summaryrefslogtreecommitdiff
path: root/NEWS
diff options
context:
space:
mode:
authorrpj <rpj>2005-04-25 14:55:16 +0000
committerrpj <rpj>2005-04-25 14:55:16 +0000
commita06bad7ea9a24a3ea93fe1e7b9b99155481ada3a (patch)
tree5d8d210afdda4bdaea3c6a6e4a7a7ccdab4b3bba /NEWS
parent262895c989b395cb0c79512a34a697b1e56bcbd7 (diff)
''
Diffstat (limited to 'NEWS')
-rw-r--r--NEWS1786
1 files changed, 893 insertions, 893 deletions
diff --git a/NEWS b/NEWS
index 61560ff..39950a9 100644
--- a/NEWS
+++ b/NEWS
@@ -1,893 +1,893 @@
-RELEASE 1.7.0
--------------
-(2005-04-12)
-
-General
--------
-
-This release is a backport of features and bug fixes new in
-release 2.3.0. See earlier notes under Release 1.4.0/General.
-
-Bugs fixed
-----------
-
-* Fixed pthread_once potential for post once_routine cancellation
-hanging due to starvation. See comments in pthread_once.c.
-Momentary priority boosting is used to ensure that, after a
-once_routine is cancelled, the thread that will run the
-once_routine is not starved by higher priority waiting threads at
-critical times. Priority boosting occurs only AFTER a once_routine
-cancellation, and is applied only to that once_control. The
-once_routine is run at the thread's normal base priority.
-
-New tests
----------
-
-* once4.c: Aggressively tests pthread_once() under realtime
-conditions using threads with varying priorities. Windows'
-random priority boosting does not occur for threads with realtime
-priority levels.
-
-
-RELEASE 1.6.0
--------------
-(2005-04-04)
-
-General
--------
-
-This release is a backport of features and bug fixes new in
-release 2.2.0. See earlier notes under Release 1.4.0/General.
-
-* Added makefile targets to build static link versions of the library.
-Both MinGW and MSVC. Please note that this does not imply any change
-to the LGPL licensing, which still imposes specific conditions on
-distributing software that has been statically linked with this library.
-
-Bugs fixed
-----------
-
-* Fix integer overrun error in sem_timedwait().
-Kevin Lussier
-
-* Fix preprocessor directives for static linking.
-Dimitar Panayotov
-
-
-RELEASE 1.5.0
--------------
-(2005-03-16)
-
-General
--------
-
-This release is a backport of features and bug fixes new in
-release 2.1.0. See earlier notes under Release 1.4.0/General.
-
-Bugs fixed
-----------
-
-* Reverse change to pthread_setcancelstate() made in 1.4.0.
-
-
-RELEASE 1.4.0
--------------
-(2005-03-16)
-
-General
--------
-
-Version 2.0.0 is being released and version 1.4.0 back-ports the
-functionality included in that release. In future, bug fixes
-will be back-ported if possible and where appropriate, however,
-this will be at the maintainer's descretion. New features will
-not generally be back-ported, although some may if they
-provide major improvements in compatibility with the standard.
-
-This is intended as a courtesy to application developers, but as
-time moves on, the imperative will fade.
-
-The package naming has changed, replacing the snapshot date with
-the version number + descriptive information. E.g. this
-release is "pthreads-w32-1-4-0-release".
-
-Bugs fixed
-----------
-
-* pthread_setcancelstate() no longer checks for a pending
-async cancel event if the library is using alertable async
-cancel. See the README file (Prerequisites section) for info
-on adding alertable async cancelation.
-
-New features
-------------
-
-* pthread_once() now supports init_routine cancellability.
-
-New tests
----------
-
-* Agressively test pthread_once() init_routine cancellability.
-
-
-SNAPSHOT 2005-03-08
--------------------
-Version 1.3.0
-
-Bug reports (fixed)
--------------------
-
-* Implicitly created threads leave Win32 handles behind after exiting.
-- Dmitrii Semii
-
-* pthread_once() starvation problem.
-- Gottlob Frege
-
-New tests
----------
-
-* More intense testing of pthread_once().
-
-
-SNAPSHOT 2005-01-25
--------------------
-Version 1.2.0
-
-Bug fixes
----------
-
-* Attempted acquisition of a recursive mutex could cause waiting threads
-to not be woken when the mutex was released.
-- Ralf Kubis <RKubis at mc.com>
-
-* Various package omissions have been fixed.
-
-
-SNAPSHOT 2005-01-03
--------------------
-Version 1.1.0
-
-Bug fixes
----------
-
-* Unlocking recursive or errorcheck mutexes would sometimes
-unexpectedly return an EPERM error (bug introduced in
-snapshot-2004-11-03).
-- Konstantin Voronkov <beowinkle at yahoo.com>
-
-
-SNAPSHOT 2004-11-22
--------------------
-Version 1.0.0
-
-This snapshot primarily fixes the condvar bug introduced in
-snapshot-2004-11-03. DLL versioning has also been included to allow
-applications to runtime check the Microsoft compatible DLL version
-information, and to extend the DLL naming system for ABI and major
-(non-backward compatible) API changes. See the README file for details.
-
-Bug fixes
----------
-
-* Condition variables no longer deadlock (bug introduced in
-snapshot-2004-11-03).
-- Alexander Kotliarov and Nicolas at saintmac
-
-* DLL naming extended to avoid 'DLL hell' in the future, and to
-accommodate the ABI change introduced in snapshot-2004-11-03. Snapshot
-2004-11-03 will be removed from FTP sites.
-
-New features
-------------
-
-* A Microsoft-style version resource has been added to the DLL for
-applications that wish to check DLL compatibility at runtime.
-
-* Pthreads-win32 DLL naming has been extended to allow incompatible DLL
-versions to co-exist in the same filesystem. See the README file for details,
-but briefly: while the version information inside the DLL will change with
-each release from now on, the DLL version names will only change if the new
-DLL is not backward compatible with older applications.
-
-The versioning scheme has been borrowed from GNU Libtool, and the DLL
-naming scheme is from Cygwin. Provided the Libtool-style numbering rules are
-honoured, the Cygwin DLL naming scheme automatcally ensures that DLL name
-changes are minimal and that applications will not load an incompatible
-pthreads-win32 DLL.
-
-Those who use the pre-built DLLs will find that the DLL/LIB names have a new
-suffix (1) in this snapshot. E.g. pthreadVC1.dll etc.
-
-* The POSIX thread ID reuse uniqueness feature introduced in the last snapshot
-has been kept as default, but the behaviour can now be controlled when the DLL
-is built to effectively switch it off. This makes the library much more
-sensitive to applications that assume that POSIX thread IDs are unique, i.e.
-are not strictly compliant with POSIX. See the PTW32_THREAD_ID_REUSE_INCREMENT
-macro comments in config.h for details.
-
-Other changes
--------------
-Certain POSIX macros have changed.
-
-These changes are intended to conform to the Single Unix Specification version 3,
-which states that, if set to 0 (zero) or not defined, then applications may use
-sysconf() to determine their values at runtime. Pthreads-win32 does not
-implement sysconf().
-
-The following macros are no longer undefined, but defined and set to -1
-(not implemented):
-
- _POSIX_THREAD_ATTR_STACKADDR
- _POSIX_THREAD_PRIO_INHERIT
- _POSIX_THREAD_PRIO_PROTECT
- _POSIX_THREAD_PROCESS_SHARED
-
-The following macros are defined and set to 200112L (implemented):
-
- _POSIX_THREADS
- _POSIX_THREAD_SAFE_FUNCTIONS
- _POSIX_THREAD_ATTR_STACKSIZE
- _POSIX_THREAD_PRIORITY_SCHEDULING
- _POSIX_SEMAPHORES
- _POSIX_READER_WRITER_LOCKS
- _POSIX_SPIN_LOCKS
- _POSIX_BARRIERS
-
-The following macros are defined and set to appropriate values:
-
- _POSIX_THREAD_THREADS_MAX
- _POSIX_SEM_VALUE_MAX
- _POSIX_SEM_NSEMS_MAX
- PTHREAD_DESTRUCTOR_ITERATIONS
- PTHREAD_KEYS_MAX
- PTHREAD_STACK_MIN
- PTHREAD_THREADS_MAX
-
-
-SNAPSHOT 2004-11-03
--------------------
-
-DLLs produced from this snapshot cannot be used with older applications without
-recompiling the application, due to a change to pthread_t to provide unique POSIX
-thread IDs.
-
-Although this snapshot passes the extended test suite, many of the changes are
-fairly major, and some applications may show different behaviour than previously,
-so adopt with care. Hopefully, any changed behaviour will be due to the library
-being better at it's job, not worse.
-
-Bug fixes
----------
-
-* pthread_create() no longer accepts NULL as the thread reference arg.
-A segfault (memory access fault) will result, and no thread will be
-created.
-
-* pthread_barrier_wait() no longer acts as a cancelation point.
-
-* Fix potential race condition in pthread_once()
-- Tristan Savatier <tristan at mpegtv.com>
-
-* Changes to pthread_cond_destroy() exposed some coding weaknesses in several
-test suite mini-apps because pthread_cond_destroy() now returns EBUSY if the CV
-is still in use.
-
-New features
-------------
-
-* Added for compatibility:
-PTHREAD_RECURSIVE_MUTEX_INITIALIZER,
-PTHREAD_ERRORCHECK_MUTEX_INITIALIZER,
-PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP,
-PTHREAD_ERRORCHECK_MUTEX_INITIALIZER_NP
-
-* Initial support for Digital Mars compiler
-- Anuj Goyal <anuj.goyal at gmail.com>
-
-* Faster Mutexes. These have been been rewritten following a model provided by
-Alexander Terekhov that reduces kernel space checks, and eliminates some additional
-critical sections used to manage a race between timedlock expiration and unlock.
-Please be aware that the new mutexes do not enforce strict absolute FIFO scheduling
-of mutexes, however any out-of-order lock acquisition should be very rare.
-
-* Faster semaphores. Following a similar model to mutexes above, these have been
-rewritten to use preliminary users space checks.
-
-* sem_getvalue() now returns the number of waiters.
-
-* The POSIX thread ID now has much stronger uniqueness characteristics. The library
-garrantees not to reuse the same thread ID for at least 2^(wordsize) thread
-destruction/creation cycles.
-
-New tests
----------
-
-* semaphore4.c: Tests cancelation of the new sem_wait().
-
-* semaphore4t.c: Likewise for sem_timedwait().
-
-* rwlock8.c: Tests and times the slow execution paths of r/w locks, and the CVs,
-mutexes, and semaphores that they're built on.
-
-
-SNAPSHOT 2004-05-16
--------------------
-
-Attempt to add Watcom to the list of compilers that can build the library.
-This failed in the end due to it's non-thread-aware errno. The library
-builds but the test suite fails. See README.Watcom for more details.
-
-Bug fixes
----------
-* Bug and memory leak in sem_init()
-- Alex Blanco <Alex.Blanco at motorola.com>
-
-* ptw32_getprocessors() now returns CPU count of 1 for WinCE.
-- James Ewing <james.ewing at sveasoft.com>
-
-* pthread_cond_wait() could be canceled at a point where it should not
-be cancelable. Fixed.
-- Alexander Terekhov <TEREKHOV at de.ibm.com>
-
-* sem_timedwait() had an incorrect timeout calculation.
-- Philippe Di Cristo <philipped at voicebox.com>
-
-* Fix a memory leak left behind after threads are destroyed.
-- P. van Bruggen <pietvb at newbridges.nl>
-
-New features
-------------
-* Ported to AMD64.
-- Makoto Kato <raven at oldskool.jp>
-
-* True pre-emptive asynchronous cancelation of threads. This is optional
-and requires that Panagiotis E. Hadjidoukas's QueueUserAPCEx package be
-installed. This package is included in the pthreads-win32 self-unpacking
-Zip archive starting from this snapshot. See the README.txt file inside
-the package for installation details.
-
-Note: If you don't use async cancelation in your application, or don't need
-to cancel threads that are blocked on system resources such as network I/O,
-then the default non-preemptive async cancelation is probably good enough.
-However, pthreads-win32 auto-detects the availability of these components
-at run-time, so you don't need to rebuild the library from source if you
-change your mind later.
-
-All of the advice available in books and elsewhere on the undesirability
-of using async cancelation in any application still stands, but this
-feature is a welcome addition with respect to the library's conformance to
-the POSIX standard.
-
-SNAPSHOT 2003-09-18
--------------------
-
-Cleanup of thread priority management. In particular, setting of thread
-priority now attempts to map invalid Win32 values within the range returned
-by sched_get_priority_min/max() to useful values. See README.NONPORTABLE
-under "Thread priority".
-
-Bug fixes
----------
-* pthread_getschedparam() now returns the priority given by the most recent
-call to pthread_setschedparam() or established by pthread_create(), as
-required by the standard. Previously, pthread_getschedparam() incorrectly
-returned the running thread priority at the time of the call, which may have
-been adjusted or temporarily promoted/demoted.
-
-* sched_get_priority_min() and sched_get_priority_max() now return -1 on error
-and set errno. Previously, they incorrectly returned the error value directly.
-
-
-SNAPSHOT 2003-09-04
--------------------
-
-Bug fixes
----------
-* ptw32_cancelableWait() now allows cancelation of waiting implicit POSIX
-threads.
-
-New test
---------
-* cancel8.c tests cancelation of Win32 threads waiting at a POSIX cancelation
-point.
-
-
-SNAPSHOT 2003-09-03
--------------------
-
-Bug fixes
----------
-* pthread_self() would free the newly created implicit POSIX thread handle if
-DuplicateHandle failed instead of recycle it (very unlikely).
-
-* pthread_exit() was neither freeing nor recycling the POSIX thread struct
-for implicit POSIX threads.
-
-New feature - Cancelation of/by Win32 (non-POSIX) threads
----------------------------------------------------------
-Since John Bossom's original implementation, the library has allowed non-POSIX
-initialised threads (Win32 threads) to call pthreads-win32 routines and
-therefore interact with POSIX threads. This is done by creating an on-the-fly
-POSIX thread ID for the Win32 thread that, once created, allows fully
-reciprical interaction. This did not extend to thread cancelation (async or
-deferred). Now it does.
-
-Any thread can be canceled by any other thread (Win32 or POSIX) if the former
-thread's POSIX pthread_t value is known. It's TSD destructors and POSIX
-cleanup handlers will be run before the thread exits with an exit code of
-PTHREAD_CANCELED (retrieved with GetExitCodeThread()).
-
-This allows a Win32 thread to, for example, call POSIX CV routines in the same way
-that POSIX threads would/should, with pthread_cond_wait() cancelability and
-cleanup handlers (pthread_cond_wait() is a POSIX cancelation point).
-
-By adding cancelation, Win32 threads should now be able to call all POSIX
-threads routines that make sense including semaphores, mutexes, condition
-variables, read/write locks, barriers, spinlocks, tsd, cleanup push/pop,
-cancelation, pthread_exit, scheduling, etc.
-
-Note that these on-the-fly 'implicit' POSIX thread IDs are initialised as detached
-(not joinable) with deferred cancelation type. The POSIX thread ID will be created
-automatically by any POSIX routines that need a POSIX handle (unless the routine
-needs a pthread_t as a parameter of course). A Win32 thread can discover it's own
-POSIX thread ID by calling pthread_self(), which will create the handle if
-necessary and return the pthread_t value.
-
-New tests
----------
-Test the above new feature.
-
-
-SNAPSHOT 2003-08-19
--------------------
-
-This snapshot fixes some accidental corruption to new test case sources.
-There are no changes to the library source code.
-
-
-SNAPSHOT 2003-08-15
--------------------
-
-Bug fixes
----------
-
-* pthread.dsp now uses correct compile flags (/MD).
-- Viv <vcotirlea@hotmail.com>
-
-* pthread_win32_process_detach_np() fixed memory leak.
-- Steven Reddie <Steven.Reddie@ca.com>
-
-* pthread_mutex_destroy() fixed incorrect return code.
-- Nicolas Barry <boozai@yahoo.com>
-
-* pthread_spin_destroy() fixed memory leak.
-- Piet van Bruggen <pietvb@newbridges.nl>
-
-* Various changes to tighten arg checking, and to work with later versions of
-MinGW32 and MsysDTK.
-
-* pthread_getschedparam() etc, fixed dangerous thread validity checking.
-- Nicolas Barry <boozai@yahoo.com>
-
-* POSIX thread handles are now reused and their memory is not freed on thread exit.
-This allows for stronger thread validity checking.
-
-New standard routine
---------------------
-
-* pthread_kill() added to provide thread validity checking to applications.
-It does not accept any non zero values for the signal arg.
-
-New test cases
---------------
-
-* New test cases to confirm validity checking, pthread_kill(), and thread reuse.
-
-
-SNAPSHOT 2003-05-10
--------------------
-
-Bug fixes
----------
-
-* pthread_mutex_trylock() now returns correct error values.
-pthread_mutex_destroy() will no longer destroy a recursively locked mutex.
-pthread_mutex_lock() is no longer inadvertantly behaving as a cancelation point.
-- Thomas Pfaff <tpfaff@gmx.net>
-
-* pthread_mutex_timedlock() no longer occasionally sets incorrect mutex
-ownership, causing deadlocks in some applications.
-- Robert Strycek <strycek@posam.sk> and Alexander Terekhov <TEREKHOV@de.ibm.com>
-
-
-SNAPSHOT 2002-11-04
--------------------
-
-Bug fixes
----------
-
-* sem_getvalue() now returns the correct value under Win NT and WinCE.
-- Rob Fanner <rfanner@stonethree.com>
-
-* sem_timedwait() now uses tighter checks for unreasonable
-abstime values - that would result in unexpected timeout values.
-
-* ptw32_cond_wait_cleanup() no longer mysteriously consumes
-CV signals but may produce more spurious wakeups. It is believed
-that the sem_timedwait() call is consuming a CV signal that it
-shouldn't.
-- Alexander Terekhov <TEREKHOV@de.ibm.com>
-
-* Fixed a memory leak in ptw32_threadDestroy() for implicit threads.
-
-* Fixed potential for deadlock in pthread_cond_destroy().
-A deadlock could occur for statically declared CVs (PTHREAD_COND_INITIALIZER),
-when one thread is attempting to destroy the condition variable while another
-is attempting to dynamically initialize it.
-- Michael Johnson <michaelj@maine.rr.com>
-
-
-SNAPSHOT 2002-03-02
--------------------
-
-Cleanup code default style. (IMPORTANT)
-----------------------------------------------------------------------
-Previously, if not defined, the cleanup style was determined automatically
-from the compiler/language, and one of the following was defined accordingly:
-
- __CLEANUP_SEH MSVC only
- __CLEANUP_CXX C++, including MSVC++, GNU G++
- __CLEANUP_C C, including GNU GCC, not MSVC
-
-These defines determine the style of cleanup (see pthread.h) and,
-most importantly, the way that cancelation and thread exit (via
-pthread_exit) is performed (see the routine ptw32_throw() in private.c).
-
-In short, the exceptions versions of the library throw an exception
-when a thread is canceled or exits (via pthread_exit()), which is
-caught by a handler in the thread startup routine, so that the
-the correct stack unwinding occurs regardless of where the thread
-is when it's canceled or exits via pthread_exit().
-
-In this and future snapshots, unless the build explicitly defines (e.g.
-via a compiler option) __CLEANUP_SEH, __CLEANUP_CXX, or __CLEANUP_C, then
-the build NOW always defaults to __CLEANUP_C style cleanup. This style
-uses setjmp/longjmp in the cancelation and pthread_exit implementations,
-and therefore won't do stack unwinding even when linked to applications
-that have it (e.g. C++ apps). This is for consistency with most
-current commercial Unix POSIX threads implementations. Compaq's TRU64
-may be an exception (no pun intended) and possible future trend.
-
-Although it was not clearly documented before, it is still necessary to
-build your application using the same __CLEANUP_* define as was
-used for the version of the library that you link with, so that the
-correct parts of pthread.h are included. That is, the possible
-defines require the following library versions:
-
- __CLEANUP_SEH pthreadVSE.dll
- __CLEANUP_CXX pthreadVCE.dll or pthreadGCE.dll
- __CLEANUP_C pthreadVC.dll or pthreadGC.dll
-
-E.g. regardless of whether your app is C or C++, if you link with
-pthreadVC.lib or libpthreadGC.a, then you must define __CLEANUP_C.
-
-
-THE POINT OF ALL THIS IS: if you have not been defining one of these
-explicitly, then the defaults as described at the top of this
-section were being used.
-
-THIS NOW CHANGES, as has been explained above, but to try to make this
-clearer here's an example:
-
-If you were building your application with MSVC++ i.e. using C++
-exceptions and not explicitly defining one of __CLEANUP_*, then
-__CLEANUP_C++ was automatically defined for you in pthread.h.
-You should have been linking with pthreadVCE.dll, which does
-stack unwinding.
-
-If you now build your application as you had before, pthread.h will now
-automatically set __CLEANUP_C as the default style, and you will need to
-link with pthreadVC.dll. Stack unwinding will now NOT occur when a thread
-is canceled, or the thread calls pthread_exit().
-
-Your application will now most likely behave differently to previous
-versions, and in non-obvious ways. Most likely is that locally
-instantiated objects may not be destroyed or cleaned up after a thread
-is canceled.
-
-If you want the same behaviour as before, then you must now define
-__CLEANUP_C++ explicitly using a compiler option and link with
-pthreadVCE.dll as you did before.
-
-
-WHY ARE WE MAKING THE DEFAULT STYLE LESS EXCEPTION-FRIENDLY?
-Because no commercial Unix POSIX threads implementation allows you to
-choose to have stack unwinding. Therefore, providing it in pthread-win32
-as a default is dangerous. We still provide the choice but unless
-you consciously choose to do otherwise, your pthreads applications will
-now run or crash in similar ways irrespective of the threads platform
-you use. Or at least this is the hope.
-
-
-WHY NOT REMOVE THE EXCEPTIONS VERSIONS OF THE LIBRARY ALTOGETHER?
-There are a few reasons:
-- because there are well respected POSIX threads people who believe
- that POSIX threads implementations should be exceptions aware and
- do the expected thing in that context. (There are equally respected
- people who believe it should not be easily accessible, if it's there
- at all, for unconditional conformity to other implementations.)
-- because pthreads-win32 is one of the few implementations that has
- the choice, perhaps the only freely available one, and so offers
- a laboratory to people who may want to explore the effects;
-- although the code will always be around somewhere for anyone who
- wants it, once it's removed from the current version it will not be
- nearly as visible to people who may have a use for it.
-
-
-Source module splitting
------------------------
-In order to enable smaller image sizes to be generated
-for applications that link statically with the library,
-most routines have been separated out into individual
-source code files.
-
-This is being done in such a way as to be backward compatible.
-The old source files are reused to congregate the individual
-routine files into larger translation units (via a bunch of
-# includes) so that the compiler can still optimise wherever
-possible, e.g. through inlining, which can only be done
-within the same translation unit.
-
-It is also possible to build the entire library by compiling
-the single file named "pthread.c", which just #includes all
-the secondary congregation source files. The compiler
-may be able to use this to do more inlining of routines.
-
-Although the GNU compiler is able to produce libraries with
-the necessary separation (the -ffunction-segments switch),
-AFAIK, the MSVC and other compilers don't have this feature.
-
-Finally, since I use makefiles and command-line compilation,
-I don't know what havoc this reorganisation may wreak amongst
-IDE project file users. You should be able to continue
-using your existing project files without modification.
-
-
-New non-portable functions
---------------------------
-pthread_num_processors_np():
- Returns the number of processors in the system that are
- available to the process, as determined from the processor
- affinity mask.
-
-pthread_timechange_handler_np():
- To improve tolerance against operator or time service initiated
- system clock changes.
-
- This routine can be called by an application when it
- receives a WM_TIMECHANGE message from the system. At present
- it broadcasts all condition variables so that waiting threads
- can wake up and re-evaluate their conditions and restart
- their timed waits if required.
- - Suggested by Alexander Terekhov
-
-
-Platform dependence
--------------------
-As Win95 doesn't provide one, the library now contains
-it's own InterlockedCompareExchange() routine, which is used
-whenever Windows doesn't provide it. InterlockedCompareExchange()
-is used to implement spinlocks and barriers, and also in mutexes.
-This routine relies on the CMPXCHG machine instruction which
-is not available on i386 CPUs. This library (from snapshot
-20010712 onwards) is therefore no longer supported on i386
-processor platforms.
-
-
-New standard routines
----------------------
-For source code portability only - rwlocks cannot be process shared yet.
-
- pthread_rwlockattr_init()
- pthread_rwlockattr_destroy()
- pthread_rwlockattr_setpshared()
- pthread_rwlockattr_getpshared()
-
-As defined in the new POSIX standard, and the Single Unix Spec version 3:
-
- sem_timedwait()
- pthread_mutex_timedlock() - Alexander Terekhov and Thomas Pfaff
- pthread_rwlock_timedrdlock() - adapted from pthread_rwlock_rdlock()
- pthread_rwlock_timedwrlock() - adapted from pthread_rwlock_wrlock()
-
-
-pthread.h no longer includes windows.h
---------------------------------------
-[Not yet for G++]
-
-This was done to prevent conflicts.
-
-HANDLE, DWORD, and NULL are temporarily defined within pthread.h if
-they are not already.
-
-
-pthread.h, sched.h and semaphore.h now use dllexport/dllimport
---------------------------------------------------------------
-Not only to avoid the need for the pthread.def file, but to
-improve performance. Apparently, declaring functions with dllimport
-generates a direct call to the function and avoids the overhead
-of a stub function call.
-
-Bug fixes
----------
-* Fixed potential NULL pointer dereferences in pthread_mutexattr_init,
-pthread_mutexattr_getpshared, pthread_barrierattr_init,
-pthread_barrierattr_getpshared, and pthread_condattr_getpshared.
-- Scott McCaskill <scott@magruder.org>
-
-* Removed potential race condition in pthread_mutex_trylock and
-pthread_mutex_lock;
-- Alexander Terekhov <TEREKHOV@de.ibm.com>
-
-* The behaviour of pthread_mutex_trylock in relation to
-recursive mutexes was inconsistent with commercial implementations.
-Trylock would return EBUSY if the lock was owned already by the
-calling thread regardless of mutex type. Trylock now increments the
-recursion count and returns 0 for RECURSIVE mutexes, and will
-return EDEADLK rather than EBUSY for ERRORCHECK mutexes. This is
-consistent with Solaris.
-- Thomas Pfaff <tpfaff@gmx.net>
-
-* Found a fix for the library and workaround for applications for
-the known bug #2, i.e. where __CLEANUP_CXX or __CLEANUP_SEH is defined.
-See the "Known Bugs in this snapshot" section below.
-
-This could be made transparent to applications by replacing the macros that
-define the current C++ and SEH versions of pthread_cleanup_push/pop
-with the C version, but AFAIK cleanup handlers would not then run in the
-correct sequence with destructors and exception cleanup handlers when
-an exception occurs.
-
-* Cancelation once started in a thread cannot now be inadvertantly
-double canceled. That is, once a thread begins it's cancelation run,
-cancelation is disabled and a subsequent cancel request will
-return an error (ESRCH).
-
-* errno: An incorrect compiler directive caused a local version
-of errno to be used instead of the Win32 errno. Both instances are
-thread-safe but applications checking errno after a pthreads-win32
-call would be wrong. Fixing this also fixed a bad compiler
-option in the testsuite (/MT should have been /MD) which is
-needed to link with the correct library MSVCRT.LIB.
-
-
-SNAPSHOT 2001-07-12
--------------------
-
-To be added
-
-
-SNAPSHOT 2001-07-03
--------------------
-
-To be added
-
-
-SNAPSHOT 2000-08-13
--------------------
-
-New:
-- Renamed DLL and LIB files:
- pthreadVSE.dll (MS VC++/Structured EH)
- pthreadVSE.lib
- pthreadVCE.dll (MS VC++/C++ EH)
- pthreadVCE.lib
- pthreadGCE.dll (GNU G++/C++ EH)
- libpthreadw32.a
-
- Both your application and the pthread dll should use the
- same exception handling scheme.
-
-Bugs fixed:
-- MSVC++ C++ exception handling.
-
-Some new tests have been added.
-
-
-SNAPSHOT 2000-08-10
--------------------
-
-New:
-- asynchronous cancelation on X86 (Jason Nye)
-- Makefile compatible with MS nmake to replace
- buildlib.bat
-- GNUmakefile for Mingw32
-- tests/Makefile for MS nmake replaces runall.bat
-- tests/GNUmakefile for Mingw32
-
-Bugs fixed:
-- kernel32 load/free problem
-- attempt to hide internel exceptions from application
- exception handlers (__try/__except and try/catch blocks)
-- Win32 thread handle leakage bug
- (David Baggett/Paul Redondo/Eyal Lebedinsky)
-
-Some new tests have been added.
-
-
-SNAPSHOT 1999-11-02
--------------------
-
-Bugs fixed:
-- ctime_r macro had an incorrect argument (Erik Hensema),
-- threads were not being created
- PTHREAD_CANCEL_DEFERRED. This should have
- had little effect as deferred is the only
- supported type. (Ross Johnson).
-
-Some compatibility improvements added, eg.
-- pthread_setcancelstate accepts NULL pointer
- for the previous value argument. Ditto for
- pthread_setcanceltype. This is compatible
- with Solaris but should not affect
- standard applications (Erik Hensema)
-
-Some new tests have been added.
-
-
-SNAPSHOT 1999-10-17
--------------------
-
-Bug fix - Cancelation of threads waiting on condition variables
-now works properly (Lorin Hochstein and Peter Slacik)
-
-
-SNAPSHOT 1999-08-12
--------------------
-
-Fixed exception stack cleanup if calling pthread_exit()
-- (Lorin Hochstein and John Bossom).
-
-Fixed bugs in condition variables - (Peter Slacik):
- - additional contention checks
- - properly adjust number of waiting threads after timed
- condvar timeout.
-
-
-SNAPSHOT 1999-05-30
--------------------
-
-Some minor bugs have been fixed. See the ChangeLog file for details.
-
-Some more POSIX 1b functions are now included but ony return an
-error (ENOSYS) if called. They are:
-
- sem_open
- sem_close
- sem_unlink
- sem_getvalue
-
-
-SNAPSHOT 1999-04-07
--------------------
-
-Some POSIX 1b functions which were internally supported are now
-available as exported functions:
-
- sem_init
- sem_destroy
- sem_wait
- sem_trywait
- sem_post
- sched_yield
- sched_get_priority_min
- sched_get_priority_max
-
-Some minor bugs have been fixed. See the ChangeLog file for details.
-
-
-SNAPSHOT 1999-03-16
--------------------
-
-Initial release.
-
+RELEASE 1.7.0
+-------------
+(2005-04-12)
+
+General
+-------
+
+This release is a backport of features and bug fixes new in
+release 2.3.0. See earlier notes under Release 1.4.0/General.
+
+Bugs fixed
+----------
+
+* Fixed pthread_once potential for post once_routine cancellation
+hanging due to starvation. See comments in pthread_once.c.
+Momentary priority boosting is used to ensure that, after a
+once_routine is cancelled, the thread that will run the
+once_routine is not starved by higher priority waiting threads at
+critical times. Priority boosting occurs only AFTER a once_routine
+cancellation, and is applied only to that once_control. The
+once_routine is run at the thread's normal base priority.
+
+New tests
+---------
+
+* once4.c: Aggressively tests pthread_once() under realtime
+conditions using threads with varying priorities. Windows'
+random priority boosting does not occur for threads with realtime
+priority levels.
+
+
+RELEASE 1.6.0
+-------------
+(2005-04-04)
+
+General
+-------
+
+This release is a backport of features and bug fixes new in
+release 2.2.0. See earlier notes under Release 1.4.0/General.
+
+* Added makefile targets to build static link versions of the library.
+Both MinGW and MSVC. Please note that this does not imply any change
+to the LGPL licensing, which still imposes specific conditions on
+distributing software that has been statically linked with this library.
+
+Bugs fixed
+----------
+
+* Fix integer overrun error in sem_timedwait().
+Kevin Lussier
+
+* Fix preprocessor directives for static linking.
+Dimitar Panayotov
+
+
+RELEASE 1.5.0
+-------------
+(2005-03-16)
+
+General
+-------
+
+This release is a backport of features and bug fixes new in
+release 2.1.0. See earlier notes under Release 1.4.0/General.
+
+Bugs fixed
+----------
+
+* Reverse change to pthread_setcancelstate() made in 1.4.0.
+
+
+RELEASE 1.4.0
+-------------
+(2005-03-16)
+
+General
+-------
+
+Version 2.0.0 is being released and version 1.4.0 back-ports the
+functionality included in that release. In future, bug fixes
+will be back-ported if possible and where appropriate, however,
+this will be at the maintainer's descretion. New features will
+not generally be back-ported, although some may if they
+provide major improvements in compatibility with the standard.
+
+This is intended as a courtesy to application developers, but as
+time moves on, the imperative will fade.
+
+The package naming has changed, replacing the snapshot date with
+the version number + descriptive information. E.g. this
+release is "pthreads-w32-1-4-0-release".
+
+Bugs fixed
+----------
+
+* pthread_setcancelstate() no longer checks for a pending
+async cancel event if the library is using alertable async
+cancel. See the README file (Prerequisites section) for info
+on adding alertable async cancelation.
+
+New features
+------------
+
+* pthread_once() now supports init_routine cancellability.
+
+New tests
+---------
+
+* Agressively test pthread_once() init_routine cancellability.
+
+
+SNAPSHOT 2005-03-08
+-------------------
+Version 1.3.0
+
+Bug reports (fixed)
+-------------------
+
+* Implicitly created threads leave Win32 handles behind after exiting.
+- Dmitrii Semii
+
+* pthread_once() starvation problem.
+- Gottlob Frege
+
+New tests
+---------
+
+* More intense testing of pthread_once().
+
+
+SNAPSHOT 2005-01-25
+-------------------
+Version 1.2.0
+
+Bug fixes
+---------
+
+* Attempted acquisition of a recursive mutex could cause waiting threads
+to not be woken when the mutex was released.
+- Ralf Kubis <RKubis at mc.com>
+
+* Various package omissions have been fixed.
+
+
+SNAPSHOT 2005-01-03
+-------------------
+Version 1.1.0
+
+Bug fixes
+---------
+
+* Unlocking recursive or errorcheck mutexes would sometimes
+unexpectedly return an EPERM error (bug introduced in
+snapshot-2004-11-03).
+- Konstantin Voronkov <beowinkle at yahoo.com>
+
+
+SNAPSHOT 2004-11-22
+-------------------
+Version 1.0.0
+
+This snapshot primarily fixes the condvar bug introduced in
+snapshot-2004-11-03. DLL versioning has also been included to allow
+applications to runtime check the Microsoft compatible DLL version
+information, and to extend the DLL naming system for ABI and major
+(non-backward compatible) API changes. See the README file for details.
+
+Bug fixes
+---------
+
+* Condition variables no longer deadlock (bug introduced in
+snapshot-2004-11-03).
+- Alexander Kotliarov and Nicolas at saintmac
+
+* DLL naming extended to avoid 'DLL hell' in the future, and to
+accommodate the ABI change introduced in snapshot-2004-11-03. Snapshot
+2004-11-03 will be removed from FTP sites.
+
+New features
+------------
+
+* A Microsoft-style version resource has been added to the DLL for
+applications that wish to check DLL compatibility at runtime.
+
+* Pthreads-win32 DLL naming has been extended to allow incompatible DLL
+versions to co-exist in the same filesystem. See the README file for details,
+but briefly: while the version information inside the DLL will change with
+each release from now on, the DLL version names will only change if the new
+DLL is not backward compatible with older applications.
+
+The versioning scheme has been borrowed from GNU Libtool, and the DLL
+naming scheme is from Cygwin. Provided the Libtool-style numbering rules are
+honoured, the Cygwin DLL naming scheme automatcally ensures that DLL name
+changes are minimal and that applications will not load an incompatible
+pthreads-win32 DLL.
+
+Those who use the pre-built DLLs will find that the DLL/LIB names have a new
+suffix (1) in this snapshot. E.g. pthreadVC1.dll etc.
+
+* The POSIX thread ID reuse uniqueness feature introduced in the last snapshot
+has been kept as default, but the behaviour can now be controlled when the DLL
+is built to effectively switch it off. This makes the library much more
+sensitive to applications that assume that POSIX thread IDs are unique, i.e.
+are not strictly compliant with POSIX. See the PTW32_THREAD_ID_REUSE_INCREMENT
+macro comments in config.h for details.
+
+Other changes
+-------------
+Certain POSIX macros have changed.
+
+These changes are intended to conform to the Single Unix Specification version 3,
+which states that, if set to 0 (zero) or not defined, then applications may use
+sysconf() to determine their values at runtime. Pthreads-win32 does not
+implement sysconf().
+
+The following macros are no longer undefined, but defined and set to -1
+(not implemented):
+
+ _POSIX_THREAD_ATTR_STACKADDR
+ _POSIX_THREAD_PRIO_INHERIT
+ _POSIX_THREAD_PRIO_PROTECT
+ _POSIX_THREAD_PROCESS_SHARED
+
+The following macros are defined and set to 200112L (implemented):
+
+ _POSIX_THREADS
+ _POSIX_THREAD_SAFE_FUNCTIONS
+ _POSIX_THREAD_ATTR_STACKSIZE
+ _POSIX_THREAD_PRIORITY_SCHEDULING
+ _POSIX_SEMAPHORES
+ _POSIX_READER_WRITER_LOCKS
+ _POSIX_SPIN_LOCKS
+ _POSIX_BARRIERS
+
+The following macros are defined and set to appropriate values:
+
+ _POSIX_THREAD_THREADS_MAX
+ _POSIX_SEM_VALUE_MAX
+ _POSIX_SEM_NSEMS_MAX
+ PTHREAD_DESTRUCTOR_ITERATIONS
+ PTHREAD_KEYS_MAX
+ PTHREAD_STACK_MIN
+ PTHREAD_THREADS_MAX
+
+
+SNAPSHOT 2004-11-03
+-------------------
+
+DLLs produced from this snapshot cannot be used with older applications without
+recompiling the application, due to a change to pthread_t to provide unique POSIX
+thread IDs.
+
+Although this snapshot passes the extended test suite, many of the changes are
+fairly major, and some applications may show different behaviour than previously,
+so adopt with care. Hopefully, any changed behaviour will be due to the library
+being better at it's job, not worse.
+
+Bug fixes
+---------
+
+* pthread_create() no longer accepts NULL as the thread reference arg.
+A segfault (memory access fault) will result, and no thread will be
+created.
+
+* pthread_barrier_wait() no longer acts as a cancelation point.
+
+* Fix potential race condition in pthread_once()
+- Tristan Savatier <tristan at mpegtv.com>
+
+* Changes to pthread_cond_destroy() exposed some coding weaknesses in several
+test suite mini-apps because pthread_cond_destroy() now returns EBUSY if the CV
+is still in use.
+
+New features
+------------
+
+* Added for compatibility:
+PTHREAD_RECURSIVE_MUTEX_INITIALIZER,
+PTHREAD_ERRORCHECK_MUTEX_INITIALIZER,
+PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP,
+PTHREAD_ERRORCHECK_MUTEX_INITIALIZER_NP
+
+* Initial support for Digital Mars compiler
+- Anuj Goyal <anuj.goyal at gmail.com>
+
+* Faster Mutexes. These have been been rewritten following a model provided by
+Alexander Terekhov that reduces kernel space checks, and eliminates some additional
+critical sections used to manage a race between timedlock expiration and unlock.
+Please be aware that the new mutexes do not enforce strict absolute FIFO scheduling
+of mutexes, however any out-of-order lock acquisition should be very rare.
+
+* Faster semaphores. Following a similar model to mutexes above, these have been
+rewritten to use preliminary users space checks.
+
+* sem_getvalue() now returns the number of waiters.
+
+* The POSIX thread ID now has much stronger uniqueness characteristics. The library
+garrantees not to reuse the same thread ID for at least 2^(wordsize) thread
+destruction/creation cycles.
+
+New tests
+---------
+
+* semaphore4.c: Tests cancelation of the new sem_wait().
+
+* semaphore4t.c: Likewise for sem_timedwait().
+
+* rwlock8.c: Tests and times the slow execution paths of r/w locks, and the CVs,
+mutexes, and semaphores that they're built on.
+
+
+SNAPSHOT 2004-05-16
+-------------------
+
+Attempt to add Watcom to the list of compilers that can build the library.
+This failed in the end due to it's non-thread-aware errno. The library
+builds but the test suite fails. See README.Watcom for more details.
+
+Bug fixes
+---------
+* Bug and memory leak in sem_init()
+- Alex Blanco <Alex.Blanco at motorola.com>
+
+* ptw32_getprocessors() now returns CPU count of 1 for WinCE.
+- James Ewing <james.ewing at sveasoft.com>
+
+* pthread_cond_wait() could be canceled at a point where it should not
+be cancelable. Fixed.
+- Alexander Terekhov <TEREKHOV at de.ibm.com>
+
+* sem_timedwait() had an incorrect timeout calculation.
+- Philippe Di Cristo <philipped at voicebox.com>
+
+* Fix a memory leak left behind after threads are destroyed.
+- P. van Bruggen <pietvb at newbridges.nl>
+
+New features
+------------
+* Ported to AMD64.
+- Makoto Kato <raven at oldskool.jp>
+
+* True pre-emptive asynchronous cancelation of threads. This is optional
+and requires that Panagiotis E. Hadjidoukas's QueueUserAPCEx package be
+installed. This package is included in the pthreads-win32 self-unpacking
+Zip archive starting from this snapshot. See the README.txt file inside
+the package for installation details.
+
+Note: If you don't use async cancelation in your application, or don't need
+to cancel threads that are blocked on system resources such as network I/O,
+then the default non-preemptive async cancelation is probably good enough.
+However, pthreads-win32 auto-detects the availability of these components
+at run-time, so you don't need to rebuild the library from source if you
+change your mind later.
+
+All of the advice available in books and elsewhere on the undesirability
+of using async cancelation in any application still stands, but this
+feature is a welcome addition with respect to the library's conformance to
+the POSIX standard.
+
+SNAPSHOT 2003-09-18
+-------------------
+
+Cleanup of thread priority management. In particular, setting of thread
+priority now attempts to map invalid Win32 values within the range returned
+by sched_get_priority_min/max() to useful values. See README.NONPORTABLE
+under "Thread priority".
+
+Bug fixes
+---------
+* pthread_getschedparam() now returns the priority given by the most recent
+call to pthread_setschedparam() or established by pthread_create(), as
+required by the standard. Previously, pthread_getschedparam() incorrectly
+returned the running thread priority at the time of the call, which may have
+been adjusted or temporarily promoted/demoted.
+
+* sched_get_priority_min() and sched_get_priority_max() now return -1 on error
+and set errno. Previously, they incorrectly returned the error value directly.
+
+
+SNAPSHOT 2003-09-04
+-------------------
+
+Bug fixes
+---------
+* ptw32_cancelableWait() now allows cancelation of waiting implicit POSIX
+threads.
+
+New test
+--------
+* cancel8.c tests cancelation of Win32 threads waiting at a POSIX cancelation
+point.
+
+
+SNAPSHOT 2003-09-03
+-------------------
+
+Bug fixes
+---------
+* pthread_self() would free the newly created implicit POSIX thread handle if
+DuplicateHandle failed instead of recycle it (very unlikely).
+
+* pthread_exit() was neither freeing nor recycling the POSIX thread struct
+for implicit POSIX threads.
+
+New feature - Cancelation of/by Win32 (non-POSIX) threads
+---------------------------------------------------------
+Since John Bossom's original implementation, the library has allowed non-POSIX
+initialised threads (Win32 threads) to call pthreads-win32 routines and
+therefore interact with POSIX threads. This is done by creating an on-the-fly
+POSIX thread ID for the Win32 thread that, once created, allows fully
+reciprical interaction. This did not extend to thread cancelation (async or
+deferred). Now it does.
+
+Any thread can be canceled by any other thread (Win32 or POSIX) if the former
+thread's POSIX pthread_t value is known. It's TSD destructors and POSIX
+cleanup handlers will be run before the thread exits with an exit code of
+PTHREAD_CANCELED (retrieved with GetExitCodeThread()).
+
+This allows a Win32 thread to, for example, call POSIX CV routines in the same way
+that POSIX threads would/should, with pthread_cond_wait() cancelability and
+cleanup handlers (pthread_cond_wait() is a POSIX cancelation point).
+
+By adding cancelation, Win32 threads should now be able to call all POSIX
+threads routines that make sense including semaphores, mutexes, condition
+variables, read/write locks, barriers, spinlocks, tsd, cleanup push/pop,
+cancelation, pthread_exit, scheduling, etc.
+
+Note that these on-the-fly 'implicit' POSIX thread IDs are initialised as detached
+(not joinable) with deferred cancelation type. The POSIX thread ID will be created
+automatically by any POSIX routines that need a POSIX handle (unless the routine
+needs a pthread_t as a parameter of course). A Win32 thread can discover it's own
+POSIX thread ID by calling pthread_self(), which will create the handle if
+necessary and return the pthread_t value.
+
+New tests
+---------
+Test the above new feature.
+
+
+SNAPSHOT 2003-08-19
+-------------------
+
+This snapshot fixes some accidental corruption to new test case sources.
+There are no changes to the library source code.
+
+
+SNAPSHOT 2003-08-15
+-------------------
+
+Bug fixes
+---------
+
+* pthread.dsp now uses correct compile flags (/MD).
+- Viv <vcotirlea@hotmail.com>
+
+* pthread_win32_process_detach_np() fixed memory leak.
+- Steven Reddie <Steven.Reddie@ca.com>
+
+* pthread_mutex_destroy() fixed incorrect return code.
+- Nicolas Barry <boozai@yahoo.com>
+
+* pthread_spin_destroy() fixed memory leak.
+- Piet van Bruggen <pietvb@newbridges.nl>
+
+* Various changes to tighten arg checking, and to work with later versions of
+MinGW32 and MsysDTK.
+
+* pthread_getschedparam() etc, fixed dangerous thread validity checking.
+- Nicolas Barry <boozai@yahoo.com>
+
+* POSIX thread handles are now reused and their memory is not freed on thread exit.
+This allows for stronger thread validity checking.
+
+New standard routine
+--------------------
+
+* pthread_kill() added to provide thread validity checking to applications.
+It does not accept any non zero values for the signal arg.
+
+New test cases
+--------------
+
+* New test cases to confirm validity checking, pthread_kill(), and thread reuse.
+
+
+SNAPSHOT 2003-05-10
+-------------------
+
+Bug fixes
+---------
+
+* pthread_mutex_trylock() now returns correct error values.
+pthread_mutex_destroy() will no longer destroy a recursively locked mutex.
+pthread_mutex_lock() is no longer inadvertantly behaving as a cancelation point.
+- Thomas Pfaff <tpfaff@gmx.net>
+
+* pthread_mutex_timedlock() no longer occasionally sets incorrect mutex
+ownership, causing deadlocks in some applications.
+- Robert Strycek <strycek@posam.sk> and Alexander Terekhov <TEREKHOV@de.ibm.com>
+
+
+SNAPSHOT 2002-11-04
+-------------------
+
+Bug fixes
+---------
+
+* sem_getvalue() now returns the correct value under Win NT and WinCE.
+- Rob Fanner <rfanner@stonethree.com>
+
+* sem_timedwait() now uses tighter checks for unreasonable
+abstime values - that would result in unexpected timeout values.
+
+* ptw32_cond_wait_cleanup() no longer mysteriously consumes
+CV signals but may produce more spurious wakeups. It is believed
+that the sem_timedwait() call is consuming a CV signal that it
+shouldn't.
+- Alexander Terekhov <TEREKHOV@de.ibm.com>
+
+* Fixed a memory leak in ptw32_threadDestroy() for implicit threads.
+
+* Fixed potential for deadlock in pthread_cond_destroy().
+A deadlock could occur for statically declared CVs (PTHREAD_COND_INITIALIZER),
+when one thread is attempting to destroy the condition variable while another
+is attempting to dynamically initialize it.
+- Michael Johnson <michaelj@maine.rr.com>
+
+
+SNAPSHOT 2002-03-02
+-------------------
+
+Cleanup code default style. (IMPORTANT)
+----------------------------------------------------------------------
+Previously, if not defined, the cleanup style was determined automatically
+from the compiler/language, and one of the following was defined accordingly:
+
+ __CLEANUP_SEH MSVC only
+ __CLEANUP_CXX C++, including MSVC++, GNU G++
+ __CLEANUP_C C, including GNU GCC, not MSVC
+
+These defines determine the style of cleanup (see pthread.h) and,
+most importantly, the way that cancelation and thread exit (via
+pthread_exit) is performed (see the routine ptw32_throw() in private.c).
+
+In short, the exceptions versions of the library throw an exception
+when a thread is canceled or exits (via pthread_exit()), which is
+caught by a handler in the thread startup routine, so that the
+the correct stack unwinding occurs regardless of where the thread
+is when it's canceled or exits via pthread_exit().
+
+In this and future snapshots, unless the build explicitly defines (e.g.
+via a compiler option) __CLEANUP_SEH, __CLEANUP_CXX, or __CLEANUP_C, then
+the build NOW always defaults to __CLEANUP_C style cleanup. This style
+uses setjmp/longjmp in the cancelation and pthread_exit implementations,
+and therefore won't do stack unwinding even when linked to applications
+that have it (e.g. C++ apps). This is for consistency with most
+current commercial Unix POSIX threads implementations. Compaq's TRU64
+may be an exception (no pun intended) and possible future trend.
+
+Although it was not clearly documented before, it is still necessary to
+build your application using the same __CLEANUP_* define as was
+used for the version of the library that you link with, so that the
+correct parts of pthread.h are included. That is, the possible
+defines require the following library versions:
+
+ __CLEANUP_SEH pthreadVSE.dll
+ __CLEANUP_CXX pthreadVCE.dll or pthreadGCE.dll
+ __CLEANUP_C pthreadVC.dll or pthreadGC.dll
+
+E.g. regardless of whether your app is C or C++, if you link with
+pthreadVC.lib or libpthreadGC.a, then you must define __CLEANUP_C.
+
+
+THE POINT OF ALL THIS IS: if you have not been defining one of these
+explicitly, then the defaults as described at the top of this
+section were being used.
+
+THIS NOW CHANGES, as has been explained above, but to try to make this
+clearer here's an example:
+
+If you were building your application with MSVC++ i.e. using C++
+exceptions and not explicitly defining one of __CLEANUP_*, then
+__CLEANUP_C++ was automatically defined for you in pthread.h.
+You should have been linking with pthreadVCE.dll, which does
+stack unwinding.
+
+If you now build your application as you had before, pthread.h will now
+automatically set __CLEANUP_C as the default style, and you will need to
+link with pthreadVC.dll. Stack unwinding will now NOT occur when a thread
+is canceled, or the thread calls pthread_exit().
+
+Your application will now most likely behave differently to previous
+versions, and in non-obvious ways. Most likely is that locally
+instantiated objects may not be destroyed or cleaned up after a thread
+is canceled.
+
+If you want the same behaviour as before, then you must now define
+__CLEANUP_C++ explicitly using a compiler option and link with
+pthreadVCE.dll as you did before.
+
+
+WHY ARE WE MAKING THE DEFAULT STYLE LESS EXCEPTION-FRIENDLY?
+Because no commercial Unix POSIX threads implementation allows you to
+choose to have stack unwinding. Therefore, providing it in pthread-win32
+as a default is dangerous. We still provide the choice but unless
+you consciously choose to do otherwise, your pthreads applications will
+now run or crash in similar ways irrespective of the threads platform
+you use. Or at least this is the hope.
+
+
+WHY NOT REMOVE THE EXCEPTIONS VERSIONS OF THE LIBRARY ALTOGETHER?
+There are a few reasons:
+- because there are well respected POSIX threads people who believe
+ that POSIX threads implementations should be exceptions aware and
+ do the expected thing in that context. (There are equally respected
+ people who believe it should not be easily accessible, if it's there
+ at all, for unconditional conformity to other implementations.)
+- because pthreads-win32 is one of the few implementations that has
+ the choice, perhaps the only freely available one, and so offers
+ a laboratory to people who may want to explore the effects;
+- although the code will always be around somewhere for anyone who
+ wants it, once it's removed from the current version it will not be
+ nearly as visible to people who may have a use for it.
+
+
+Source module splitting
+-----------------------
+In order to enable smaller image sizes to be generated
+for applications that link statically with the library,
+most routines have been separated out into individual
+source code files.
+
+This is being done in such a way as to be backward compatible.
+The old source files are reused to congregate the individual
+routine files into larger translation units (via a bunch of
+# includes) so that the compiler can still optimise wherever
+possible, e.g. through inlining, which can only be done
+within the same translation unit.
+
+It is also possible to build the entire library by compiling
+the single file named "pthread.c", which just #includes all
+the secondary congregation source files. The compiler
+may be able to use this to do more inlining of routines.
+
+Although the GNU compiler is able to produce libraries with
+the necessary separation (the -ffunction-segments switch),
+AFAIK, the MSVC and other compilers don't have this feature.
+
+Finally, since I use makefiles and command-line compilation,
+I don't know what havoc this reorganisation may wreak amongst
+IDE project file users. You should be able to continue
+using your existing project files without modification.
+
+
+New non-portable functions
+--------------------------
+pthread_num_processors_np():
+ Returns the number of processors in the system that are
+ available to the process, as determined from the processor
+ affinity mask.
+
+pthread_timechange_handler_np():
+ To improve tolerance against operator or time service initiated
+ system clock changes.
+
+ This routine can be called by an application when it
+ receives a WM_TIMECHANGE message from the system. At present
+ it broadcasts all condition variables so that waiting threads
+ can wake up and re-evaluate their conditions and restart
+ their timed waits if required.
+ - Suggested by Alexander Terekhov
+
+
+Platform dependence
+-------------------
+As Win95 doesn't provide one, the library now contains
+it's own InterlockedCompareExchange() routine, which is used
+whenever Windows doesn't provide it. InterlockedCompareExchange()
+is used to implement spinlocks and barriers, and also in mutexes.
+This routine relies on the CMPXCHG machine instruction which
+is not available on i386 CPUs. This library (from snapshot
+20010712 onwards) is therefore no longer supported on i386
+processor platforms.
+
+
+New standard routines
+---------------------
+For source code portability only - rwlocks cannot be process shared yet.
+
+ pthread_rwlockattr_init()
+ pthread_rwlockattr_destroy()
+ pthread_rwlockattr_setpshared()
+ pthread_rwlockattr_getpshared()
+
+As defined in the new POSIX standard, and the Single Unix Spec version 3:
+
+ sem_timedwait()
+ pthread_mutex_timedlock() - Alexander Terekhov and Thomas Pfaff
+ pthread_rwlock_timedrdlock() - adapted from pthread_rwlock_rdlock()
+ pthread_rwlock_timedwrlock() - adapted from pthread_rwlock_wrlock()
+
+
+pthread.h no longer includes windows.h
+--------------------------------------
+[Not yet for G++]
+
+This was done to prevent conflicts.
+
+HANDLE, DWORD, and NULL are temporarily defined within pthread.h if
+they are not already.
+
+
+pthread.h, sched.h and semaphore.h now use dllexport/dllimport
+--------------------------------------------------------------
+Not only to avoid the need for the pthread.def file, but to
+improve performance. Apparently, declaring functions with dllimport
+generates a direct call to the function and avoids the overhead
+of a stub function call.
+
+Bug fixes
+---------
+* Fixed potential NULL pointer dereferences in pthread_mutexattr_init,
+pthread_mutexattr_getpshared, pthread_barrierattr_init,
+pthread_barrierattr_getpshared, and pthread_condattr_getpshared.
+- Scott McCaskill <scott@magruder.org>
+
+* Removed potential race condition in pthread_mutex_trylock and
+pthread_mutex_lock;
+- Alexander Terekhov <TEREKHOV@de.ibm.com>
+
+* The behaviour of pthread_mutex_trylock in relation to
+recursive mutexes was inconsistent with commercial implementations.
+Trylock would return EBUSY if the lock was owned already by the
+calling thread regardless of mutex type. Trylock now increments the
+recursion count and returns 0 for RECURSIVE mutexes, and will
+return EDEADLK rather than EBUSY for ERRORCHECK mutexes. This is
+consistent with Solaris.
+- Thomas Pfaff <tpfaff@gmx.net>
+
+* Found a fix for the library and workaround for applications for
+the known bug #2, i.e. where __CLEANUP_CXX or __CLEANUP_SEH is defined.
+See the "Known Bugs in this snapshot" section below.
+
+This could be made transparent to applications by replacing the macros that
+define the current C++ and SEH versions of pthread_cleanup_push/pop
+with the C version, but AFAIK cleanup handlers would not then run in the
+correct sequence with destructors and exception cleanup handlers when
+an exception occurs.
+
+* Cancelation once started in a thread cannot now be inadvertantly
+double canceled. That is, once a thread begins it's cancelation run,
+cancelation is disabled and a subsequent cancel request will
+return an error (ESRCH).
+
+* errno: An incorrect compiler directive caused a local version
+of errno to be used instead of the Win32 errno. Both instances are
+thread-safe but applications checking errno after a pthreads-win32
+call would be wrong. Fixing this also fixed a bad compiler
+option in the testsuite (/MT should have been /MD) which is
+needed to link with the correct library MSVCRT.LIB.
+
+
+SNAPSHOT 2001-07-12
+-------------------
+
+To be added
+
+
+SNAPSHOT 2001-07-03
+-------------------
+
+To be added
+
+
+SNAPSHOT 2000-08-13
+-------------------
+
+New:
+- Renamed DLL and LIB files:
+ pthreadVSE.dll (MS VC++/Structured EH)
+ pthreadVSE.lib
+ pthreadVCE.dll (MS VC++/C++ EH)
+ pthreadVCE.lib
+ pthreadGCE.dll (GNU G++/C++ EH)
+ libpthreadw32.a
+
+ Both your application and the pthread dll should use the
+ same exception handling scheme.
+
+Bugs fixed:
+- MSVC++ C++ exception handling.
+
+Some new tests have been added.
+
+
+SNAPSHOT 2000-08-10
+-------------------
+
+New:
+- asynchronous cancelation on X86 (Jason Nye)
+- Makefile compatible with MS nmake to replace
+ buildlib.bat
+- GNUmakefile for Mingw32
+- tests/Makefile for MS nmake replaces runall.bat
+- tests/GNUmakefile for Mingw32
+
+Bugs fixed:
+- kernel32 load/free problem
+- attempt to hide internel exceptions from application
+ exception handlers (__try/__except and try/catch blocks)
+- Win32 thread handle leakage bug
+ (David Baggett/Paul Redondo/Eyal Lebedinsky)
+
+Some new tests have been added.
+
+
+SNAPSHOT 1999-11-02
+-------------------
+
+Bugs fixed:
+- ctime_r macro had an incorrect argument (Erik Hensema),
+- threads were not being created
+ PTHREAD_CANCEL_DEFERRED. This should have
+ had little effect as deferred is the only
+ supported type. (Ross Johnson).
+
+Some compatibility improvements added, eg.
+- pthread_setcancelstate accepts NULL pointer
+ for the previous value argument. Ditto for
+ pthread_setcanceltype. This is compatible
+ with Solaris but should not affect
+ standard applications (Erik Hensema)
+
+Some new tests have been added.
+
+
+SNAPSHOT 1999-10-17
+-------------------
+
+Bug fix - Cancelation of threads waiting on condition variables
+now works properly (Lorin Hochstein and Peter Slacik)
+
+
+SNAPSHOT 1999-08-12
+-------------------
+
+Fixed exception stack cleanup if calling pthread_exit()
+- (Lorin Hochstein and John Bossom).
+
+Fixed bugs in condition variables - (Peter Slacik):
+ - additional contention checks
+ - properly adjust number of waiting threads after timed
+ condvar timeout.
+
+
+SNAPSHOT 1999-05-30
+-------------------
+
+Some minor bugs have been fixed. See the ChangeLog file for details.
+
+Some more POSIX 1b functions are now included but ony return an
+error (ENOSYS) if called. They are:
+
+ sem_open
+ sem_close
+ sem_unlink
+ sem_getvalue
+
+
+SNAPSHOT 1999-04-07
+-------------------
+
+Some POSIX 1b functions which were internally supported are now
+available as exported functions:
+
+ sem_init
+ sem_destroy
+ sem_wait
+ sem_trywait
+ sem_post
+ sched_yield
+ sched_get_priority_min
+ sched_get_priority_max
+
+Some minor bugs have been fixed. See the ChangeLog file for details.
+
+
+SNAPSHOT 1999-03-16
+-------------------
+
+Initial release.
+