was still wrongish for the OS4 library. So I had to change
  it *again*. Which probably means that the 68k library will
  need further changes...

- Moved stdlib_main.o into the regular libc.a, at least for
  the 68k build. The PowerPC build may follow later, provided
  I manage to get the specs file fixed. Actually, stdlib_main.o
  is in the libc.a library already. Now about that specs file...

- Moved the check for the presence of an FPU into the
  math_init.c code. I am far from certain whether this will
  have the desired effect, though. Due to how the GNU ld linker
  works, libraries are scanned once only. And the FPU check will
  be pulled in only if something references the HUGE_VAL
  constant.

- Activated the dormant thread-safe standard input/output/error
  handling code.

- Small fixes to fcntl() and select() to cover the thread-safe
  fd->fd_DefaultFile == ZERO case.


c.lib 1.189 (5.3.2005)

- Rewrote the __translate_unix_to_amiga_path_name() function to
  translate patterns such as "foo/bar/../../baz" properly, and to
  use strlen() a lot less.

- Major, major changes! Moved most of the monolithic code out of
  the file descriptor hook and into the respective functions,
  such as dup2(), fchmod(), fchown(), fcntl(), fdatasync(), fstatfs(),
  fsync(), ftruncate() and lseek(). Code which is not strictly
  required will no longer find its way into your programs if you
  link with the updated library.

  NOTE: these changes require that the entire library is rebuilt!

- The buffered and unbuffered file hook code is now invoked through
  function pointers alone. The utility.library/CallHookPkt mechanism
  is no longer required.

- Moved the entire lseek() code relevant for files into the hook
  function.

- Simplified the close() function which now just calls into the
  hook code to perform whatever is necessary. The hook code is
  responsible for cleaning up after aliases, etc. This change in
  turn made it possible to greatly simplify the hook code for
  buffered files which now bypasses close/read/write/lseek and
  directly invokes the hook code for unbuffered files.

- Added various floating point constants to <math.h>, courtesy
  of Henning Nielsen Lund. Thank you very much!

- When using the wildcard expansion code for command line
  parameters (which is by default linked in with libunix.a),
  regular expressions can no longer prompt dos.library requesters
  to appear. However, to be on the safe side, if you are expecting
  to pass regular expressions on the command line, do not use
  the wildcard expansion code such as by overriding the library
  symbols with dummy functions such as are used in the file
  "stdlib_wildcard_expand.c".

- Added a new variable '__open_locale'which can be used to
  restrict all library functions to use the "C" language locale
  rather than the current system locale settings. In addition
  to that, two new functions __locale_exit() and __locale_init()
  can be used to close and (re-)open the system locale at a
  later time.

- Local ("static") functions are now identified by the STATIC
  qualifier. This was done in preparation for changes that will
  deal with global and local data and the issue of thread safety.

- Added stdio thread locking functions flockfile(), funlockfile(),
  and ftrylockfile().

- Modified the internal FILE structure to allow for thread locking.
  Note that this again requires that the library is rebuilt!

- Added or modified macros for getc_unlocked(), getchar_unlocked(),
  putc_unlocked() and putchar_unlocked().

- Added rand_r().

- Added flockfile()/funlockfile() wrappers around all stdio
  functions.

- Added more semaphore locking around the basic stdio, memory, locale
  and dirent data operations. That should do it! While the library is
  not reentrant (this is not ixemul.library) it should be thread-safe
  now. Thread-safe in the sense of POSIX 1003.1c-1995.

- The thread-safety code is now subject to conditional compilation.
  Both the library and the user code need to be rebuilt with the
  preprocessor symbol __THREAD_SAFE defined to get thread-safe
  code.

- Extended the thread-safety locking to the file descriptors.

  NOTE: these changes require that the entire library is rebuilt!

- The translation of Unix to Amiga path names now silently accepts
  absolute Amiga path names passed to it and will use them without
  changing them.

- Added atoll(), ffs(), ftw(), nftw(), lstat() and uname() code
  contributed by Peter Bengtsson. Thank you very much!

- Reworked the code that handles quoting for the wildcard expansion
  routine. We no longer allocate memory and then modify it, but
  call a function for each quoted parameter which does whatever is
  necessary.

- The shell command parameter parser now considers the non-breaking
  space character (ISO code 160) to be a blank space character, too.

- Moved the signal semaphore allocation/initialization/deallocation
  into a dedicated module. This also has the advantage that it's
  harder to break code by accidentally forgetting to call
  InitSemaphore() after having allocated the memory for it.

- Rewrote the code that allocates the file descriptor and file
  buffer tables so that all the memory allocations are in one
  place and it's possible to specify exactly how many table
  entries are required at a time.

- Creation and initialization of semaphores now uses the AmigaOS4
  specific functions for this purpose, if available.

- In the thread-safe variant, the library now tries to allow
  multiple concurrent callers to use the socket functions. Note
  that this works only with the Roadshow TCP/IP stack, and the
  results with other TCP/IP stacks are rather unpredictable.


c.lib 1.188 (7.2.2005)

- Folded duplicate code in "stdio_init_exit.c" into a common function.

- Simplified the code in "time_asctime_r.c" which builds the time
  string. It gracefully handles buffer sizes which are too short
  by returning an empty string.

- Moved the "tm->tm_wday" initialization out of the hook function
  in "time_strftime.c" since it was to be called only once anyway.

- Lost a few compiler warnings in "unistd_time_delay.c" and
  "time_gettimeofday.c".

- Folded duplicate code in "time_mktime.c"; also, errno is no longer
  modified unless the library is built with the "CHECK_FOR_NULL_POINTERS"
  option.

- We now allocate the AnchorPath used in the unistd_wildcard_expand.c
  code. Also, the contents of the AnchorPath structure are no longer
  modified between calls. MatchEnd() has to be sufficient.

- Moved redundant code out of the readdir()/opendir()/closedir()
  functions which is not required unless the code is built for
  Unix compatibility mode.

- Lost the __not_a_number and __infinity variables, including the
  code which initialized them.

- Reading/changing the errno variable is no longer done directly, but
  involves accessor functions.

- References to the HUGE_VAL quantity now involve an accessor function, too.

- Changed the manner in which the __huge_val constant is initialized by
  the __math_init() function. The new approach should be more portable.

- The library no longer sends ACTION_DISK_INFO packets to the console
  handler. The side-effects were too varied and irritating after all.

- Added the fsync() and fdatasync() functions and the <stdint.h> and
  <inttypes.h> header files contributed by Peter Bengtsson. Thank
  you very much!

- Tweaked the build makefiles to produce fewer meaningless warnings.


c.lib 1.187 (29.1.2005)

- The default console output window opened when a program is launched
  from Workbench would open and stay open. This was not intended to
  happen and is a side-effect of the new stdio initialization code which
  checks if the stdio streams are in non-blocking mode. Fixed.

- Moved the common DateStamp to time_t conversion code into a shared
  function.

- The fall-back function for converting time into a string in strftime()
  now calls itself for the "%c", "%x" and "%X" format specifiers.

- mktime() is supposed to convert the time specification, given as local
  time, into the number of seconds since January 1st, 1970, relative to
  UTC. This didn't really work up until now since the time value returned
  was given as local time.

- Plugged in a different algorithm for calculating the day of the week in
  strftime() and asctime_r(). This one isn't sensitive to the effects of
  adding/subtracting the local time zone.

- Changed the algorithm that calculates the number of days that have passed
  so far as used by the the __convert_time() function and the conversion
  code in strftime().

- Also changed the algorithm used by strftime() to produce the week numbers
  (the '%U' and '%W' format specifiers). The new method is much simpler
  than the old one.

- Made the code that converts a 'time_t' value into the 'struct DateStamp',
  as used by strftime() and utime(), into its own function. This also
  fixes a bug in the code strftime() would use which was completely
  unaware of the local time zone settings.


c.lib 1.186 (14.1.2005)

- Redirecting stderr to a file no longer has the effect of showing error
  messages and assertion failure notifications as requesters. The exception
  is in redirecting stderr to NIL: which will prompt the requester use.

- gettimeofday() now calls GetSysTime() rather than DateStamp() to obtain
  the current system time. This resolves granularity issues since the
  DateStamp() result was only accurate by 1/50 of a second.

- The "ptrdiff_t" definition in <stddef.h> now defaults to type 'int' rather
  than 'long int'.

- The "char" limits in <limits.h> are now set up according to the current
  compiler settings, which can either default to an unsigned or
  signed definition.

- Changed the rules again for the use of stderr/stdout redirection when
  printing error messages. It is always safe to redirect them now and
  no requester will appear unless you specifically set the value of the
  "__no_standard_io" variable to TRUE in your program.

- Removed a misplaced IsInteractive() from the stdio initialization
  function. Now this could have been big trouble...

- Removed tests for FileHandle->fh_Type != NULL which used to precede
  all IsInterative() tests. I verified that IsInteractive() will always
  return FALSE for NIL: type file handles.

- Dropped the special flag variable used by the abort() function that
  tracks whether or not console output is possible. We now use the
  global "__no_standard_io" instead.

- Made the <ctype.h> macros more robust.

- Removed the "NIL:" file handle tests preceding the Open("CONSOLE:",..)
  calls. As of Kickstart 2.x and beyond these are no longer a source of
  trouble.

- The V37/V40 compatibility code is no longer built for the AmigaOS4
  version of the library.

- Switched over the last use of DeviceProc() to GetDeviceProc(), etc.

- open() no longer examines a file after opening it in order to figure
  out whether read/write accesses are permitted. This decision is now
  for the file system to make.

- Whether or not stdio console streams are blocking or non-blocking
  is now determined at initialization time. The I/O mode is restored before
  the program exits. Previously, any changes to the I/O mode would persist.

- Lost some more code that is not required for AmigaOS 4.x and can be
  handled conveniently through conditional compilation.

- close() did not reset the non-blocking file property, as it should
  have. This only worked for files which were closed anyway, but not
  for the stdio streams. Fixed.

- Added a missing definition to stdio_init_exit.c which is part of
  the OS4 header files, but not of the older header file distributions.



c.lib 1.185 (2.1.2005)

- Moved the environment variable cleanup code into a destructor function.

- Fixed a typo in the hstrerror() function.

- The common error reporting function __show_error() could throw Enforcer
  hits if the program was not launched from Shell. Fixed.

- Moved the memory initialization and cleanup functions into the
  malloc/free code itself and updated the alloca code to do its
  own data management.

- Finally optimized the alloca() memory cleanup code.

- Tried to make it possible to have debug versions of the memory
  management code in the default link library along with the
  normal versions. Let's see how well this works out in the GCC
  version. SAS/C does not seem to like it...

- Changed the definition of alloca() for GCC, which now defaults
  to the built-in function.

- Updated the stdio.h, stdlib.h, string.h and unistd.h header files
  to declare function prototypes for the __MEM_DEBUG versions of the
  library functions only if that preprocessor symbol is defined.

- If a SIGINT signal is caught and processed by the default signal
  handler, the localized "*** BREAK" string will be printed rather
  than the built-in one.

- The data structure alignment (file I/O buffer) is now configurable
  at compile time. The default used to be 16 bytes, which is appropriate
  for the 68040/68060 but not for the PowerPC, which uses 32 or 128
  bytes per cache line.

- fwrite() now flushes the entire "buffer" for unbuffered files. The
  exception are "interactive" files such as console windows. For these
  line buffered output is used.

- Whether or not a file is bound to an interactive device, such as a
  console window, is now checked and remembered after a file descriptor
  has been associated with it.


c.lib 1.184 (28.11.2004)

- Added asctime_r(), ctime_r(), gmtime_r(), localtime_r() and strtok_r().

- Added stubs for the Rexx Variables Interface code that used to
  be part of amiga.lib. While comparable functionality is available
  in rexxsyslib.library V45, the new stubs might be helpful during
  porting. Care must be taken since these functions don't work
  exactly like the originals.

- Integrated strlcpy() and strlcat() which are intended to be safer
  replacements for strncpy() and strncat().

- The program's task priority is now configurable through an external
  variable '__priority'.

- The process name to be used when detaching can be configured through
  the new '__process_name' variable.

- The minimum required operating system version can be configured
  through the new '__minimum_os_lib_version' variable; a matching
  error message can be provided through the new '__minimum_os_lib_error'
  variable.

- The default console window specification can be overriden through
  the new '__stdio_window_specification' variable.

- The socket initialization code did not set up a reference to the
  'h_errno' variable correctly. This had the effect of making name
  and address resolution errors trash the 'errno' variable instead
  and leaving 'h_errno' always set to 0. Fixed.

- For sockets, ioctl() and fcntl() now interact on the FIONBIO/FIOASYNC
  requests (ioctl) and the O_NOBLOCK/O_ASYNC flags (fcntl).

- popen() now accepts "rb" and "wb" as mode parameters. However, "r+",
  "w+" and variants thereof are still unsupported due to the
  unidirectional pipe support in the standard "PIPE:" device.


c.lib 1.183 (13.11.2004)

- Cleaned up the OS4 build makefile, losing redundant libraries,
  adding more startup object code and ultimatively making the whole
  rebuild logic work again: if code changes and dependencies are
  set up correctly, it will now get rebuilt. Previously, such
  changes went unnoticed and you had to rebuild the entire library
  from scratch.

- Added stubs for CreatePort(), DeletePort(), CreateTask(), DeleteTask()
  and NewList() which have equivalents in exec.library V50 but for which
  it might be useful if ported code didn't have to reference these
  explicitly.

- mktemp() was broken in libunix.a with Unix path semantics enabled.
  This was because the name template was translated and translated
  back again, overwriting the translation buffer. This, funny enough,
  broke Samba's printing feature. Fixed by translating the name only
  before each test for "uniqueness" is made. The new code also handles
  empty "" templates gracefully, which was a problem with both the
  "standard" and the Unix path semantics flavour.

  Why is it that I find bugs like this always after having just
  released another library update?


c.lib 1.182 (8.11.2004)

- Changed the error abort condition for the %s conversion of the
  scanf() family. It now matches the abort conditions for all other
  conversions and no longer ignores whether any other parameters were
  converted before. This was a quirk in the older implementation.

- The scanf() family now accepts %E and %G in place of %f and %X in
  place of %x.

- Simplified the common code that fopen(), freopen() and fdopen()
  share and which has to figure out by looking at a file access
  mode specification which parameters should be used.

- Dropped error detection in the scanf() family. The EOF has to be
  good enough. Also, ungetc() failure still leads to error handling.



c.lib 1.181 (26.10.2004)

- The scanf() family now always returns the number of assignments made
  unless an error occured or an EOF was hit during conversion before the
  first assignment could be made.

- strtok() did not terminate properly if the last token in the string
  did not end with a separator character but with a '\0' byte. Fixed.

- The directory scanning functions opendir/closedir did not get the global
  directory data tracking data structure initialized which later led to
  Enforcer hits and maybe trashed memory. Fixed.


c.lib 1.180 (23.10.2004)

- The printf() family now produces no output at all for %e, %f and %g if the
  floating point support code is disabled. Previously, a minimum field width
  specification could take effect, printing a series of 0 or blank space
  characters where no output should have been produced.


c.lib 1.179 (22.10.2004)

- The scanf() family no longer assumes that a leading '0' indicates that the
  following digits form an octal number if the conversion type has been
  specified as already '%x' already.


c.lib 1.178 (7.10.2004)

- The OS4 version had floating point math support code enabled in all
  libraries and not just "libm.a". Fixed.


c.lib 1.177 (29.9.2004)

- Moved the locale initialization/cleanup code into constructors
  and destructors.

- The socket cleanup function is now a destructor.

- The math cleanup function is now a destructor.

- The wildcard cleanup function is now a destructor.

- The stdio cleanup function is now a destructor.

- The stack extension cleanup function is now a destructor.

- The code that cleans up after the program's current directory
  was changed is now a destructor function.

- Moved the initialization/cleanup code for unlink() into constructors
  and destructors.

- Moved the initialization/cleanup code for usergroup.library into
  constructors and destructors.

- Added usleep(), and created wrapper code that both sleep() and
  usleep() can use.

- Added strtoll() and strtoull(), with further changes to <limits.h>
  and <stdlib.h>.

- The socket exit code now calls the common stdio function which
  flushes and shuts down all buffered and unbuffered files.

- Fixed the stack swapping function which, for reasons unknown, ceased
  to work...

- The start time used by clock() is now initialized by a constructor
  function.

- NOTE THAT ALL THE CHANGES WITH REGARD TO USE OF DESTRUCTOR AND
  CONSTRUCTOR FUNCTIONS REQUIRE A COMPLETE REBUILD OF THE LIBRARY! IF
  YOU DO NOT DO THIS, THE CONSTRUCTOR/DESTRUCTOR FUNCTIONS MAY NOT
  GET CALLED.


c.lib 1.175 (10.9.2004)

- <stdarg.h> now tries to use the compiler supplied, machine specific
  <stdarg.h> file and also includes the OS4-specifc <sys/amigaos-va.h>
  file, if necessary.

- Added another test to <sys/time.h> to check if the "timeval" definition is
  already in place. If it is, <exec/types.h> will not be included.

- <sys/time.h> can be made to define the timeval data structure locally if
  the __USE_CLIB2_TIMEVAL preprocessor symbol is defined. In this case no
  <exec/types.h> or <devices/timer.h> file will be read.

- The sprintf() family did not properly handle empty strings as format
  specifications. This would result in a necessary buffer flush action
  getting skipped, which consequently did not put the string termination
  character into the output buffer.


c.lib 1.174 (27.8.2004)

- The parameters of atan2() were swapped. Fixed.

- Merged fdlibm 5.3 changes with __kernel_tan(), __exp() and __pow()
  functions.

- Replaced the rint() and rintf() functions with the fdlibm code.


c.lib 1.173 (25.8.2004)

- The <assert.h> header file was missing the C++ 'extern "C" { .. }'
  declarations.

- Added a new function __get_default_file() which allows direct access
  to the low level file handle/socket descriptor associated with a
  file descriptor.


c.lib 1.172 (21.8.2004)

- acos() now returns 0 for a domain error.

- asin() now returns 0 for a domain error.

- atan2() now returns 0 for a domain error.

- fmod() now returns x if y == 0 and sets a
  domain error.

- sqrt() now returns 0 for a domain error.

- Added NaN and +Inf constants to the math library
  initialization code which could come in handy later.


c.lib 1.171 (16.8.2004)

- Added "math_hypot.c"

- The 68k build makefile now builds the vfprintf/vfscanf functions
  with %lld/%llu support.

- Split the release notes file into changes and actual release notes,
  which may overlap but differ in technical content.


c.lib 1.170 (14.8.2004)

- Added "math_logb.c", and it appears to work, too.


c.lib 1.169 (8.8.2004)

- Since the 'long double' data type is not really supported by the GCC
  versions we use (and not supported by SAS/C either) building the
  library with support code for it is now a configurable option. Unless
  enabled, 'long double' is now treated like 'double'.

- Reworked the HUGE_VAL definition which previously would default to
  Infinity. The new code properly defines this to the largest
  representable floating point number. The 'great' thing about the
  HUGE_VAL definition in this library is that it's a reference to
  a binary constant initialized by the library startup code. So any
  application checking for HUGE_VAL merely needs to be relinked
  rather than recompiled.


c.lib 1.168 (7.8.2004)

- Ditched __is_infinity() and __is_not_a_number(), brought them back
  as isinf() and isnan().

- Updated code and header files to state exactly which ISO 'C' version
  is meant by the "The following is not part of the ISO 'C' standard."
  warning.

- Integrated rint() and rintf().


c.lib 1.167 (29.7.2004)

- Modified __is_infinity() and __is_not_a_number() to use more portable
  methods for accessing the low level representation of the IEEE 754
  numbers they work on.

- Fixed a compiler warning in h_strerror()


c.lib 1.166 (28.7.2004)

- Added h_strerror() function to libnet.a; there's a global 'h_errno'
  variable available, too.

- <signal.h> now defines a type 'sig_t'.

- <unistd.h> now allows the 'fd_set' type to be referenced as
  'struct fd_set', too.


c.lib 1.165 (26.7.2004)

- In printf(), if the precision is 0 and the value to be printed for
  the %d, %i, %u, %o and %x conversion types is 0, no output should
  be produced. Fixed.

- Fixed the rounding for %g (again). The maximum number of significant
  digits produced by a precision given for %g is different for numbers
  < 1.0 and those >= 1.0: the leading 0 for numbers < 1.0 does not count
  as a significant digit.

- The calculation of the number of leading digits in a floating point
  number now uses a mixture of log10() and a fall-back just in case
  log10() returns gunk.

- For floating point numbers, the sign, the digits, trailing zeroes
  and the exponent are now set up and 'assembled' separately. This
  removes the artificial limitation of floating point numbers to
  a maximum of 78 digits to follow the decimal point.


c.lib 1.164 (25.7.2004)

- The precision given for the printf() %g conversion types determines the
  number of digits to generate, not the number of digits to follow the
  decimal point. Fixed. The GCC FPU code generation now works again.
  Previously, the floating point number conversion crashed cc1 since the
  output amounted to 32 digits plus a terminating NUL, which overran the
  30 byte buffer sprintf() stored the output in. My, this was very hard
  to find :(

- Restored the floor() loop to figure out the size of the integer part
  of a floating point number. For some strange reason, the log10() function
  can return HUGE_VAL for numbers >= MAX_INT if mathieeedoubtrans.library
  is involved. This crashes the conversion, which has the effect of making
  the number come out as 0 rather than MAX_INT. Now that was nasty, too :(


c.lib 1.163 (25.7.2004)

- Simplified the %e/%f/%g conversion some more. We now call log10 and pow
  instead of looping and scaling the floating point number to be printed.


c.lib 1.162 (25.7.2004)

- Simplified the floating point number output code in vfprintf().
  Instead of two separate branches handling output for the %e and
  %f conversion types, the two now share the same code.


c.lib 1.161 (24.7.2004)

- Dropped all the non-public source code fragments from the library,
  including the memory pools, ARexx variable support and asyncio
  functionality.


c.lib 1.160 (24.7.2004)

- fputc() never flushed the output buffer if the stream in question
  was unbuffered.

- Rewrote puts() not to call fputs(). This avoids flushing the output
  buffer twice if stdout is unbuffered.


c.lib 1.159 (23.7.2004)

- Trying to open a directory like a file now reliably produces the
  error code EISDIR. Previously, the error code could have come
  out as ENOTDIR instead.


c.lib 1.158 (21.7.2004)

- freopen() was broken since it tried to reuse data from an already
  closed stream. Fixed.

- Fixed a number of small compiler problems resulting from debug
  output code that had never been exercised, or was not adapted to
  subsequent code changes.

- You can now enable debug output and have assertions come out
  correctly.


c.lib 1.157 (21.7.2004)

- The file hook message action "file_hook_action_write_and_flush" was
  not handled properly by the socket and scanf/printf hooks. Fixed.

- Removed a few assertions from the select() wrapper code which
  prevented it from working properly.

- Changed the stdio buffered I/O flushing code so that data can stay
  in the cache, if necessary, rather than get written to disk. This
  should help the asynchronous I/O feature.


c.lib 1.156 (21.7.2004)

- Brought back the __unix_path_semantics variable. If you link with
  -lunix and set this variable to FALSE, you'll still get Unix-like
  behaviour from certain routines, only the translations of Unix<->Amiga
  path name semantics will be disabled, leaving you with Amiga path
  semantics.

- sscanf(), asprintf(), vasprintf(), snprintf(), vsnprintf(), sprintf()
  and vsprintf() now use short local buffers to speed up conversion rather
  than a single measly byte.

- opendir() now avoids returning duplicate volume names.



c.lib 1.155 (21.7.2004)

- Lost all special case code which took care of unbuffered output
  and input. The new implementation works by using one byte of
  buffer, which is part of the iob structure. Unbuffered output
  is produced by 'flushing' the buffer once data has been written
  to it. The new code also has the advantage of making ungetc()
  work on unbuffered streams.

  NOTE THAT THIS MEANS THAT THE ENTIRE LIBRARY NEEDS TO BE REBUILT!

- Each FILE buffer now sports back-link information which can be
  tested to verify that it is valid.


c.lib 1.154 (21.7.2004)

- fwrite() has to call fputc() for each byte to be written. Oops...
  fwrite() no longer changes the file position if a write access
  went wrong. Trying to write to a stream which is not available
  for writing now counts as an error.

- fread() has to call fgetc() for each byte to be read. fread() no
  longer changes the file position if a partial read would yield
  fewer data than is necessary to fill an element. Trying to read
  from a stream which is not available for reading now counts as
  an error.

- Lost all the code that tested if the hook pointer embedded in a
  file descriptor table entry or buffered file was NULL. It
  simply has to be a valid pointer.


c.lib 1.153 (19.7.2004)

- Oops: cli->cli_DefaultStack does not necessarily match the size
  of the program's stack if it was launched from the shell and the
  "$STACK:" keyword embedded in the program overrides the default
  shell command stack size.

- Stack space allocated locally is now public rather than private
  memory (although it probably should be private).


c.lib 1.152 (19.7.2004)

- Simplified the code that figures out how much stack space was
  provided to the program when it was launched. The number does
  not need to be absolutely accurate at the time it is calculated.

- vfprintf() now prints a pointer (%p) as if it were a hexadecimal
  number with 8 digits (0x%08x).


c.lib 1.151 (18.7.2004)

- Oops: I mixed up the file descriptor and buffered I/O flags
  in fflush() and __fill_iob_read_buffer().


c.lib 1.150 (18.7.2004)

- Checked the dos.library/InternalRunCommand() code to see if
  the fiddling with Process->pr_ReturnAddresss is really necessary to
  figure out how much stack space is available. Turns out that it
  isn't under Kickstart 2.0 and beyond. Since this is what the library
  requires, only the task lower/upper bound and the current stack
  pointer address are consulted now.

- isascii() is now a macro.


c.lib 1.149 (17.7.2004)

- Reworked the ctype table layout and how the table is used by
  the optional ctype macros.


c.lib 1.148 (17.7.2004)

- Changed the flag definitions used by buffered files.

  NOTE THAT THIS MEANS THAT THE ENTIRE LIBRARY NEEDS TO BE REBUILT!

- Added new macros to <stdio.h> and <ctype.h> which replace functions
  of the same name and which might be more efficient.


c.lib 1.147 (17.7.2004)

- Removed the isascii() macro and replaced it with a real function.


c.lib 1.146 (17.7.2004) <jlangner>

- added a missing isascii() define to the ctype.h include file
  as it is quiet a common function macro and should be included


c.lib 1.145 (16.7.2004)

- Oops, I broke strncpy() *again*. Fixed.


c.lib 1.144 (16.7.2004)

- The file descriptor duplication code now sits in a separate
  function rather than in the different file/socket descriptor
  hooks.

- The code which reduces AmigaDOS path names by dropping path
  components followed by two slashes now sits in a separate
  function. Previously, it appeared twice in the Amiga<->Unix
  path name translation code.

- Simplified the select() function considerably.


c.lib 1.143 (16.7.2004) <jlangner>

- added missing hook_entry object files to GNU OS4 makefile

- added required newlines "\n" to the 68k GNU inline assembler
  functions to make them compile with GCC version 3 and upwards.


c.lib 1.142 (15.7.2004)

- ferror() does not return the code of the error for the stream
  which ran into trouble, it merely signals that an error occured.
  There goes another iob state variable.

  NOTE THAT THIS MEANS THAT THE ENTIRE LIBRARY NEEDS TO BE REBUILT!

- Removed all the code that set the value of 'errno' to 0.

- Likewise, removed all the code that reset the error condition
  that might have been set for a file. The only function which
  can do that is clearerr().

- No functions other but clearerr(), ungetc() and fseek() clear the
  'eof detected' flag now.

- Simplified vfprintf() some more, merging the %x formatting code
  with the other integer conversion code.

- vfscanf() now decides that with the %i conversion, a number is
  in octal notation if the first digit converted is a 0 and there
  is no 'x' or 'X' character following it.

- vfscanf() no longer counts '%n' as an assignment/conversion operation.

- Rewrote strtol() and strtoul() again. The overflow detection is
  now different in both routines and they avoid performing a
  conversion if the base is 1 (which would lead to an infinite
  loop). Also, the possible base values are limited to <= 36 and
  letters in the range a..z are allowed, too.

- All buffered streams which hold data and which operate in line buffering
  mode are now flushed before the buffer of a line buffered input stream
  is filled.

- fflush() now flushes all buffered files if the stream parameter
  is NULL.

- Unlike fgets(), gets() does *not* put the terminating '\n' character
  into the buffer. Fixed.

- rewind() now clears both the 'end of file' and error indicator flags.

- Replaced all memcpy() calls with memmove().

- Cleaned up error code translation and forwarding within the hook
  functions. Instead of mucking around with the global errno variable,
  the hook functions now return the error code as part of the hook
  message. The error code translation is now done entirely through
  two functions rather than one function, which fiddled with the
  global errno variable, and local code which, if necessary, translated
  ERROR_OBJECT_WRONG_TYPE into ENOTDIR.

  NOTE THAT THIS, TOO, MEANS THAT THE ENTIRE LIBRARY NEEDS TO BE REBUILT!

- Lost the unconditional address alignment test in the qsort() data
  swapping routine. If the code is built for an 68020 and beyond, it
  is replaced by a no-op.


c.lib 1.141 (14.7.2004)

- Rewrote the buffered file I/O code, and in particular how the read
  access interacts with the file's state information. Rather than
  counting how many bytes can be read from the buffer still, the
  new code checks the buffer position against the number of bytes
  stored in the buffer. This makes maintenance easier since only
  one state variable needs to be modified at a time.

  Note that this means that code which would use the getc() macro
  needs to be rebuilt!

- The pid_t type used by kill() and returned by getpid() is now
  really a Process address pointer. It used to be a CLI number.


c.lib 1.140 (12.7.2004)

- The stack alignment code used when reallocating new stack space and
  swapping it with the current configuration pushed the end of the
  stack beyond the boundaries of the allocated stack. This crashed
  in no time at all, but sometimes it just about worked without
  blowing up. The fixed code is simpler than before and now initializes
  both the lower and the upper stack boundaries to be aligned to
  multiples of 16 bytes.


c.lib 1.139 (11.7.2004)

- snprintf/vsnprintf were broken: instead of limiting output to
  the size given as a parameter, data beyond the end of the buffer
  was trashed.


c.lib 1.138 (11.7.2004)

- Greatly simplified the vfscanf() code. Added new code which
  initializes each parameter before the conversion begins.
  This should help software which tosses data at scanf()
  and expects it to come back initialized without checking for
  the function's return code.

- The vfscanf() '%f' conversion method could have consumed one
  more character than it was really allowed to. Fixed.

- If vfscanf() is built without floating point support, the
  '%f' conversion method now properly pulls the respective
  floating point number from the stack but performs no
  conversion and does not modify the contents of the
  variable.


c.lib 1.137 (11.7.2004)

- Moved the initialization for the stack size checking out
  of the _main() code and into its own initialization function
  where it belongs. That way, the test also works for programs
  that detach from the shell.

- The '%g' and '%G' vfprintf() format specifiers now have the
  effect of stripping trailing 0 and . characters in the number
  string produced. Previously, this effect was also enabled for
  the '%f', '%e' and '%E' formats.

- Greatly simplified the vfprintf() code and verified that the
  output matches the requirements.


c.lib 1.136 (9.7.2004)

- Dropped in loads more of assertions to check the validity of
  files and descriptors.


c.lib 1.135 (6.7.2004)

- Fixed the last two issues in puts() and fputs() caused by the
  bug fix I made to fputc() which would not return the character
  value passed in. Occasionally, this could cause puts() to omit
  the trailing '\n'.


c.lib 1.134 (4.7.2004)

- Simplified and reordered the init/exit procedure in "stdlib_main.c".
  Initializations which could be done on the spot when required are
  no longer performed that early on. The stdio initialization is
  performed earlier so that error messages can be printed using
  stdio methods.

- A stack overflow can no longer turn into an infinite loop.

- Removed some more redundant data from the memory management data
  structures.

- The temporary file cleanup in fclose() did not properly turn off
  profiling. Fixed.

- Both close() and fclose() now fill their data structures with 0
  bytes before they return.

- The tables for the file descriptors and buffered files are now
  allocated dynamically.

- If the asynchronous I/O options is not in effect, then the
  global file data structures will not include a pointer to
  the AsyncFile.

- Lost some duplicate code from the locale cleanup routines.

- Simplified the code that picks a vacant file descriptor table
  slot.

- NOTE: YOU NEED TO REBUILD THE ENTIRE LIBRARY IF YOU WANT TO USE
        THESE CHANGES!


c.lib 1.133 (4.7.2004)

- The memory routines no longer look up the management data structures
  before they deallocate or modify an allocated memory chunk. This is
  now done only in the debug version of the library.

- Moved the code that performs memory deallocations in the debug and
  nodebug versions of the library. Instead of separate functions, a
  single common routine is used.

- When launched from shell, the program's name is no longer allocated
  through malloc() but through AllocVec(), and released separately later.
  This helps the memory debug code which otherwise wouldn't be able to
  print meaningful information when starting up and shutting down.

- You can now get additional information out of the memory debugging
  code, covering the locations where memory is allocated and released,
  including how much memory is involved.

- The month and day names used by the time conversion code are now
  declared as 'const'.

- Replaced all uses of the EPERM error code with EACCES.



c.lib 1.132 (26.6.2004)

- abort() and raise() now call a common termination routine.

- Chopped the math_kernel.c code into individual files. Some local constant
  values had to be duplicated, but so there...

- Lost "smakefile.old" and "math_kernel.c".

- Got rid of more compiler warnings...

- Fixed a typo in the "stdlib_headers.h" header file.

- In libdebug.a, for 'va_arg()' to work with 'char' type parameters, these must
  be picked up as type 'int' from the stack. Otherwise we get a linker error later...
  This should finally make the PowerPC 'native' kprintf() function work.


c.lib 1.131 (21.6.2004)

- The '%i' conversion method in the scanf() family did not take
  a leading '0' properly into account when updating the number of
  letters to be read and processed.

- If the '%c' conversion method reads fewer than the required
  characters, it stops the conversion rather than aborting it
  with an error.

- The scanf() family now watches more carefully what data it still
  requires to satisfy a conversion. For example, while it may accept
  a '-' character to indicate that a sign is given to the number to
  follow, if there is no number to follow will trigger an error
  condition.


c.lib 1.130 (20.6.2004)

- The scanf() family now returns EOF if an error was detected
  (not just an end of file) or if the stream ended before a
  conversion could be performed.

- The file type reported by stat() did not really take pipes into
  account. Fixed.

- Cleaned up the data structure definitions and references. This turned
  up a bug in the Unix<->Amiga path name translation which would use
  a local data structure which was larger than the public version. Fixed.


c.lib 1.129 (18.6.2004)

- Added a free memory threshold test feature to malloc(); how it works
  is described in <dos.h>.

- The PowerPC debug.lib now calls RawPutChar() and RawMayGetChar()
  through the 68k LVO. This should finally make kprintf() and friends
  usable on the PowerPC.

- The asyncio code now defaults to allocate at least twice as much
  memory as a buffered file would require (so that double buffering
  can be used). Also, the fundamental asyncio support functions require
  that enough memory is allocated to allow for four disk blocks to
  be read at a time.

- When opening a file for asynchronous I/O access, both read-only and
  write-only access is supported now. Previously, the functionality
  was limited to read-only access.


c.lib 1.128 (14.6.2004)

- Moved the stack size usage reporting code again so that the name of
  the program can be printed correctly. Also, the data from the
  initial stack size swapping is now cached until it is used.

- Reordered the resource allocations and initializations in tmpname()
  to make error recovery more robust.


c.lib 1.127 (13.6.2004)

- Truncating a file to size 0 was, if successful, considered a failure.
  Fixed.

- ftruncate() no longer changes the file position.

- Advisory record locking now checks if the lock should be shared or
  exclusive and verifies that the file open mode supports this access
  mode.

- Negative record lengths are now supported in the advisory record
  locking code.


c.lib 1.126 (13.6.2004)

- Unless Unix path semantics are in use, lseek() no longer defaults
  to 'seek and extend' operation.

- There are no longer two separate functions to take care of 'seek'
  and 'seek and extend' operations.

- The Unix-like lseek() behaviour now takes advantage of cached
  file seek position information. Previously, it just ignored it.

- The 'truncate' file action now makes sure that once the size of
  the file has been trimmed, the new file position refers to the
  end of the file. There also was a problem with growing the size
  of the file which could have started at the wrong file offset.

- The advisory file locking code now handles integer overflows
  gracefully when the end of a locked region is to be calculated.
  In this case, the end is assumed to be the (2^31)-1st byte.
  Also, when removing a lock the search for the locked region in
  question keys off the start position and the length rather than
  the start position and the adjusted end position.

- Added more sanity checks to file system operations which invariably
  have to deal with FileHandle->fh_Type.

- Rewrote the Unix->Amiga path name translation function again.


c.lib 1.125 (13.6.2004)

- Rewrote the Amiga->Unix path name conversion code. Double slashes are
  translated more robustly into '../' and the translation of the volume
  name and the path to follow it are now done separately. The current
  directory name ("") could wind up getting translated to a string pointing
  into a local buffer. Fixed. Also, the new code knows how to translate
  ":" to a path relative to the current directory (if that directory is
  defined by an absolute path name).

- The code that parsed and initialized the command line parameter
  array to be passed to the main() function made different assumptions
  during the counting and the initialization of the different parameters.
  Fixed.

- snprintf() and vsnprintf() now handle zero length buffer
  sizes properly.

- The stack size usage measuring code got broken when the
  constructor/destructor mechanism was reimplemented. It
  now works again.


c.lib 1.124 (12.6.2004)

- Dropped the fseek() optimization for files with a write buffer. This
  feature had little impact, but complicated buffer management enormously.

- In fputc(), __fputc(), fwrite(), __flush(), etc. and the functions which
  build upon them flushing the buffer now always takes place before new data
  is pushed into the buffer. The exception is fputc()/__fputc() which can
  flush the buffer if the file is in line buffering mode and a line feed
  character has just been written. Also, failure to flush the buffer now
  implies that the character/data to be stored in the respective function
  was not successful. Previously, the different functions either stored the
  data in the cache before and/or after the buffer was flushed, making it
  difficult to tell if it had been written. Which in turn made it impossible
  to recover from the failure since you could not be sure if another attempt
  to store that data would really store it at the proper file position.

- mktime() now insists that the year must be 1978 or later.

- In debug.lib, KPutFmt() and the functions which build upon it now
  should work even on the PowerPC. This works by analyzing the format
  string and by converting the parameters to fit into a table whose
  contents RawDoFmt() can handle.

- For the asyncio code, all write accesses immediately flushed the buffer,
  thus nullifying any advantage the asynchronous I/O might have provided.
  Fixed.

- The optimized memset(), memcpy() and memmove() functions now use
  displacement addressing rather than post- and preincrement in the
  unrolled loop. This should help strictly pipelined processors. It
  does not seem to make any difference at all for the 68060, though.

- The locale cleanup code now separately closes the specific locales
  and the 'all' locale.


c.lib 1.123 (12.6.2004)

- The distribution license no longer requires that for binaries using
  the library a mention of the license text is mandatory.

- closedir() now checks for a NULL pointer and backs out if it finds
  one.

- opendir() now exits with an error if it cannot allocate enough memory
  for a volume name to be added to the list.

- readdir() now checks for a NULL pointer and backs out if it finds one.

- readdir() no longer builds with volume list scanning code unless
  Unix path compatibility is enabled.

- rewinddir() now ignores a NULL directory pointer.

- strncmp() now treats the parameters as strings of unsigned characters.

- strdup() no longer traverses the string to be copied twice to figure
  out how much data needs to be copied.

- strerror() now knows how to produce an error message for error
  code 0x80000000.

- strtol() and strtoul() always need to initialize the 'ptr'
  parameter before they exit, not just when they succeed.

- fputc() did not return the character value written, but 0. Fixed.

- fputs() now returns EOF if it couldn't flush the write buffer,
  if that was necessary.

- fseek() no longer changes the value of the global 'errno' variable,
  as required by the specification.

- fseek() now sets the buffer position to 0 if the number of bytes that
  can be read from the buffer drops to 0. This avoids clashes with write
  access to the same file.

- ftell() now always sets the global 'errno' variable to a well-defined
  value.

- fwrite() now tries to seek back to the original file position if
  flushing the buffer failed. This complements similar code in fread().

- fwrite() no longer flushes the buffer before it returns. It appears
  that this behaviour is not mandated by the specifications. The
  buffer will only be flushed upon entry, and even then only to make
  room for new data to be stored.

- fread() now preserves error status information when recovering from
  a failed buffer fill attempt. Previously, the error status information
  would revert to 'no error' during this task.

- mktemp() now also adds the current Task address to the seed value that
  goes into the creation of the temporary name.

- If dup2() fails to close a file, if that is necessary, it will return
  an error rather than ignoring the failure.

- Lost a return statement in dup2(). Also, growing the size of the
  file descriptor table can now be aborted with ^C.

- In fcntl(), growing the size of the file descriptor table can now be
  aborted with ^C.


c.lib 1.122 (1.6.2004)

- Lost the NULL tests in front of all UnLock() and FreeDeviceProc() calls.

- The command line is now trimmed, dropping leading and trailing
  blank spaces.

- The ftruncate() functionality did not write 0 bytes to the file
  being extended. Fixed.

- popen() allocated a few bytes less than necessary for the command
  name and the parameters.

- Neither popen() nor system() now quote the command name unless it
  is really necessary.


c.lib 1.121 (30.5.2004)

- Made the memory debugging options a little more flexible and easier to
  configure.


c.lib 1.120 (30.5.2004)

- The asynchronous I/O code was never actually used by the library. Fixed.

- Tweaked the realloc() code to take the memory allocation granularity into
  account. This means, for example, that for small allocation sizes < 8 bytes
  no reallocation will ever be performed since it wouldn't amount to anything.
  This complements similiar code in malloc() which also rounds up allocation
  sizes to the OS memory allocation granularity limits.


c.lib 1.119 (27.5.2004)

- Got rid of a few compiler warnings involving signed/unsigned integer
  comparisons.

- Fixed a bug in the Unix->Amiga translation routine which would cause
  file names such as './../foo' come out as '/foo' and end up being
  translated to 'foo:'. The new code now takes care of removing leading
  './' and embedded '/./' segments before further processing is
  performed on the name string.

- Fixed up the 68k GNUmakefile so that it builds properly.


c.lib 1.118 (26.5.2004)

- Rewrote those putc/__putc macros again...

- The vfprintf()/fputs() code did not properly test the result of
  the __putc() function call. There should have been a test for
  '!= EOF' rather than '== OK'.


c.lib 1.117 (26.5.2004)

- realloc() now checks the size of blocks to be shrunk. If the block
  shrinks by less than half its current size, then no new memory
  is allocated for it. This saves the need to allocate memory and to
  move the data.

- Implemented an option to have the allocated memory chunks managed
  by a balanced tree, sorted by address. This should help applications
  which claim a lot of memory, possibly in small chunks, which get
  allocated/released a lot. While this is still not quite as good as
  having the low level memory management code use buddy lists or
  slab allocation, it might still do the trick. This feature is
  activated in "stdlib_headers.h" via the __USE_MEM_TREES preprocessor
  symbol. You will need to rebuild the files stdlib_data.o, stdlib_free.o,
  stdlib_init_exit.o, stdlib_malloc.o, and stdlib_red_black.o if you
  change this.

- The size of the physical memory allocation chunks is now taken into
  account when making memory allocations. This means that allocations
  are rounded up to multiples of 8 bytes each.


c.lib 1.116 (26.5.2004)

- Reworked the putc/getc macro definitions, moved them behind the
  function prototype declarations in <stdio.h>.

- The putc() macro did not return the same result as the equivalent
  fputc() function would have done after flushing the buffer. The
  new macro now calls a special flush routine which itself safely
  returns the last character written.


c.lib 1.115 (24.5.2004)

- Prior to performing ^C checking, the respective routines now verify first
  if ^C checking is enabled in the first place. This saves a function call.

- Replaced all ^C handling code that would drop straight into raise(SIGINT)
  with a proper call to __check_abort(), which the user may have set up to
  his own liking.

- The wildcard matching argument list build code no longer defaults to
  allow for ^C checking. This is permitted only if the global checking
  flag is enabled.


c.lib 1.114 (24.5.2004)

- Replaced most of the internal single character I/O calls with macros.

- I/O buffers are now allocated so that their start addresses are aligned
  to cache line boundaries.

- Buffer size and position are now both signed integers. This fixes a
  tricky issue with the newly-introduced macros and zero length buffer
  sizes.


c.lib 1.113 (23.5.2004)

- The 'FILE' structure now has a public part. Flags and layout have changed!

- Reimplemented getc() and putc() as macros, which helps a lot with some
  applications.

- Added more of the SAS/C internal floating point support code, which makes
  sprintf() and friends build even with floating point options enabled.

- Replaced CheckSignal() with SetSignal(), which helped the ^C checking
  code a lot.


c.lib 1.112 (23.5.2004)

- Plugged in the SAS/C code profiling functionality. This works by measuring
  calling time intervals and should be very portable.

- Duplicating file descriptors could lead to trouble if the descriptor
  number in question exactly matched the number of descriptors. Fixed.

- Reactivated code in open() which was disabled for OS4 but which
  should have stayed active.

- Reduced the overhead of the single character input/output routines,
  especially when called from within the library. This helped fgets(),
  scanf(), fputs() printf() and others.


c.lib 1.111 (12.5.2004)

- chdir() now checks if the object in question is really a directory.

- Added special tests to a number of Open(), Lock() and Delete() calls
  to catch attempts to open "subdirectories" of files. If it happens,
  e.g. trying to open <utility/tagitem.h> when "utility" is a header
  file on the search list, errno will be set to ENOTDIR.


c.lib 1.110 (10.4.2004)

- memmove() no longer tries to move large chunks of data if the source
  and destination overlap by fewer bytes than can be moved at a time.
  The original code should not have done this in the first place, but
  it's safer to check rather than to assume that it will work out fine.


c.lib 1.109 (9.4.2004)

- Fixed a bug in the PPC version of setjmp that would trash floating
  point registers.

- Fixed a bug in dup2: If the second fd didn't exist, it would return -1,
  which is wrong (it should not close the second fd when it doesn't exist).

- Added F_DUPFD to fcntl


c.lib 1.108 (15.3.2004)

- Fixed a bug in the printf() family which affected how the %e
  output was formed (and by extension, %E, %g and %G). The %e
  format must put exactly one digit in front of the decimal
  point, followed by the mantissa and the exponent. This worked
  fine unless the round-off fudge added to the number caused
  an overflow, producing two digits in front of the decimal point.
  This has been fixed by checking for an overflow after having
  added the round-off fudge and rescaling the result, if
  necessary [Andrija Antonijevic].


c.lib 1.107 (15.2.2004)

- The setjmp() code now actually manages to save and restore the
  FPU state in the 68k library form.

- The SAS/C constructor/destructor invocation finally works.

- Added an option to cache the current file position, which should
  help seek performance.

- Added targets "version" and "cvs-tag" to the makefiles.


c.lib 1.106 (14.2.2004)

- index() now calls strchr() and rindex() calls strrchr().

- strchr() and strrchr() now read each byte of the string to
  be searched exactly once.

- Added optimized versions of memchr() and memcmp().

- The optimized versions of memchr(), memcmp(), memcpy(), memmove()
  and memset() check first if the amount of data involved justifies
  the expense of gearing up to perform an optimized operation.

- strstr() no longer calls strlen() even once.


c.lib 1.105 (12.2.2004)

- Replaced memcpy(), memmove() and memset() with optimized portable
  versions.


c.lib 1.104 (11.2.2004)

- When open() calls Lock() and fails to obtain a correct result, the
  error code checking now also covers ERROR_ACTION_NOT_KNOWN (which is
  ignored). This should help handlers which are no file systems, such
  as PAR: and PIPE:.

- The file->iob_BufferWriteBytes and file->iob_BufferPosition counters
  no longer have to be in sync. In fact, the position can be smaller
  than the number of bytes waiting to be written. The buffer flushing
  has to take this into account now (it may have to seek back after
  having dumped the buffer, which is nasty).

- Worked Jrg's changes into the fseek() code. This covers two fields:
  if possible, an absolute positioning command is converted into a
  relative position change; if the current file buffer allows for it,
  the buffer access position is updated rather than forcing the buffer
  to be flushed to disk/reread.


c.lib 1.103 (8.2.2004)

- The bug fixes in version 1.101 broke ungetc(). Fixed. Also reworked the
  sanity checks in ungetc() which previously were performed either out
  of order or were entirely missing.

- Renamed socket_get_socket_descriptor.c to socket_get_descriptor.c to
  make it easier to build the library on the Amiga.


c.lib 1.102 (22.1.2004) <jlangner>

- modified string.h to also include strings.h within the non ISO 'C' standard
  part. This should make porting of existing applications a bit easier as many
  sources only include <string.h>. As the current string.h anyway defines the
  symbols of i.e. bcmp() and bcopy() etc, which should normally also be
  in strings.h, I don't see any problem with that.

- added function macros for stricmp() and strnicmp() to non ISO 'C' standard
  part of <strings.h> to make it easier to port applications relying on
  stricmp()/strnicmp().


c.lib 1.101 (18.1.2004)

- Draining an stdio buffer via fgetc/fread, etc. completely before switching
  over to using the same buffer for writing data to a file would flop because
  one of two important variables did not get reset to zero. This would also
  end up trashing the buffer management. New code makes sure that if a read
  access ends up setting the number of bytes available for reading from the
  buffer to 0, the buffer position is reset to 0 as well.

- fwrite() now pushes the buffer contents to disk if it has completely
  filled up rather than allowing the full buffer to stick around until
  another opportunity presents itself to flush it to disk.


c.lib 1.100 (18.1.2004)

- abort() and raise() no longer call fprintf(stderr,..) but rather
  use fputs(..,stderr) instead.

- Updated the copyright text.


c.lib 1.99 (28.12.2003)

- Updated "stdlib_constructor_begin.c" to invoke the SAS/C constructors
  and destructors and to make the OS4 specific destructor handling
  safer: destructors calling exit() will now cause the next following
  destructor function to be called.


c.lib 1.98 (23.12.2003)

- Fixed up Hans-Jrg's previous fix for strncat(). Added comments
  explaining as to what this function is supposed to accomplish in
  order to avoid making the same mistake again and again.

- accept() now returns a properly wrapped socket descriptor (thanks to
  Marcus!).

- TODO: fix up the PowerPC destructor code; currently, a destructor
  function calling exit() will cause all subsequent destructors to
  be either skipped or the entire destructor list to be rerun ad
  inifinitum (haven't checked this in detail yet). Change the code to
  work more like the 68k version, which plants a setjmp() at the beginning
  of the loop and remembers which one was the last destructor called.


c.lib 1.97 (15.12.2003)

- The library can now use constructors and destructors internally.

- Added the realpath() function.

- Put the dirent functions back in libc.a.


c.lib 1.96 (13.12.2003)

- gettimeofday() no longer reports microsecond values > 1000000.


c.lib 1.95 (13.12.2003)

- The default memory pool and puddle sizes can now be configured at link
  time by defining the __default_pool_size and __default_puddle_size
  variables in your program.

- strncat() no longer eats the last character copied.


c.lib 1.94 (9.12.2003)

- <utime.h> now includes <time.h> rather than <sys/types.h>, if
  necessary.

- Changed DoTimer() to allocate the MsgPort and the timerequest from
  public memory rather than from the stack (which could be allocated
  as private memory).


c.lib 1.93 (29.11.2003)

- Fixed up the Unix->Amiga path name translation code to map "volume:/" style
  paths to the virtual root directory.

- The file operations (open, chmod, mkdir, rmdir, mktemp, chown, unlink,
  utime and rename) now watch out for the root directory to come out of a
  Unix->Amiga path name translation and will react according to it.

- access() now also works on the virtual root directory.


c.lib 1.92 (28.11.2003)

- Replaced qsort() with a different, iterative implementation. In terms of
  performace it seems to beat everything out there.

- Tinkered with the select() code again. The local FD_ZERO() implementation
  was too smart. It would attempt to optimize the operation by breaking it
  down in to series of bytes to be set to zero, followed by individual bits.
  This went wrong because this isn't how the fd_set bits are organized.
  The code now clears all the long words involved.

- Updated the fix_datestamp() function which is part of the select()
  wrapper. The updated version should be less vulnerable against
  overflows.


c.lib 1.91 (25.11.2003)

- Fixed up the dependencies in the makefiles. The libunix.a routines which
  would involve the asynchronous I/O code would not get rebuilt if the
  associated header file would change.

- Added extra code to the fd hook function which looks after files that
  should be unlinked, but which didn't get deleted at the first attempt
  because they were still in use. That extra code checks, after a file is
  closed, if the file in question was set up for deletion. This seems to
  help the Samba port a lot, which no longer accumulates scores of lock
  files in "Samba:log/locks" which get deleted only when smbd exits.


c.lib 1.90 (24.11.2003)

- The wildcard pattern expansion for command line arguments no longer
  drops parameters for which there is no match. Instead, the argument
  is used as is. With the old behaviour, things like
  "-D__interrupt=__attribute__((__interrupt__))" would expand to nothing,
  breaking the GCC preprocessor.

- The special definitions & code that help GCC to build itself are now
  enabled through a single line in a local header file.

- The file descriptor hook now sets errno to EBADF for file descriptors
  which are not in use (this used to be ENOSYS).


c.lib 1.89 (23.11.2003)

- The asynchronous file I/O support can now be enabled at library build
  time by uncommenting a definition in a single header file. The makefile
  dependencies then do the rest...

- Fixed a bug in the Unix->Amiga path name conversion code. A trailing
  "/." in a path name is now stripped.


c.lib 1.88 (22.11.2003)

- alloca() is now included in the library even if GCC is used as the
  host compiler.


c.lib 1.87 (17.11.2003)

- truncate() and ftruncate() now safely extend the size of a file by
  adding 0 bytes at the end.


c.lib 1.86 (17.11.2003)

- Fixed a bug in the Unix->Amiga name translation which could lead to
  a path beginning with "../" to be mistakenly identified as being an
  absolute path name starting with "/".


c.lib 1.85 (16.11.2003)

- The realignment code was missing in the FlushAsync() routine. Fixed.

- The alignment position was not calculated correctly in SeekAsync()
  as the wrong variable was involved. Fixed.

- __umodsi3, __udivsi3, __udivsi4, __modsi3, __divsi3, __divsi4 would
  not refer to the proper __UtilityBase library symbol. Fixed.


c.lib 1.84 (12.11.2003)

- Updated the optional asyncio code to return to a block-aligned
  file position after seeking in a write-only file. Also added
  code to check if read/write access is actually allowed.

- Multiple inclusion of debug.h no longer leads to trouble.

- Merged duplicate file advisory lock testing code.

- The dirent functions are now a part of libunix.a and no longer
  appear in libc.a.

- Advisory file locks can now start beyond the end of the (current)
  file and it is possible to lock a region until the end of the file.
  Previously, the latter was rejected as an invalid range.

- If several exactly overlapping regions are locked on a file,
  they all are now unlocked by a single call.

- Replaced the qsort() code.

- The assertion handling has moved into its own routine which
  tries to make sure that the failure message is shown.

- Rewrote clock() (again) to be much simpler in operation.

- The opendir/readdir/closedir cleanup code now actually invokes the
  closedir() function rather than doing the cleanup job manually.

- open() no longer tests for delete protection after the file has
  been already opened. If it's open, only a possible write protection
  may be problematic and ought to be flagged.

- __drop_iob_read_buffer() now checks if the buffered file is actually
  in use.

- If fseek(..,0,SEEK_CUR) is is called, it no longer results in the
  read/write buffers to be dropped.

- The snprintf() family now returns the proper number of characters
  that would have gone into the buffer, if enough room were available.
  Previously, it would stop counting after the buffer had filled up.

- Merged duplicate code around free().

- The memory cleanup code now calls the free functions if memory
  debugging code is enabled.

- If the memory debugging code is enabled, allocations and deallocations
  will trigger consistency checks now and then (with a likelyhood of
  about 6% per operation).

- mktime() now properly fills in the day of the week.

- strftime() now expects the format hook code to provide for the NUL-
  termination.

- __translate_amiga_to_unix_path_name() now performs a sanity check
  before it proceeds to use the supplied path name. The check is
  to make sure that the local buffer does not overflow. Rearranged
  code in __translate_unix_to_amiga_path_name() to perform a similar
  test.

- Rewrote the command line wildcard expansion code to use standard
  Exec lists rather than the tricky singly linked list stuff, coupled
  with tricky pointer arithmetic. If the code crashes, it's now much
  easier to tell why.

- The __usergroup_init() function now returns an error code rather
  than dropping into exit() in case of trouble.

- They 'key' parameter of bsearch() is now of type "void *" rather
  than "char *".


c.lib 1.83 (5.11.2003)

- The scanf() family now skips over initial whitespace preceding
  data to be converted. Exceptions are the %c, %n and %[ conversions
  which are documented not to bother.


c.lib 1.82 (4.11.2003)

- Fixed another bug in the scanf() family. Floating point numbers were
  not converted correctly, only the integral part was picked up.


c.lib 1.81 (4.11.2003)

- Added assertions to track trashing of certain global variables.

- Rewrote the select() wrapper again. It no longer makes assumptions about
  the size of the fd_set data structures passed in. The new code is, however,
  significantly more complex than it used to be. It uses far less stack space,
  though...


c.lib 1.80 (29.10.2003)

- More tweaking for the optional asyncio support. If requested, it's
  always enabled for read-only files bound to descriptors, but never
  for write-only or read/write access. It's always enabled for buffered
  files standard I/O files because there's an fflush() to take care
  of pushing the buffer contents to disk. There is no fsync() for
  file descriptors which would accomplish the same.


c.lib 1.79 (29.10.2003)

- The optional asyncio code now also sports a function similiar
  to fflush() which will write the current contents of the buffer
  to the file.


c.lib 1.78 (27.10.2003)

- The asyncio code support is now an optional feature which can be
  enabled at library build time by defining the ASYNCIO_SUPPORT
  preprocessor symbol.

- Cleaned up the code that uses the buffered I/O data structure
  'iob', which used to be a much bigger data structure with
  a hard-coded buffer attached to it. The buffer is now allocated
  dynamically when needed.

  Note that this change requires that the library is rebuilt or the
  I/O buffer data structures will cause trouble.


c.lib 1.77 (26.10.2003)

- Worked an updated version of the asyncio code into the library.
  File I/O performance will benefit from this if, and only if, a
  file is opened either for read or write access only. If both
  read and write access are requested for a file, the asyncio layer
  will not kick in. Also, in order to enable it, you have to declare
  a '__asyncio_buffer_size' variable (see "stdio_asynciobuffersize.c")
  somewhere in your program.

  Note that this change requires that the library is rebuilt or the
  file descriptor data structures will cause trouble.


c.lib 1.76 (22.10.2003)

- Added <sys/amigaos-va.h> to define the macros va_startlinear() and
  va_getlinearva().


c.lib 1.75 (20.10.2003)

- __initialize_fd() and __initialize_iob() now perform the bulk of the
  data structure initialization rather than just setting the data to
  zero and plugging in a hook function.


c.lib 1.74 (19.10.2003)

- Renamed the global __fd_table variable to __fd to match the buffered
  I/O data structure name.

- Further modifications to make the debug version of the library build
  without errors.

- Moved the stack size usage checking code from _main() into separate
  routines which are found in the linker library.


c.lib 1.73 (19.10.2003)

- __grow_fd_table(), __grow_iob_table(), __flush_iob_write_buffer(),
  __drop_iob_read_buffer(), __fill_iob_read_buffer() and __open_iob()
  now returns 0 for success, -1 otherwise.

- __grow_fd_table() and __grow_iob_table() now both set errno in
  case of failure.

- Lost the unnecessary file descriptor and buffered I/O table
  initializations.


c.lib 1.72 (19.10.2003)

- fopen(), freopen() and fdopen() now all call the same common
  routine to set up a buffered stdio stream. Previously, the
  fdopen() code was seperate and did things differently from
  the rest.

- Deletion of temporary files upon program exit could lead to
  a crash (worst case), but never actually deleted anything. Fixed.

- The stdio_exit() cleanup function now actually calls into
  fclose() and close() rather than doing the same job all over
  again manually.

- Lost code which didn't make much sense in the socket descriptor
  hook. Same thing for the temp file cleanup code.

- dup() now invokes dup2() with the second parameter set to -1.

- Simplified the duplicate cleanup in close().


c.lib 1.71 (19.10.2003)

- Added missing prototype for setbuf() in <stdio.h>.

- Put the definition of HUGE_VAL in <math.h> inside #ifndef, to allow
  GCC to do its thing.


c.lib 1.70 (18.10.2003)

- The tests to verify that a file descriptor was valid all had a
  problem. This test is now done in a separate function which
  performs the necessary error checking.

- The dup2() function did not close the second file descriptor
  prior to duplicating the first one. Also, the dup(..,-1) case
  did not return the number of the new file descriptor but -1
  instead.

- The tests to check whether a Hook is attached to a file
  descriptor now set errno to EBADF if no Hook is found.

- All file descriptor flags are now cleared when a descriptor
  is closed. Same thing for buffered files.


c.lib 1.69 (14.10.2003)

- Not all conversions performed by the scanf() family would respect
  the maximum field width. Fixed.

- The local sscanf() file hook function had the last two parameters
  reversed. Fixed.

- strftime() now fills in the week day if it's not provided.


c.lib 1.68 (13.10.2003)

- To activate the memory debugging code, use "#define __MEM_DEBUG".
  Other debugging code (assertions) is deactivated through
  "#define NDEBUG".


c.lib 1.67 (11.10.2003)

- Reworked the program startup code. Instead of rattling down a number of
  initialization functions, one at a time, the whole thing is now performed
  through a 'constructor style' table. This ought to be replaced by the
  real thing once it's safely possible.


c.lib 1.66 (11.10.2003)

- UtilityBase, LocaleBase, SocketBase and UserGroupBase no longer override
  any user defined global variables. The library now keeps them under
  private names.

- The wildcard expansion code would not return an empty argument list in
  case no match was found. Fixed.

- Cleaned up the stdlib_main.c code, moved all the data and function
  references into local header files.

- Completed the usergroup.library API integration for the OS4 variant.

- The constructors are now called in forward order and the destructors
  in reverse order.

- Moved the constructor/destructor invocation code into a separate file.


c.lib 1.65 (9.10.2003)

- In <sys/types.h> dev_t is now an unsigned long.


c.lib 1.64 (9.10.2003)

- Bum. The old and the new clock() code was wrong. The clock() function
  is supposed to return how much processor time the current process
  consumed since it was started, which is a monotonously increasing
  number (or something like it, unless the system time is reset).
  The new code should deliver proper data, although it might be better
  to switch it over to timer.device/ReadEClock.


c.lib 1.63 (8.10.2003)

- Ouch, the old clock() code was correct, it just didn't look like it.
  I rewrote it to make it clearer what happens and why it happens like
  that [Andrija Antonijevic].


c.lib 1.62 (6.10.2003)

- A "+" and a "*" had been mixed up in clock(). Fixed.


c.lib 1.61 (5.10.2003)

- Now that main.o is a platform neutral startup code file, it's possible
  to do some machine type checks in "stdlib_machine_test.c". This code
  checks for presence of an FPU and an MC68020 CPU (or better). It may
  not work for -fbaserel32 which by the time the test is performed
  will have already crashed.

- Added the missing stdlib_dosbase.c, stdlib_localebase.c, stdlib_sysbase.c,
  stdlib_utilitybase.c and math_init_exit.c files to the repository.


c.lib 1.60 (5.10.2003)

- Renamed the source code files which implement the strtod(), strtol(),
  strtoul(), atof(), atoi() and atol() functions.

- Moved main() out of libc.a and made it into a startup object file
  which needs to be added to the GCC specs file in the "*startfile"
  section, right at the end of the list. This solves practically all
  linking problems I encountered so far. Which means that -lunix
  and -lsocket actually work and that you don't have to tweak the
  specs file to build the compiler.


c.lib 1.59 (4.10.2003)

- The math initialization/test function is now always invoked by the
  _main() function. The math init/exit code defaults to stubs, though,
  unless libm.a is linked in.


c.lib 1.58 (4.10.2003)

- Moved the global library base/interface definitions from stdlib_main.c
  into separate files, each of which defines a base/interface pair.

- Moved the math library initialization/testing into libm.a where it
  really belongs. The MathIEEE library base definitions went there,
  too.


c.lib 1.57 (1.10.2003)

- Calling dup2() on a stdio file reinitialized the source fd instead of the
  destination fd.  Fixed.

- Added const declaration to first argument of sscanf().

- <assert.h> no longer enforces the inclusion of <dos.h>.


c.lib 1.56 (30.9.2003)

- Moved the scalbn() function prototype into math_pow.c, where it is
  used. This is not an ISO 'C' function to start with.

- strtod(), strtol(), strtoul(), atof(), atoi() and atol() moved
  from <string.h> into <stdlib.h>.

- fputs() no longer adds a newline character at the end of the line.
  This is now done only by puts().

- fdopen(), fopen(), freopen() and popen() no longer mix up file
  access mode options such as "r" and "r+".

- popen() now adds the current time stamp to the name of the
  temporary files it opens.


c.lib 1.55 (21.9.2003)

- Added support for 'long long' quantities in the printf/scanf
  families. This is not yet enabled by default since I couldn't
  seem to make it work on the 68k platform.


c.lib 1.54 (18.9.2003)

- fputs()/puts() didn't add the '\n' character at the end of the
  line. Fixed.


c.lib 1.53 (25.8.2003)

- Added hooks for code to perform wildcard expansion for all
  command line parameters. Excluded are those parameters which
  are surrounded by quotes (").


c.lib 1.52 (19.8.2003)

- fpos_t is now a scalar data type.

- size_t is now an unsigned integer.

- Lost the file descriptor table usage counter. Whether or not a
  file descriptor is in use is now indicated by a flag.

- Temporary buffered files now have a dynamically allocated temporary
  file name string attached. Previously, a statically allocated
  buffer was part of each file buffer data structure.

- The order of the hook parameters for the file descriptor and
  buffered files was wrong.

- All the buffered file routines now check if the file in question
  is actually in use before they proceed.

NOTE: the last four changes require a rebuild of the entire library,
      just to be safe!


c.lib 1.51 (12.8.2003)

- setvbuf() now sets the buffer size to 0 when switching a stream
  into unbuffered mode. Also, with buffering mode _IONBF (no
  buffering) no memory is allocated and no supplied buffer is
  used.

- The stdio cleanup code no longer runs all file descriptors and
  files through the shutdown code. Only those which are actually
  in use.

- Fixed a typo in the BSD style license text.

- Cosmetic changes...

- fopen("a+", ..) did not actually append any data.

- The unlink() cleanup code didn't unlock the temporary locks
  it maintained.


c.lib 1.50 (2.8.2003)

- Rewrote the entire Amiga->Unix and Unix->Amiga name/path translation
  code from scratch. It's now much more robust than it used to be.

- Restored the original difftime() code.


c.lib 1.49 (1.8.2003)

- vfprintf() now treats the '%' format specifier as an escape
  character and produces the next following letter as is if that
  letter does not indicate a specific format.


c.lib 1.48 (31.7.2003)

- vfprintf() now sets up the buffer it uses for converting numbers
  into text every time it goes through the main loop. This means
  that a terminating '\0' is placed at the end.

- vfprintf() now first checks if a floating point number is
  +/-infinity or not a number and then performs the '%g' check
  to figure out which output format should be used.

- Fixed the alternate conversion option for precision == 0 and
  the %e/%f/%g conversions.

- In vfprintf(), if there is no digit following the precision
  specifier, the precision is now set to 0.

- In vfscanf() only %c, %s and %[ respond to the maximum field
  width option, all other conversion methods ignore it.

- readdir() again returns the ".." entry for the root
  directory.


c.lib 1.47 (31.7.2003)

- Fixed vfscanf() and strtol() again to accept very small negative
  integers (-2147483648). The previous check to avoid overflows
  was too strict.


c.lib 1.46 (31.7.2003)

- strftime() now makes sure that the string generated is always
  NUL-terminated. The formatting code checks if the dates are
  in range to avoid accessing invalid day/month table data.


c.lib 1.45 (30.7.2003)

- lseek() could end up freeing a local buffer twice. Fixed.

- The sscanf() wrapper no longer emulates seek/write.

- Added some more error checking to the vasprintf() code.

- Fixed the %h qualifier in vfprintf().

- The %d conversion of negative integers in vfprintf() was relying
  upon a non-portable operation: the % operator used on negative
  numbers was expected to yield a negative result. The new code
  tries to be more careful and uses the % operator only on
  non-negative numbers.

- vfprintf() now handles floating point number exponents
  gracefully which are larger than 100.

- Made the varargs handling vfprintf() more portable.

- vfscanf() could produce an error if a character in the
  format string did not exactly match the corresponding
  character in the input stream. Now it just pushes the
  character back and stops scanning.

- The %c conversion in vfscanf() never stored more than
  one single character.

- The automatic %d octal number conversion in vfscanf() lost
  the first digit following the introducing '0'. Fixed.

- vfscanf() now properly handles the 0x/0X prefix for the
  %x conversion.

- vfscanf() now only returns EOF if no arguments could be
  converted and the EOF status of the stream involved has
  been asserted.

- vsnprintf() no longer tries to skip the trailing NUL byte if
  the resulting string is larger than the buffer it would
  hold.

- The memory allocation cleanup code now also properly handles
  the extra large debug allocations and cleans up the
  alloca() tracking list.

- Fixed strncpy() and strncat() again not to write too much
  data. Also strncat() does not pad the destination buffer.
  Neither function is supposed to do anything if the
  length parameter is negative.

- strncmp() now returns 0 if the length parameter is negative.


amiga.lib 1.7 (30.7.2003)

- None of the 68k assembly language files appear in the PowerPC build
  any more. Also, none of the functions which use the local CallHook()
  appear in the PowerPC build either.


c.lib 1.44 (28.7.2003)

- Files created through open() now get their 'executable' protection
  bits cleared when they are closed.

- mkdir() now sets the protection bits on the directory created.

- The record locking code now uses shared semaphore locking where
  necessary.

- Added workarounds for NIL: style file handles inthe record locking
  code.

- Preparations to get rid of the file descriptor use counter...

- __open_iob() no longer calls lseek().

- popen() and fdopen() now also ignore the extra access options
  separated by a comma.

- tmpnam() now creates names with up to 10 characters.

- Lost the special case code for zero length memory allocations.
  They now produce NULL result codes.

- Lost the DosLibrary, LocaleBase and ExecBase definitions where
  they were not actually required.

- strncat() didn't pad the destination string with NUL bytes,
  according to spec. Fixed.

- Fixed the weird strxfrm() to work according to spec.

- Simplified the asctime() and strftime() code.

- The "%j" format specifier did not produce exactly three digits
  in strftime(). Fixed.


amiga.lib 1.6 (28.7.2003)

- Removed NULL pointer checks from "amiga.lib". Added assertions in
  their places.

- Added tests for IconBase and CxBase to the "amiga.lib" routines.


c.lib 1.43 (27.7.2003)

- readlink() didn't free the DeviceProc structure it could use.

- More fixes to the record locking code. If a lock/unlock action was
  performed on a duplicate of a file descriptor, it wasn't always
  guaranteed that all the duplicate/original file descriptors were
  affected.

- Fixed a potential bug in the strncpy() code which could drop into
  an infinite loop.

- Replaced almost all do .. while loops with while .. loops.


c.lib 1.42 (27.7.2003)

- Changed the memcpy(), bcopy(), memmove() and bcopy() code
  not to use any assembly language bits. I'm wondering whether
  this will have an effect on overall stability.

- The internal number->string conversion routine used by the
  time.lib code had two of its parameters swapped. Fixed.

- The main() startup code had a bug in the character escape
  handling. Everything escaped other than 'e' or 'n' would
  terminate the respective command line parameter in place.

- The shell escape character (default is '*') is now configurable
  at compile time, which can help applications ported from the
  Unix world.Note that you cannot override the '*' escape character,
  you can merely define another valid escape character.

- Changed how double quotes are used by the main() startup code.
  A quoted string is now terminated if there is a blank space
  following the final double quote or if the end of the command
  line is reached.


c.lib 1.41 (26.7.2003)

- In vfscanf() the '%e', '%f' and '%g' conversions now default again
  to 'float' unless a format size specifier is used. Also, the function
  no longer 'forgets' that the default conversion is being used, which
  could happen if several parameters were to be converted.

- asctime() now tries to make sure that the conversion process
  does not overflow the local buffer.

- readdir() no longer returns the ".." entry for the root
  directory. Also simplified the readdir() cleanup code.

- open(), stat() and lstat() could leave dangling FileLocks behind. Fixed.

- dup2() now returns the new file descriptor rather than 0. It also
  tries to close the second descriptor if that one is in use and
  easy to close.

- fileno() never returned the associated file descriptor but the
  table index of the FILE stream. Fixed.

- Increased the internal maximum directory/file path length to
  1024 characters. Note that this means that the entire library
  will have to be rebuilt!

- setvbuf() now frees its custom allocated buffer if flushing the
  write buffer fails.

- setenv() now longer permits a NULL 'value' parameter.

- fdopen() now 'claims' a file descriptor rather than performing
  an 'internal' dup() on it.

- vfscanf() now stops converting the mantissa when it finds
  an exponent character in the input stream.

- vfscanf() now makes sure that the strings it reads are
  all NUL terminated while they are being read. If the
  conversion is aborted, this will still produce a properly
  NUL terminated string.

- The '%' conversion in vfscanf() no longer exits with an
  error if the next character read does not match the
  '%s' character.

- vfprintf() now watches the size of the buffer it uses for
  converting numeric data into strings.

- The file descriptor hook code now initializes the duplicate
  when it performs a dup() operation.

- The record locking support now also extends to duplicated
  file descriptors.

TODO: replace struct fd.fd_UseCount with a flag.


net.lib 1.8 (25.7.2003)

- The select() function was broken in a number of ways. Always was, even
  in its original form in the Samba 2.0.7 Amiga emulation library. First,
  the except_fd bits could have been set even for files (unlikely, if not
  impossible to happen). Second, if sockets and real files were used in
  parallel, sockets were preferred and the FD_SETs for the files would
  come out as indicating that data was available for reading/writing
  (which was wrong). Third, the test loop which checked whether the files
  had data waiting to be read never initialized the data structure it
  had to use. As a bonus, the new code now also checks if a file is
  writable and will then flag its FD_SET bit as "ready for writing".

- More fixes to select(): in each iteration of the polling code, the
  set bits would be cleared that might later be used to check for
  new input. This no longer happens.


c.lib 1.40 (25.7.2003)

- fopen() now ignores any open options following the mode string,
  separated by comma.


c.lib 1.39 (23.7.2003)

- Ouch! vfprintf() didn't work right for floating point number conversion
  (again). The '%g' conversion code didn't work on the PowerPC because
  the method employed to figure out which format to use was not portable.
  Fixed. The '%f' conversion code would produce no output at all if the
  number to be printed was zero and the number of digits following the
  comma was zero, too. Fixed. Large numbers didn't come ought right.
  Fixed. The '%e' conversion code was ill prepared for converting
  negative numbers and sometimes failed to produce digits. Fixed.

- More trouble. strtod() and, by extension, atof() didn't convert
  numbers properly with more than one digit following the mantissa.
  Fixed. Same stupid problem in vfscanf().

- The '%e', '%f' and '%g' conversions would always default to 'float'.
  They now default to 'double'. Use the 'h' or 'L' specifiers if you
  want less/more.


c.lib 1.38 (23.7.2003)

- The '%s' and '%[' conversion code in vfscanf() never NUL-terminated
  the string just read. Fixed.


amiga.lib 1.5 (15.6.2003)

- CreatePort() didn't fill in the port priority. Fixed.


amiga.lib 1.4 (1.5.2003)

- RangeRand() didn't obey the range parameter's restrictions. Fixed.


c.lib 1.37 (3.3.2003)

- Added an option to allow a program to detach itself from the shell
  it was launched from. This option is ignored for residentable
  programs.


c.lib 1.36 (14.2.2003)

- Removed redundant code from the routines which deal with buffered
  stdio streams. The respective routines now only test for the
  buffering type and ignore the buffer size altogether, just like
  it's supposed to work.

- Fixed the vfprintf() %e format output which no longer flips the
  sign of the number to be converted until it's safe to do so.
  The code to figure out whether %g should be mapped to %e or %f
  was unreliable because it could miscount the number of digits in
  front of the comma.

- Updated the code that dumps the stdio buffer in fcntl() before the
  set/get lock function is performed. These two should have nothing
  in common, though. Let's see what it does.


c.lib 1.35 (13.2.2003)

- When a stdio or plain unbuffered stream is opened, the respective data
  structures are reinitialized with default values. Previously, the old
  settings could persist and stale data could cause unallocated or
  reallocated data to be clobbered. This seems to have been the problem
  which eventually broke the Samba port.

- Added some more code to improve debug output quality.

- The buffered stdio routines now include some extra safety checking
  code in case the buffer data structures get clobbered. Not that this
  should happen again...


c.lib 1.34 (8.2.2003)

- The program startup code now opens dos.library and utility.library
  before it starts to figure out whether it has sufficient stack space
  available or not. This allows the stack size setup callback function
  to get its bearings. It also simplifies the startup code a bit.

- The UtilityBase and LocaleBase library bases will now be cached until
  the startup code performs its cleanup duties. This is intended to help
  application software which insists on reopening these libraries by
  itself all over again.

- The stdio initialization code now uses the shell's standard error output
  stream if one is available. Otherwise, it will open "CONSOLE:" instead,
  thereby cloning the console file handle.


c.lib 1.33 (2.2.2003)

- Rewrote the vfprintf() floating point conversion code not to rely
  upon floating point to integer conversions which could overflow the
  capacity of 32 bit integers. The new code now converts the individual
  digits as floating point numbers which then end up getting converted
  into integers, which should be safe enough.

- Simplified the fgets() code a bit.


c.lib 1.32 (26.1.2003)

- The fgets() EOF detection was flawed. Fixed [Andrea Valinotto].
  It now returns NULL if no data was read yet and an EOF condition
  was detected. Previously, it would keep returning an empty string.


c.lib 1.31 (14.1.2003)

- Rewrote the floating point conversion code in vfscanf(). The
  EOF handling for the data to be parsed was not robust enough.
  It aborted scanning altogether if the last data item also
  included the last character in the data stream.

- Now properly counts all characters read in vfscanf() even if
  some of them have to be pushed back.

- Invalid character matches are now pushed back by vfscanf().


c.lib 1.30 (28.12.2002)

- ungetc() now returns the character pushed back, or EOF if it
  could not push anything back. It also clears the end-of-file
  indicator.

- ftell() no longer drops the read/write buffer contents. It now
  checks where the current file access pointer is and then adjusts
  that position with regard to what's in the read/write buffer.


net.lib 1.7 (28.12.2002)

- The socket file descriptor hook now rejects seek accesses,
  setting the proper errno code ESPIPE.


net.lib 1.6 (21.12.2002)

- Turns out that the daemon socket bound to the stdio streams must be
  marked as available for reading and writing or Samba won't work.


stack.lib 1.2 (18.12.2002)

- Added code that collects information on how much memory is allocated
  for stack extension. That data is printed by the debug library before
  it exits.


c.lib 1.29 (18.12.2002)

- strxfrm() and strcoll() did not use the locale.library pragma/inline
  header files and thus would not link properly. Fixed.

- Changed the entry assertions for bcopy(), bsearch(), memcpy(),
  memmove(), memset() and bzero() to ignore zero length arrays.
  This fixes the GCC port which had alloca(0) followed by bzero(..,0)
  in it.

- malloc() and alloca() now return NULL again for a zero length
  allocation. Not sure if this is the way to go...


c.lib 1.28 (14.12.2002)

- setenv() trashed the variable value to be used if the variable and
  value names were given in the "variable=value" form. Fixed.

- Reworked the overflow checking for strtol() and the scanf() family
  again. This time it hopefully works as expected.

- Trying to allocate a memory chunk of size 0 now returns a pointer
  to a unique address rather than NULL. I wonder whether this is
  really the right thing to do, but it seems to help the compiler.

- Calling realloc() on a memory chunk with exactly the same size it
  was allocated with now returns a pointer to that chunk rather than
  NULL.

- The alloca() memory cleanup code is now triggered from within the
  library by a separate function which can be told apart from user
  code making the call.

- Now remembers which local variables were set with putenv()/setenv()
  and makes sure that before it exits that they all get removed first.


c.lib 1.27 (13.12.2002)

- The strtol() fix didn't actually fix anything broken, but broke some
  working things instead. Fixed.

- Reworked the overflow checking code in scanf() and strtol() again to
  compile without producing warnings.


c.lib 1.26 (13.12.2002)

- strtol() did not set the converted number to the smallest possible
  negative integer if the number to convert had a sign. Fixed.


c.lib 1.25 (13.12.2002)

- Added special case code to handle conversion of the smallest long
  integer -2147483648 from ASCII string to number.


c.lib 1.24 (12.12.2002)

- vfprintf() couldn't convert -2147483648 into a string. This is what
  caused GCC to fail rebuilding itself. Fixed.


c.lib 1.23 (12.12.2002)

- setvbuf() now consistently accepts a size value of 0, whereupon
  it will default to "no buffering" mode.

- Ouch. My "optimization" for unbuffered streams did not work if the
  buffer size was set to zero. Fixed.


c.lib 1.22 (12.12.2002)

- The buffered file I/O routines no longer take care of appending
  data to the end of the file. This is now done entirely by the
  file descriptor layer.

- fread() could in unbuffered mode end up reading partial elements
  from the file. It now backs up to change the file position to take
  the exact number of records into account that should have been
  read.

- setvbuf() now frees the custom file buffer before replacing it
  with a user supplied buffer.

- stderr is no longer buffered by default (as it should be). However,
  optimizations are in place to make vfprintf() and fputs() queue their
  output until they return.


net.lib 1.5 (12.12.2002)

- Replaced the inline header files.

- When launched as a daemon from inetd or similar, the daemon socket
  passed to the startup code is now adopted for stdin. Previously,
  three new sockets were duplicated from the daemon socket. Also,
  the three stdio streams are no longer buffered by default. The
  file descriptor setup now matches the standard I/O setup in that
  stdin can be read from and stdout/stderr can be written to.


c.lib 1.21 (10.12.2002)

- The path name translation code could end up clobbering the original
  string. Fixed.

- The command name translation in system() did not work right. It
  ended up translating the entire string which bombed with very
  long command lines. Same problem for popen(), same fix.

- Updated the Unix name translation code to bail out of the string
  to translate is too long to be translated in the first place.
  Previously, it just trashed the stack :(

- system() now calls SystemTagList() with SYS_UserShell=TRUE.

- Fixed up the header files so that function pointers for the debug
  routines actually work. This became apparent when I was trying to
  rebuild the binutils with the clib2.

- The <sys/time.h> header file no longer pulls in <devices/timer.h>
  by default. If it hasn't been included yet, then only the timeval
  structure is defined.


c.lib 1.20 (8.12.2002)

- The stdio buffered file I/O routines now clear their error indicators
  before work begins.

- Simplified the code to back the lseek() function. It no longer calls
  Seek() twice if the file position does not change.

- fgetc() and fputc() now set errno to EBADF if the file in question
  cannot be read from or written to, respectively.


amiga.lib 1.3 (8.12.2002)

- The name of the memory pool is now placed at an odd memory location.


c.lib 1.19 (4.12.2002)

- index()/rindex()/strchr()/strrch() now perform signed byte
  comparisons. These used to be unsigned.

- Added a cast to memchr() to make sure that the comparison
  is performed on unsigned bytes.

- fseek() and fsetpos() now clear the file EOF condition.


c.lib 1.18 (29.11.2002)

- ftruncate() no longer works for files which aren't open for writing.

- Simplified the __show_error() code.

- When allocating a custom buffer, setvbuf() didn't actually fill in
  the allocated buffer's address, causing data to be trashed. Fixed.
  Setting the buffer size to zero now works only for the
  "don't buffer" mode. If the buffer size was changed, the previous
  buffer was not flushed to disk before.

- If an attempt to flush a file write buffer fails, the next attempt
  to put data into the buffer will retry flushing the buffer.
  Previously, data would have been written beyond the bounds of the
  buffer.

- If fgets() hits an error, it now NUL terminates the string
  to be on the safe side.

- Lost the (disabled) asyncio code.

- The startup file for SMALL_DATA32 mode did not set up register A4
  at all. Ouch.


amiga.lib 1.2 (29.11.2002)

- Fixed the RVI code to look & work more like the original implementation.


m.lib 1.2 (7.11.2002)

- pow() called the IEEEDPPow() LVO with the x/y parameters reversed.
  Actually, IEEEDPPow() seems to be the one to blame [Andrija Antonijevic].

- __fixunsdfsi() and __fixunssfsi() had a cast in the wrong place.
  The original libnix functions these were based on were still
  correct [Andrija Antonijevic].


m881.lib 1.2 (7.11.2002)

- __fixunsdfsi() and __fixunssfsi() had a cast in the wrong place.
  The original libnix functions these were based on were still
  correct [Andrija Antonijevic].


c.lib 1.17 (7.11.2002)

- ftell() now returns the current file position rather than invoking
  fseek(), which does something different altogether [Andrija Antonijevic].

- fgetpos() now calls ftell() [Andrija Antonijevic].

- fsetpos() didn't properly evaluate the return value of fseek()
  which would have been -1 in case of error [Andrija Antonijevic].

- fseek() now returns 0 for success and -1 for error [Andrija Antonijevic].


c.lib 1.16 (5.11.2002)

- memcpy() and memmove() had a debug switch wired the wrong way
  round, causing the wrong kind of code to be emitted. Code which
  could call exec.library/CopyMem and shouldn't have called it.
  Fixed. It no longer depends upon exec.library.

- Removed all the abort signal routines from the string formatting
  code. This kind of action is now done solely by the I/O routines.


c.lib 1.15 (4.11.2002)

- sscanf() ended up calling vfprintf(). Ouch.

- Practically none of the numeric conversion operations in the scanf()
  family actually worked. Fixed.

- scanf() and friends now push a character back upon detecting an
  overflow rather than using it and setting the number converted to
  MAXINT.


c.lib 1.14 (21.10.2002)

- Rewrote strcasecmp() and strncasecmp() to work more like strcmp() and
  strncmp(). Previously, these functions would drop into the utility.library
  routines Stricmp() and Strnicmp() which don't work like strcmp() or
  strncmp(), respectively.

- Rewrote the code that closes file descriptors. It should now be much smarter
  in handling duplicates.

- fchmod() invoked the wrong hook action which could lead to a crash.

- Added wrapper code for ExamineFH() to avoid crashing due to a bug in
  dos.library V40 and below with NIL: file handles.

- The temporary directory locks set up for temporary files were never
  unlocked.

- Submitting an untracked memory region to be freed did not produce the
  intended error output.


c.lib 1.13 (7.10.2002)

- The vfprintf() family didn't produce any sign characters at all unless
  a field width specifier was used [Andrija Antonijevic].


c.lib 1.12 (28.9.2002)

- The vfprintf() family got the hexadecimal number conversion (and the pointer
  conversion, too) wrong because the conversion string had an extra zero digit
  in it [Detlef Wrkner].


c.lib 1.11 (15.9.2002)

- In the printf() family, the default case for '%c' (plain character
  output without padding) did not update the length of the string
  created [Andrija Antonijevic].


c.lib 1.10 (11.9.2002)

- In gettimeofday() the important DateStamp() call was missing [Andrija
  Antonijevic].

- Removed the replacement remainder calculation code. It's not necessary.


c.lib 1.9 (7.9.2002)

- Added replacement code for the two routines the updated compiler will call
  in place of the inline '020 math which cannot be used on the '060. This
  ought to be done smarter, but this works, too.


c.lib 1.8 (28.8.2002)

- The open() code now avoids calling Lock() on NIL:, which otherwise could
  have caused the file open action to follow later to be skipped.

- Added more compiler internal code to support SAS/C.

- Increased the size of the memory debugging walls.

- When reading data off stdin character by character, the stdout and
  stderr streams are now flushed first.


c.lib 1.7 (27.8.2002)

- Discovered that ExamineFH() and ChangeMode() will crash on file
  handles which refer to "NIL:". Put a workaround for these into
  the open() code.


c.lib 1.6 (27.8.2002)

- With the standard I/O streams the vfprintf() code would still add a
  terminating NUL byte to each string printed. Fixed.

- The open() code now checks before it opens if the file in question
  can be overwritten, should this be necessary. Likewise, once the
  file in question has opened, its protection bits are examined to
  verify that the access mode it has been opened for is applicable
  in the first place.

- The asynchronous I/O option now works only for O_RDONLY and O_WRONLY
  mode access.


c.lib 1.5 (25.8.2002)

- fread() treated EOF as an error, returning with 0 characters read
  from the buffer even if there was still some data in the buffer that
  could be read.

- fread() now seeks back to the last location it could read an entire
  element from. Subsequent fgetc() calls will thus pick up the data
  not processed by the previous fread() call. Also, if fread() encounters
  an error, it seeks back to where the read was supposed to begin.


c.lib 1.4 (22.8.2002)

- Updated the ctype_headers.h, locale_headers.h, math_headers.h,
  signal_headers.h, stdio_headers.h, stdlib_headers.h, string_headers.h,
  strings_headers.h and time_headers.h header files. The extern library
  definitions are now in the SAS/C branch.

- Moved the __is_daemon definition from unistd.h into dos.h and changed
  its type as well.


net.lib 1.4 (22.8.2002)

- Updated the SAS/C compatible prototypes and #pragma definitions.


c.lib 1.3 (22.8.2002)

- Files opened in exclusive access mode are now switched back into
  shared access mode for testing.

- The asynchronous file I/O is now enabled only for read and write
  access. Appending is disabled.

- You now have more control over the asynchronous file I/O and can
  enable it on demand and specify the default buffer size.

- You can now provide for a callback function that initializes the
  default stack size to use early on during the runtime configuration
  process.


net.lib 1.3 (22.8.2002)

- getgroups() and setgroups() now ignore the second parameter if the
  number of groups involved is zero.

- Upon startup, the runtime environment now sets a global variable
  to 1 if the current program was launched by the internet superserver
  or an equivalent facility.


c.lib 1.2 (21.8.2002)

- Added sigaddset() and sigemptyset().

- isxdigit() didn't work. Fixed.


net.lib 1.2 (21.8.2002)

- Added setpwent(), getpwent(), endpwent(), setgrent(), getgrent() and
  endgrent()


----------------------------------------------------------------------------

amiga.lib 1.1 (20.08.2002)

- Now bears a version number.


c.lib 1.1 (20.08.2002)

- Now bears a version number.


debug.lib 1.1 (20.08.2002)

- Now bears a version number.


m.lib 1.1 (20.08.2002)

- Now bears a version number.


m881.lib 1.1 (20.08.2002)

- Now bears a version number.


net.lib 1.1 (20.08.2002)

- Now bears a version number.


stack.lib 1.1 (20.08.2002)

- Now bears a version number.


unix.lib 1.1 (20.08.2002)

- Now bears a version number.


----------------------------------------------------------------------------

- '%f' floating point output now also works for numbers > 9.

- <sys/stat.h> needs to include <time.h> and <stddef.h>.

- <sys/types.h> now only defines a bunch of data types, it no
  longer includes any header files.

- Activated the local bcopy() and ovbcopy() code.

- Added ACrypt().

- Added CreateTask() and DeleteTask().

- Added Sebastian's fix for vsprintf(), which did not generate
  a '-' at all appropriate times.

- Added Unix-style directory scanning with pseudo "/" directory
  support.

- Added __stdio_configure_streams() to reconfigure the stream
  I/O after the Workbench output window has been opened.

- Added a local NewList() replacement.

- Added a plain 68k version of the setjump()/longjmp() pair.

- Added an extra call to __check_abort() after each
  bsdsocket.library call. Because, if a ^C was registered,
  causing the respective function call to return, the signal
  will still be set, waiting to be picked up.

- Added assertions to vfscanf().

- Added basename(), dirname(), asprintf(), snprintf(),
  vasprintf(), vsnprintf(), mkdtemp(), mkstemp(), mktemp(),
  putenv(), setenv(), unsetenv(), strcasecmp(), strncasecmp(),
  strdup(), gettimeofday() and alloca().

- Added data and code to support SAS/C stack overflow checking
  and abort.

- Added further ^C checking to closedir(), opendir(), readdir(),
  rewinddir(), creat(), open(), read(), write(), chmod(), fchmod(),
  fstat(), lstat(), mkdir(), rmdir(), stat(), fprintf(), fputs(),
  ftell(), getc(), getchar(), gets(), __open_iob(), printf(),
  putc(), putchar(), puts(), rewind(), setbuf(), sprintf(),
  tmpfile(), tmpnam(), vfprintf(), __vfscan(), vprintf(),
  vsprintf(), getenv(), system(), strtod(), strtol(), strtoul(),
  strftime(), access(), chdir(), chown(), fchown(), fileno(),
  ftruncate(), getcwd(), getopt(), lchown(), link(), readlink(),
  sleep(), symlink(), truncate(), unlink() and utime().

- Added getopt(), kill() and getpid().

- Added grp.h and pwd.h header files.

- Added isatty().

- Added more assertions and tests to detect illegal parameters
  earlier than before.

- Added more assertions to strftime().

- Added more casts to (size_t) where appropriate.

- Added more memory debugging hooks to other routines which
  implicitly allocate memory and return it.

- Added more of the GCC specific support routines.

- Added new __get_a4() and __set_a4() routines to the startup
  code. Also put them into the <dos.h> header file.

- Added platform specific <dos.h> header file.

- Added popen() and pclose().

- Added some more assertions to qsort().

- Added some of the bsdsocket.library stubs required for
  Samba 2.0.7 to work.

- Added stack size checking support.

- Added stack swapping code.

- Added statfs() and fstatfs().

- Added stub code to handle double precision IEEE math.

- Added support code to emulate the behaviour of debug.lib.

- Added the Amiga/Unix name translation code.

- Added the GCC specific floating point runtime library routines.

- Added the GCC specific startup code.

- Added the README file.

- Added the amiga.lib RVI code.

- Added the hook and method functions from amiga.lib.

- Added the math support code.

- Added the memory allocation debugging code.

- Added the memory pools code to the amiga.lib set.

- Added the monitoring and debugging code.

- Added the remaining few amiga.lib 'C' library routines. The
  memory pools stuff is probably hopeless, but we'll what I can do.

- Added the socket I/O wrapper.

- Added the source code comments to the memory pools source code.

- Added the stack extension code. It still has to be updated to
  be more readable, though.

- Added the wide character stubs.

- Added umask().

- Added usergroup.library support.

- All debug output now uses kprintf() again.

- All memory allocations now use malloc() and free().

- Changed the manner in which the floating point support code is
  compiled into the library.

- Changed the name of the preprocessor symbol which activates
  the Unix file name semantic support.

- Changed the names of unused function parameters to match their uses.

- Changed the way the library is built. We now build seven
  different component libraries which all must be linked
  separately.

- Conditional branches no longer bear an 's' or 'w' suffix.
  The assembler is supposed to find out all by itself which
  one should be picked.

- File descriptors and I/O buffers now receive the respective
  data structures as part of the hook calling packet rather
  than having to probe the h_Data contents.

- Finally figured out what the %n format specifier is supposed
  to accomplish in fprintf().

- Finally figured out what the __main() reference was good
  for and implemented construtor and destructor support.

- Fixed a bunch of compiler warnings.

- Fixed more than one bug in sleep(). This time the code
  hopefully works as advertized.

- Fixed several bugs GCC highlighted.

- Fixed two more compiler warnings related to signed/unsigned
  comparisons.

- For Workbench startup, the stderr stream is now merely an alias
  for the stdout stream.

- Forgot the SAS/C CallHookA() stub.

- Forgot to add the stubs for the socket and usergroup components.

- Found and worked around two more code generation problems with
  WaitSelect() and recvfrom().

- GMT+0 always came out as GMT+ with strftime(). It now stays GMT.

- If possible, memory pools are used for memory management.

- In <signal.h> we now define sigset_t rather than in <sys/types.h>.

- In CreateTask(), the recovery from a failed AllocEntry() call
  was flawed. Reread the exec source code and fixed that bug.

- In alloca(), no unused memory was ever freed.

- In listen() a test to see if the file descriptor was really a
  socket was missing.

- In memmove(), the local ovbcopy() code is now smarter about the
  registers it has to save.

- In stat(), lstat() and fstat() the st_dev structure member is
  now initialized with the address of the file system's MsgPort
  the respective directory/file is stored on. This allows file
  descriptors to be compared.

- In strstr() there is now a test to make sure that the substring
  is smaller or equal in size to the big string to be searched in.

- In vfprintf() the '%g' output format always came out as '%e' if
  the precision was uninitialized. Fixed.

- Instead of passing local variables around to show errors or to
  display the current program name, the whole code now uses global
  variables.

- Integrated a modified version of the Commodore asynchronous file
  I/O code into the library, mainly for the benefit of Samba.

- Integrated the default stack size checking variables into the
  makefiles.

- Integrated the extra super elaborate advisory file locking code
  which worked so well with Samba. Or did it? With the new library
  it appears that opportunistic locks don't work any more. I'm not
  so sure if they ever did :(

- Knocked up replacement functions for amiga.lib and SAS/C runtime
  library functions.

- Looks like Samba needs the super elaborate locking method after
  all. Oh well...

- Modified the startup code and the assembly language fragments
  to support the -fbaserel32 option. Built Samba 2.0.7 with it,
  which actually seems to work OK.

- Modified the unistd.h header file to avoid clashes with the
  TCP/IP header files.

- More components are now part of libunix.a to make sure that the right replacement functions are linked in.

- Moved NewList() into its own source file. This is supposed to
  be more compatible with amiga.lib.

- Moved all the functions which are unimplemented out of the build
  files. The old build makefiles are still around, though.

- Moved the fundamental file descriptor I/O into the stdio branch.
  The fcntl branch now only contains stubs.

- Moved the locale references made in strtod() into the math
  libraries, too.

- Moved the macro definitions into a separate header file.

- Moved the string hook code into sscanf().

- No longer uses the cooked version of <assert.h> for the build.
  A local "debug.h" header file is used instead.

- Now actually compiles and links with GCC.

- Now builds code for the 68020-60 target rather than 68020-40.

- Now builds with SAS/C again.

- Only a single error message should appear during startup, not
  a long list of messages for each problem that might appear.

- Oops. Forgot the constructor/destructor defaults for GCC.

- Oops. One variable was uninitialized.

- Plugged in the optimized assembly language data copying code.

- Preparations for %ll support in vfprintf() and vfscanf().

- Rebuilt with GCC, which revealed a problem in calling
  AutoRequest(). Fixed.

- Rebuilt with new GCC inline macros which yield better code.

- Redid the BOOPSI functions in 'C'.

- Redid the qsort() code, borrowing snippets from libnix2.

- Reduced the complexity of fseek() a bit.

- Renamed stdio_unixsemantics.c to stdio_unixbehaviour.c.

- Reorganized the libdebug.a source code. Each file now contains
  only a single routine.

- Replaced all unconditional branches with jbra.

- Restored the old advisory record locking code to see if it
  makes a difference.

- Reviewed the stack extension code and added comments. Now
  uses pooled memory allocations instead of AllocMem()/FreeMem()
  if possible.

- Reworked the header file set to key off the compiler used.

- Reworked the popen() initialization code.

- Rewrote the ncrt0.S and nrcrt0.S startup code again, trying to
  make it simpler.

- Rewrote the setjmp()/longjmp() code to also save and restore
  the floating point register set if there is an FPU installed.

- Separated the code paths for alloca(), malloc(), realloc()
  and free(). They all call intermediate routines to get their
  jobs done.

- Simplified error return codes in many components.

- Simplified the build makefile.

- Stack swapping appears to work.

- Switched compiler options to large data/large code to keep
  SAS/C out of my hair.

- Tested the Unix name semantics translation.

- The 'e', 'f' and 'g' formatting commands produce no output
  the floating point support code is not built into the library.

- The Amiga/Unix name translation routines now treat "" and "."
  special.

- The FLOATING_POINT_SUPPORT macro definition never checked for
  the M68881_FLOATING_POINT_SUPPORT symbol. Fixed.

- The HUGE_VAL constant now actually refers to a constant floating
  point number which the startup code will initialize with 'positive
  infinity'.

- The IEEE math code now compiles properly.

- The IEEE math libraries are now opened requiring version 33 or
  higher. Earlier versions did not support floor() or ceil().

- The Unix name translation code now allows for files in /dev to
  be accessed. /dev/null, however, is still mapped to NIL:.

- The __swap_stack_and_call() function never actually loaded A6
  with SysBase. Guess I was lucky...

- The _main() startup code now displays error messages and is more
  careful about mucking with the program's streams and current
  directory configuration.

- The abort() and raise() code now shares the same error message.

- The abort() code now drops into _exit() rather than relying
  upon raise() to do this.

- The amiga_rexxvars code no longer uses 'fp' in place of 'a5'.

- The asyncio.c code now uses the local bcopy() routine rather
  than calling into CopyMem(). The local bcopy() is a lot faster.

- The atexit() processing now takes care of all registered
  functions, even if exit() is called by one of them.

- The constructor/destructor support code could crash if no
  such functions were present in the binary. Fixed.

- The conversion of the program startup parameters to a command
  vector now works as expected.

- The debug.h header file now also works with GCC (sort of).

- The file descriptor duplication is now done in the respective hook.

- The fundamental lseek() code could trash the current file
  position. Fixed.

- The kgetnum() code did not properly update the character
  position for the sign and the digits. Fixed.

- The libdebug.a code is now fully self contained, so that it may
  be linked in last.

- The log tag initialization in the socket setup code did not work
  properly. Fixed.

- The math routines are now compiled only if the preprocessor
  symbols are set up for them to build.

- The name translation routine in readlink() changed the wrong name.
  Fixed.

- The new test code also prints an integer and a rational number
  to test floating point output.

- The non-blocking console I/O flag was not reset as it should have
  been. Fixed.

- The numeric and collation locale references were still using
  the old locale variables. Fixed.

- The numeric conversion routine used by the time code did not
  generate proper ASCII digits. Fixed.

- The old record locking code needed to be integrated better.

- The pseudo-root directory containing all volumes is now scanned
  entirely within readdir(); opendir() now only collects the
  available volume names.

- The raw I/O routines no longer reference the SysBase variable.

- The return code of rename() was always the opposite of what
  it really should have been. Fixed.

- The socket code is now compiled conditionally.

- The special behaviour of remove() and rename() is now a compile
  time option.

- The stack extension setup and cleanup routines are actually
  called now.

- The stack size checking code is no longer built if the stack
  extension code is to be used.

- The stdlib initialization code now attempts to create a memory
  pool. This is done exactly once.

- The time conversion code no longer drops into assert() if the
  time stamp is not within the supported Amiga time window.

- The unlink() and rename() functions can now be configured to
  perform vaguely like on a Unix system.

- There is still something wrong with the built-in functions, so
  I disabled them in the makefile.

- Tried to find a way to allow for ":" to pass through the Unix
  file name translation process without being mistaken for the
  pseudo-root directory.

- Triggering a call to exit() while the program is already on the
  way out should now be considerably safer than it used to be and
  no longer end in recursion.

- Turned out that in order to be residentable, the constant
  __huge_val cannot actually be replaced at runtime with positive
  infinity. So the __huge_val is no longer constant, but a cast in
  <math.h> forces it to be.

- Turns out that the GNU linker visits each library only once.
  Bummer. Had to reorganize the source code again and now libnet.a
  contains only those special versions of the file access routines
  which use Unix-like semantics and behaviour.

- Updated several instances of signed vs. unsigned comparisons.

- Updated the GNUMakefil to be more portable and to build all the
  library flavours necessary.

- Updated the object file list.

- When using socket I/O as a daemon, no error messages are sent to
  stderr by abort() or raise().

- Whoops. Overlooked two files which should have gone into the
  repository much earlier.

- Workbench startup now opens a default output window.

- __fixunsdfsi() and __fixunssfsi() didn't actually do anything
  useful. Fixed.

- __locale_exit() should no longer close the same locale twice.

- __open_iob() now truncates files for "w+" mode.

- __stdio_exit() and close() should handle closing of an aliased
  file descriptor properly now.

- _main() now checks how much stack space was actually used and
  prints that number.

- access() now always returns an error if a file does not exist.
  This fixes open(), which could fail if a file did not exist
  even though it should have failed if it existed.

- assert() now prints the program name.

- assert() now throughs an Enforcer hit to make bug tracking easier.

- assert() now uses kprintf() to print its messages.

- bcopy() no longer calls into memcpy().

- bzero() no longer calls into memset().

- clock() did not work properly. Fixed.

- free() and malloc() now mangle the contents of the memory
  buffer before they return unless the NDEBUG symbol is defined.

- free() now also invokes the alloca() drop code.

- fseek() could end in an assertion if no data had been read from
  the file yet. Fixed.

- fstat() now also works on sockets.

- getcwd() always trashed memory if "/" was returned. Fixed.

- getcwd() now takes the / directory into account.

- index() and rindex() no longer invoke strchr() and strrch().

- lseek() returns the file position reached after the file
  pointer has been repositioned.

- mktemp() always changed the wrong string. Fixed.

- mktemp() ran over the end of the buffer, only that this didn't
  really matter up until now.

- popen() didn't actually compile. Fixed.

- printf() and friends no longer attempt to NUL terminate streams.

- qsort() and bsearch() no longer drop into abort() if their
  invalid parameters don't matter because there is no data to
  examine.

- readlink() now attempts to let dos.library resolve the path
  and then returns an absolute path name. If that doesn't work,
  it invokes ReadLink() instead.

- realloc() no longer needs to know about the manner in which
  the allocation itself is managed.

- remove() now calls unlink().

- rename() now returns -1 on error, not 0 as it could have done
  accidentally.

- setvbuf() doesn't really need to go through the memory
  allocation debugging code. The memory allocated isn't visible
  to the outside.

- size_t parameters are now tested to see if they might have been
  negative integers originally.

- sleep() tries to be nicer about the timer I/O request in progress.

- snprintf() now returns the number of characters that would
  have been written to the buffer, now matter how many could
  really be stored.

- stat() now makes up pseudo data for the pseudo root directory.

- statfs() now also knows how to treat the pseudo root directory.

- strchr() and strrchr() now perform their comparisons on
  unsigned characters.

- strings.h didn't actually compile. Fixed.

- time() did not add any seconds to the current time. Fixed.

- tmpnam() could return with AmigaDOS requesters disabled for
  the shell. Fixed.

- unlink() now checks if the file to be deleted has been set up
  for deletion before.

- vfprintf() now uses fputc() to dump strings rather than calling
  fwrite(). This has the effect of flushing the buffer properly
  if there are line feed characters in the string specified for %s.
