* harbour/include/hbstack.h
* harbour/source/vm/estack.c
* harbour/source/vm/hvm.c
* harbour/source/debug/dbgentry.c
* harbour/source/debug/debugger.prg
* moved debugRequest flag and debugger internal structure
pointer to HVM stack
* allocate debugger internal structure dynamically in thread
thread local area when PCODE with debug information is executed
* changed STATIC s_oDebugger to THREAD STATIC
* send HB_DBG_VMQUIT after each thread termination when
debugger is activated
Now debugger can be used in MT programs and each thread can be
debugged separately (each has its own debuger). Please only
remember that thread sharing console window can overwrite other
threads debugger screen. There is also one limitation which I'll
try to remove later: only main thread debugger keeps information
about file wide STATIC names and line numbers with possible break
point places. This data should be shared between debuggers.
* harbour/include/hbapi.h
* harbour/source/vm/hvm.c
* harbour/source/vm/estack.c
* disabled all debugger related bindings and variables when
HB_NO_DEBUG macro is set
% moved s_fDebuging to HVM function stack frame to eliminate
saving/restoring this variable in each function/message call
Note: if you think it's good to block linking .prg code compiled
with -b (debugger support) with HVM compiled with HB_NO_DEBUG
then it's enough to disable whole hb_dbg_SetEntry() function
in hvm.c when HB_NO_DEBUG macro is set.
* harbour/source/rdd/wacore.c
; updated comments about xbase++ behavior in dbRelease()
* harbour/include/hbvm.h
* harbour/source/vm/hvm.c
* harbour/source/vm/dynlibhb.c
* harbour/source/vm/runner.c
! added MT protection for HB_LIBLOAD()/HB_LIBFREE() and
HB_HRBLOAD()/HB_HRBUNLOAD(). Please remember that you should
not use any external LoadLibrary() functions because they are
not protected for clean PCODE modules loading/unloading even
in ST mode. Using them can cause HVM internal structure corruption.
If we have such functions in contrib then they should be redirected
to our HVM ones.
There is still one problem in MT mode which has to be resolved yet.
When dynamic library containing PCODE using static variables
is loaded 1-st time then it will force resizing of internal array
with static variables. If in exactly the same moment some other
thread operates on static variable then it may cause corruption.
To resolve this problem we will have to divide continuous memory
block common for all modules where we store all static variables
into separated blocks bound with each PCODE module (symbol table).
* harbour/source/vm/dynlibhb.c
+ added support for DLL loading/unloading in OS2 builds.
Based on xHarbour code by Maurilio Longo - please test.
* harbour/source/vm/thread.c
+ added small description for hb_threadOnce() .prg function
* harbour/include/hbthread.h
! do not set TLS in POCC and XCC builds when _MSC_VER macro is set
This compilers also have this macro but dllspec( thread) does
not work for them (at least in version 4.50.15)
* harbour/utils/hbtest/rt_hvm.prg
+ added few tests which exploit by DMC bug. Be careful with
this compiler. In some cases it gives buggy final code, f.e.
try this:
#include <stdio.h>
typedef struct
{
double dd;
long long int ll;
} UN;
const char * cmp( UN * pU )
{
return pU->dd >= pU->ll ? "OK" : "ERRROR";
}
int main( void )
{
UN u = { 10.50, 10 };
printf( "%s\n", cmp( &u ) );
return 0;
}
The same bug can be exploited also by final Harbour binaries
compiled by DMC in code like:
? iif( 10.50 >= 10, "OK", "ERROR" )
I used DMC 8.42n. As long as such things will not be fixed by
DMC developers this compiler is not supported by Harbour and
reported problems with DMC Harbour builds will have to be
ignored by us.
* harbour/include/hbclass.ch
* harbour/source/vm/classes.c
* harbour/source/rtl/tobject.prg
* harbour/source/rtl/tclass.prg
! fixed race condition when class function is called 1-st time
and new class is created. It was possible that more then one
thread execute this function simultaneously and two classes
identical classes but with different handles were created
and registered in HVM
* harbour/contrib/xhb/xhbcomp.prg
* updated hack with optional inheritance for new class code
* harbour/source/compiler/gencc.c
! fixed thread static variable offsets generated in .c code when
-gc3 is used
* harbour/source/vm/hvm.c
% removed unnecessary call to HB_STACK_TLS_PRELOAD
* harbour/source/vm/hvm.c
! execute hb_vmThreadQuit() after hb_vmStackRelease()
* harbour/source/vm/thread.c
* do not destroy critical sections on exit
* harbour/include/hbthread.h
* casting cleanup
* harbour/include/hbapigt.h
* harbour/source/rtl/gtapi.c
* harbour/source/rtl/scroll.c
* harbour/source/rtl/console.c
* harbour/source/rtl/box.c
* changed hb_gtDrawBox() and hb_gtPutText() to operate on color
number (index in some GTs) instead of color string
* use int as color number (index) in hb_gtScrollEx() hb_gtBoxEx()
* harbour/source/rtl/box.c
% changed DISPBOX() with color and frame string to work as one
GT atomic operation
* harbour/source/vm/classes.c
! fixed profiler usage
* harbour/source/vm/hvm.c
* reduced the mutex lock time in thread exit state
* harbour/source/vm/thread.c
; added note about possible behavior of hb_threadSelf() function
in one very specific situation. It's documented and expected
behavior
* harbour/source/rtl/hbgtcore.c
+ added support for retrieving current HB_GTI_NOTIFIERBLOCK code block
+ added support for removing HB_GTI_NOTIFIERBLOCK code block without
setting the new one
* harbour/source/vm/garbage.c
* unblock HVM before executing destructors and releasing blocks
* harbour/source/vm/fm.c
! do not call internal error with active lock when FM statistic
is enabled
* harbour/source/vm/thread.c
! stop other threads in hb_threadOnce() if the 1-st one
executes bAction block to be sure that it will be completed
before thread leave hb_threadOnce() function
* harbour/source/vm/set.c
! do not call hb_inkeyReset() in set release. It's not necessary
and it could change keyboard buffer size for other threads using
the same GT driver.
* harbour/source/vm/hvm.c
% allocate new thread stack without lock
* harbour/source/rtl/hbgtcore.c
* harbour/source/rtl/gtapi.c
* moved clipboard resetting from hbgtcore.c to gtapi.c
* harbour/source/rtl/filesys.c
! do not use read/write operations with given file offset in Win9X
which seems to not support such functionality
* harbour/include/hbvm.h
* harbour/source/vm/hvm.c
+ added hb_vmProcessDynLibSymbols() which works like
hb_vmProcessSymbolsEx() but always mark registered modules as
dynamic
* harbour/source/vm/maindllp.c
* redirect hb_vmProcessSymbolsEx() to hb_vmProcessDynLibSymbols()
* harbour/source/vm/thread.c
* disabled some low level locking code in non MT HVM mode
* harbour/source/vm/hvm.c
* changed the place where exception handles are set/remove.
Now it's set at the beginning of hb_vmInit() and removed
at the end of hb_vmQuit()
* harbour/source/rdd/dbcmd.c
* harbour/source/rdd/dbf1.c
! added workaround for problem with EVAL block used in PACK
command
* harbour/source/compiler/hbmain.c
* minor simplification
* harbour/include/hbtypes.h
* harbour/source/vm/maindllp.c
! updated casting for current SVN code
[TOMERGE 1.0]
* harbour/include/hbthread.h
* harbour/source/vm/hvm.c
* harbour/source/vm/thread.c
+ added .prg function
hb_threadWait( <pThID> | <apThID>, [ <nTimeOut> ] [, <lAll> ] )
=> <nThInd> | <nThCount> | 0
This function wait for <nTimeOut> seconds (default infinite wait)
until one or all (controlled by <lAll> parameter) of given HVM threads
will have finished execution in HVM area. It does not join the threads.
It returns index to 1-st thread which has finished execution (default)
or number of threads which has finished in given time period when lAll
is .T.
* source/vm/cmdarg.c
* HB_ARGV() will now return the executable name if called
without a parameter.
! HB_ARGSTRING() fixed to return "" in some error
cases instead of NIL.
* harbour/source/vm/runner.c
* harbour/source/vm/codebloc.c
* harbour/source/vm/itemapi.c
* minor casting modification
* harbour/tests/speedtst.prg
* added workaround for timeout parameter not working in xHb Subscribe()
when 3-rd parameter is passed
* harbour/include/hbstack.h
* harbour/source/vm/estack.c
* harbour/source/vm/hvm.c
% moved static USHORT uiPolls to HVM stack to not reduce the HVM
performance in MT mode. In multi CPU environment if many threads
try to change the same memory variable in the same time the overall
performance is strongly reduced on most of tested hardware - probably
some random CPU delay used to resolve the conflict.
* include/hbapierr.h
* source/vm/extrap.c
* source/rtl/errorint.c
! Fixed recent modification on Windows platform, where
GPF would generate an internal error.
It didn't launch the OS error handler.
! Fixed hb_errInternal() being marked as HB_EXPORT
in the headers only.
* harbour/include/Makefile
+ harbour/include/hbatomic.h
* harbour/include/hbthread.h
* harbour/source/vm/garbage.c
* harbour/source/vm/fm.c
* moved atomic and spinlock functions into hbatomic.h
* harbour/include/hbatomic.h
+ added atomic inc/dec for GCC and x86@64 and PPC@32
+ use OSAtomic*() and OSSpin*() functions for atomic inc/dec and
spinlocks in Darwin builds
+ added spinlocks to MS-Win builds
* harbour/source/vm/hvm.c
! fixed last commit typo in thread number allocating - all threads
where using 0 number
* harbour/source/vm/thread.c
! fixed return value in recursive call to hb_mutexLock() - was FALSE
! fixed typo in timeout parameter number
should be 2 instead of 1
* harbour/source/rtl/tobject.prg
! fixed init() return value - it should be ignored and current
object should be returned
* harbour/source/rtl/tthreadx.prg
! probably fixed timeout parameter use in signal:wait() method
I guess that like other xbase++ functions it's in 1/100th of
a second
* harbour/include/hboo.ch
* harbour/include/hbclass.ch
* harbour/source/vm/classes.c
* harbour/source/rtl/tclass.prg
+ added support for SYNC object and CLASS messages working like
in xBase++. Please remember that all locked by thread SYNC messages
objects and classes are unblocked automatically for the time which
thread spends in hb_subscribe[Now]() or signal:wait()
* harbour/include/hbclass.ch
* harbour/include/hboo.ch
+ added SYNC attribute to accepted syntax in class declaration
it allows to compile code which uses this attribute but low
level implementation is not ready yet - work in progress
* harbour/source/vm/garbage.c
! added #define INCL_DOSPROCESS for OS2 builds. For some reasons
#define INCL_DOSDATETIME I've found in DosSleep() documentation
is note enough
* harbour/tests/speedtst.prg
* marked T054 as memory tests
* harbour/source/vm/hvm.c
* harbour/source/vm/thread.c
* added hb_threadOnce( @<onceControl> [, <bAction> ] ) -> <lFirstCall>
This function allow to execute some code only once. It's usefull in
MT environment for initialization.
<onceControl> is variable which holds the execution status and have
to be initialized to NIL. In most of cases it will be simple static
variable in user code.
<bAction> is optional codeblock which is executed only once (on 1-st
call with given <onceControl>)
* harbour/source/rtl/filesys.c
! do not make any file name conversions in hb_fsNameConv() if HVM stack
is not allocated
* harbour/source/rtl/gtcrs/gtcrs.c
* casting
* harbour/tests/speedtst.prg
+ added support for --scale switch - it enables special mode
which compare scalability, f.e.:
./speedtst --thread=4 --scale
executes each test loop 4 times by single thread then
by 4 simultaneous threads, counts total real times in both
cases and presents them with the scale factor.
Please note that exactly the same loop tests are executed
simultaneously so this test should well show all places which
are not well scalable.
This test allows me to locate yet another memory variable written by
each thread on each function call just like uiPolls in main HVM loop.
It's s_bDebugging which is not used in MT mode because we do not have
any MT debugger yet but is was effectively killing the scalability
in my hardware.
If possible I would like to see some Windows results from multi CPU
machine with --scale parameter, f.e.:
speedtst --thread=8 --scale > speedtst.log
Please make some tests.
* harbour/source/vm/hvm.c
% do not assign s_bDebugging in hb_vmDo()/hb_vmSend() when debugger
is not linked
* harbour/source/vm/garbage.c
% enabled spinlocks in all x86 GCC builds
* harbour/source/vm/fm.c
! do not build DL-MM when HB_FM_WIN32_ALLOC is set
* harbour/tests/speedtst.prg
+ added support for --exclude=<tests,...> option to exclude some tests,
f.e.:
./speetest --exclude=029.030.023.025.027.040.041.043.052.053.019.022.031.032
disables all tests which allocates memory. The same tests can be also
excluded by --exclude=mem
* source/vm/extrap.c
+ Now using standard internal error call for Windows too.
+ Implemented //BATCH for OS/2. (please test)
* source/rtl/errorint.c
* Upped internal buffer.
* contrib/hbole/oleauto.prg
* contrib/hbole/ole2.c
* contrib/hbwin/win_tole.prg
* contrib/hbwin/win_ole.c
* OLESHOWEXCEPTION() implemented on the .prg level, using
Alert() rather than MessageBox().
! All MessageBox()es converted to Alert()s.
+ Added OLEEXCEPTIONSOURCE(), OLEEXCEPTIONDESCRIPTION()
to hbole.
* source/vm/maindllp.c
! Changed MessageBox() call to internal error.
* contrib/gtwvg/gtwvg.c
* source/rtl/gtwvt/gtwvt.c
! Changed MessageBox() calls to internal error.
* contrib/gtwvg/gtwvg.c
* Internal error calls synced with GTWVT.
* source/rtl/hbgtcore.c
- Removed "Internal error" text from two internal error descriptions.
* make_vc.mak
- Removed reference to MessageBox() from comment.
; NOTE: No more MessageBox() calls in Harbour.
(except some wrappers for user .prg code)
* include/hbsetup.ch
! Removed some company names from comments.
* source/vm/extrap.c
+ Added //BATCH internal switch to indicate exception handler
to suppress the Windows exception dialog when an exception
occurs. Useful for programs running without user interaction.
NOTE: This may not be the ultimate method. If we can gather
more such situations where such "batch" mode might
be useful, we may as well implement it as a Set().
- Removed Harbour specific exception MessageBox(). Besides
having gotten rid of a MessageBox() call in core VM, Harbour now
got back the standard Windows error handling dialog, which is
more standard and easy to get by than the previous default
method, where several clicks were needed to pass through
the (sometimes hidden) popups. The details of all GPFs are
still logged into hb_out.log.
+ Removed size limit for the logged exception call stack.
* harbour/source/rdd/sdf1.c
! fixed stupid typo (missing !) in last modification which broke
SDF RDD read mode. Thanks to Marco for information and example.
* harbour/include/hbstack.h
* harbour/source/vm/estack.c
* small cleanup in used macros
* harbour/include/hbset.h
* harbour/source/vm/set.c
+ divided __SETCENTURY() PRG function into two functions:
__SETCENTURY() and public C function hb_setSetCentury()
to manipulate century setting from C code.
Modyfication by Francesco Saverio Giudice
* harbour/include/hbvm.h
* harbour/include/hbthread.h
* harbour/source/vm/hvm.c
* harbour/source/vm/thread.c
+ added hb_vmThreadRegister() funtion which allows to register new
thread in HVM without HVM stack allocation
+ added hb_vmThreadRelease() which removes registered thread which
does not have allocated HVM stack yet
+ added hb_threadStateNew() which returns new thread control state
which can be used as parameter to hb_vmThreadInit(),
hb_vmThreadRegister(), hb_vmThreadRelease()
* register thread control state also for main thread
Now hb_threadSelf() returns thread pointer also for main thread
! fixed possible race condition in hb_threadWaitForAll() and
hb_threadTerminateAll() by using hb_vmThreadRegister() before
starting thread.
* harbour/source/rtl/idle.c
* interrupt hb_idleSleep() on QUIT or BREAK requests
* harbour/tests/mt/mttest09.prg
* small cleanup
* harbour/source/vm/thread.c
* release thread return value when is retrieve
* harbour/common.mak
* harbour/source/rtl/Makefile
+ harbour/source/rtl/tthreadx.prg
+ added xBase++ compatible signal class
* harbour/source/rdd/wacore.c
! added missing alias setting after attaching work area
+ harbour/tests/mt/mttest09.prg
+ added test code for using the same work area in different threads
* harbour/include/hbstack.h
* harbour/source/vm/estack.c
+ added assembler inline macros for TLS access in BCC and MinGW32
They are enabled when Harbour is compiled with -DHB_NO_TLS.
Please check the speed difference. If it will be noticeable
then we keep them probably as default. Otherwise they will be
removed.
