* harbour/source/rtl/gttrm/gttrm.c
! do not catch SIGCHLD
* harbour/include/hbapifs.h
* harbour/source/rtl/hbproces.c
+ added C function: hb_fsProcessRun()
* harbour/source/rtl/hbprocfn.c
+ added .prg function:
hb_processRun( <cCommand>, [ <cStdIn> ], [ @<cStdOut> ], ;
[ @<cStdErr> ], [ <lDetach> ] ) -> <nResult>
This function is implemented for all builds also in DOS ones where
temporary files are used to simulate pipes.
TODO: in OS2 builds it should be possible to implement this function
without temporary files just like in *nixes and MS-Windows builds.
I would like to ask OS2 users to make it.
* utils/hbmk2/hbmk2.prg
+ Added app bundle generation for OS X targets in -gui mode.
It's experimental yet. (f.e. clean will not clean it f.e.)
Thanks to Teo Fonrouge for the hints.
* Using escapting function (instead of manual escaping) when
launching Windows apps on NT.
; TOFIX: hb_FCopy()
* source/rtl/fscopy.c
+ Now it will copy flags on *nix systems (just like __COPYFILE()).
* source/rtl/copyfile.c
* Minor formatting.
; TODO: Add generic app icon support. Possible?
* INSTALL
- Deleted QT version number from example settings to
ease on maintenance without making the doc look outdated.
* source/rtl/hbmd5.c
! Minor typo in comment.
* config/win/global.cf
* config/wce/global.cf
% Removed unnecessary @ chars.
* contrib/hbmisc/Makefile
+ contrib/hbmisc/fcomma.prg
- source/rdd/usrrdd/rdds/fcomma.prg
* source/rdd/usrrdd/rdds/Makefile
* fcomma.prg moved to hbmisc, since it depends on
hbmisc features.
* examples/hbextern/hbextern.prg
- Removed special treatment of fcomma.prg.
+ harbour/include/hbtask.h
+ harbour/source/vm/task.c
* harbour/include/hbthread.h
* harbour/include/hbatomic.h
* harbour/source/vm/thread.c
* harbour/source/vm/hvm.c
* harbour/source/vm/fm.c
* harbour/source/rtl/idle.c
* harbour/source/rtl/filesys.c
+ implemented OS independent task switching system - it gives PTHREAD
compatible basic API so it can be used in HVM as alternative MT support
which does not use any OS threads. As long as Harbour does not call
any blocking OS function then it's possible to create and execute
simultaneously many threads though only one CPU is used and switched
between HVM threads. It gives similar scalability to xbase++ threads
and also similar behavior in item protection at .prg level.
Now it's possible to use HVM threads in any OS.
Of course it does not mean that Harbour adds in some magic way
thread support to OS-es which does not support threads like DOS.
It only means that HVM supports threads for .prg code just like
in native MT environment as long as some C code does not block
task switching or process execution will not be frozen by sth, i.e.
executing other process (__run()) in single process OS like DOS.
In some cases it can be interesting alternative even in OS which
have native thread support.
All tests/mttest*.prg programs and speedtst --thread=<n> --scale
are executed correctly with new task switching just like with
OS native MT support.
Compilation with task switching in hbvmmt library can be forced
by HB_TASK_THREAD macro which also disable native OS threads
support.
For task context switching two alternative methods are used:
1) getcontext()/makecontext()/swapcontext() (SUSv2, POSIX.1-2001)
which is preferable because does not need any additional
hacks but not all OS-es supports these functions.
It's enabled by default in Linux builds.
2) setjmp()/longjmp() (POSIX, ISO 9899 (C99)) otherwise.
These functions are supported by most of C compilers
but there is no function to set new stack in saved context
so it's necessary to introduce for each architecure/C compiler
peace of code which makes it. Macro HB_TASK_STACK_INIT() in
task.c makes it. I defined this macro for x86@32 in DJGPP
Linux GCC and OpenWatcom builds. I tested OpenWatcom builds only
in DOS and Linux but probably it works in all x86@32 builds.
If someone is interesting in adding support for some other
platforms which does not support ucontext.h and 1-st methods
then please define above macro for them.
Have a fun with new toy ;-)
* harbour/source/vm/Makefile
* enabled hbvmmt in DJGPP and OpenWatcom DOS builds. It works well.
Viktor if possible please add support for -mt switch in hbmk2
in all builds even if we do not compile hbvmmt by default so
it can be used with DJGPP and OW and any other builds for which
someone enable hbtask.c though OS does not support threads.
* harbour/contrib/hbmzip/hbmzip.c
! fixed '[const] char|BYTE *' casting in DOS and OS2 builds
* harbour/source/macro/macro.y
* harbour/source/macro/macro.yyc
* harbour/source/compiler/harbour.y
* harbour/source/compiler/harbour.yyc
% reduced INASSING definition - thanks to Xavi for information
* harbour/source/compiler/hbopt.c
! fixed stupid bug in logical expression optimization - thanks to Randy
for self contain example
* harbour/include/hbapi.h
* harbour/include/hbmacro.h
* harbour/include/hbexprb.c
* harbour/source/vm/macro.c
* harbour/source/vm/hvm.c
* harbour/source/macro/macro.y
* harbour/source/macro/macro.yyc
+ added direct support for pushing references to macro expressions.
It resolves the problem with Clipper compatibility (thanks to Alex
for information and example) in code like:
m->var := {0}
&("var[1]") += 10
and also many other things, f.e. now Harbour can cleanly compile
and execute this code:
proc main
local s
m->o := errorNew()
s:="o:tries"
? m->o:tries
? &s
? &(s+"+=2")
? m->o:tries
? &(s+":=3")
? m->o:tries
?
&s := 4
? m->o:tries
&s -= 2
? m->o:tries
&s++
? m->o:tries
?
s := "o"
m->o := 1
? &s
&s := &s + 4
? &s
? &s += 5
? --&s
?
s := "o[2]"
m->o := { 1, 2 }
? &s
&s := &s + 4
? &s
? &s += 5
? --&s
return
* harbour/utils/hbtest/Makefile
* force hbtest compilation with line numbers - information about line
numbers is important part of this test.
* include/hbcomp.h
* include/hbcompdf.h
* include/hbapifs.h
* include/hbexprop.h
* include/hbmacro.h
* include/hbapi.h
* include/hbapiitm.h
* include/hbdate.h
* source/pp/ppcore.c
* source/vm/itemapi.c
* source/vm/arrays.c
* source/vm/extend.c
* source/common/expropt1.c
* source/rtl/philes.c
* source/rtl/empty.c
* source/rtl/minmax.c
* source/rtl/dates.c
* source/rtl/filesys.c
* source/rdd/dbfnsx/dbfnsx1.c
* source/rdd/dbfcdx/dbfcdx1.c
* source/rdd/dbf1.c
* source/rdd/dbffpt/dbffpt1.c
* source/rdd/hbsix/sxdate.c
* source/compiler/hbmain.c
* source/compiler/complex.c
* source/compiler/harbour.yyc
* source/compiler/harbour.y
* contrib/hbct/files.c
* contrib/hbct/dattime2.c
* contrib/hbct/datetime.c
* contrib/xhb/filestat.c
* contrib/hbtip/utils.c
* examples/hboleold/win_ole.c
* LONG -> long for date/time related values
(julian, date, type, millisec)
; This is an effort to clean Harbour sources from
Windows/legacy-Clipper basic types and replace
them with own or ANSI C ones.
In the above case, usage wasn't consistent,
LONG and long were mixed.
* source/main/harbour.c
* source/vm/fm.c
* LONG -> long for some stat counters.
* examples/hbextern/hbextern.prg
* Excluding all HB_ARCHITECTURE dirs explicitly.
(this resolves the recursion problem)
* Excluding fcomma.prg explicitly.
- Deleted 'example' from exclusion list.
- source/rdd/usrrdd/example
+ tests/usrrdd
* Moved examples under test dir.
* source/rdd/usrrdd/rdds/Makefile
- Removed fcomma.prg from builds since it depends on
hbmisc contrib libs, on functions which aren't MT
compatible and have other problems too.
* contrib/Makefile
* Moved hbxbp to the end of list as a workaround for potentially
outdated hbqt.ch reference, if it's before hbqt build.
Still not prefect though. We should probably fix the ordering
of -I option in build process, but it doesn't seem trivial.
Or avoid installing headers into central include dir, but that
also leads far.
* utils/hbmk2/hbmk2.prg
* Minor step to fix a recent regression.
* source/rtl/hbzlibgz.c
* HB_GZDIRECT() is now always defined in core, if it's built against
older zlib version, it will simply do nothing.
We can discuss how to handle this situation, all have pros and cons:
1) Always define: Throw RTE at runtime.
2) Always define: Return error at runtime.
3) Don't define if low-level wrapper isn't present. Makes it
easy to catch missing dependencies on .prg level, but it
break Harbour to provide a consistent higher level API.
* examples/hbextern/hbextern.prg
! Minor fix in generated comment.
* examples/hbextern/hbextern.hbp
+ -w2 -> -w3
* ChangeLog
! Wrong header in previous entry.
* contrib/hbqt/hbqt_qimage.cpp
* contrib/hbqt/hbqt_qtextdocument.cpp
* contrib/hbqt/hbqt_qapplication.cpp
* contrib/hbqt/hbqt_qtextcursor.cpp
* contrib/hbqt/hbqt_qclipboard.cpp
* contrib/hbqt/hbqt_qurl.cpp
* contrib/hbqt/hbqt_qwebhistoryitem.cpp
* contrib/hbqt/hbqt_qvariant.cpp
* contrib/hbqt/hbqt_qfont.cpp
* contrib/hbqt/hbqt_qimagereader.cpp
* contrib/hbqt/hbqt_qcursor.cpp
* contrib/hbqt/hbqt_qfontdatabase.cpp
* contrib/hbqt/hbqt_qcoreapplication.cpp
* contrib/hbqt/hbqt_qobject.cpp
* contrib/hbqt/hbqt_qtextframe.cpp
* contrib/hbqt/hbqt_qabstractitemmodel.cpp
* Minor EOF change after regerenating them with hbqtgen.
(probably it would be better to generate only one EOLs at
EOF by hbqtgen)
* source/rtl/gtwvt/gtwvt.c
! Fixed maximize button state when HB_GTI_RESIZABLE is set to .T.
before the window is drawn.
Please test.
% Cleaned windows style definitions in source.
* include/hbstack.h
* include/hbapi.h
* Formatting in comment.
* harbour/include/hbapirdd.h
* harbour/source/rdd/workarea.c
* harbour/source/rdd/wafunc.c
* harbour/source/rdd/dbcmd.c
* harbour/source/rdd/hsx/hsx.c
* harbour/source/rdd/usrrdd/usrrdd.c
* modified RDD method declaration (DBENTRYP_*) to eliminate casting.
3-rd party RDD code have to be updated.
* modified some RDD structures to force more strict alignment.
Warning!!! AREA structured modified - all 3-rd party RDDs which
redefines AREA members instead of using AREA structure
directly have to be updated.
* harbour/include/hbrdddbf.h
* harbour/include/hbrddnsx.h
* harbour/include/hbrdddel.h
* harbour/include/hbrddsdf.h
* harbour/include/hbrddcdx.h
* harbour/include/hbrddntx.h
* harbour/source/rdd/delim1.c
* harbour/source/rdd/sdf1.c
* harbour/source/rdd/dbf1.c
* harbour/source/rdd/hbdbsort.c
* harbour/source/rdd/dbfcdx/dbfcdx1.c
* harbour/source/rdd/dbfnsx/dbfnsx1.c
* harbour/source/rdd/dbfntx/dbfntx1.c
* harbour/source/rdd/dbffpt/dbffpt1.c
* harbour/contrib/rddads/rddads.h
* harbour/contrib/rddads/ads1.c
* harbour/contrib/rddads/adsx.c
* harbour/contrib/rddads/adsfunc.c
* harbour/contrib/rddsql/hbrddsql.h
* harbour/contrib/rddsql/sqlbase.c
* harbour/contrib/rddsql/sqlmix.c
* harbour/contrib/rddsql/sddmy/mysqldd.c
* harbour/contrib/rddsql/sddpg/pgsqldd.c
* harbour/contrib/rddsql/sddfb/fbirddd.c
* harbour/contrib/rddsql/sddodbc/odbcdd.c
* harbour/contrib/hbbmcdx/bmdbfcdx.c
* harbour/contrib/hbbmcdx/hbbmcdx.h
* updated for new method declaration
* modified RDD AREA structures definition to use supper AREA structures
directly as 1-st member instead of redefining individual AREA members.
It resolves the synchronization problem so now modification in super
AREA structures does not force modifications in source code of
descendant RDDs. It also eliminates possible typos in manual structure
synchronization.
* include/hbapifs.h
* source/rtl/filebuf.c
* source/rtl/filesys.c
* Changed file I/O buffer parameters from
'( [const] BYTE * )' to '( [const] void * )'
* utils/hbmk2/hbmk2.prg
+ Added ${hb_curdir} macro support.
+ Added ${hb_first} macro support (will return first source name
specced on cmdline)
+ Added support for 'sources=' line in .hbc files. This
will accept .prg, .c/.cpp, .po/.pot, object and resources.
For resources, also wildcards are supported. Files with
no extension will be considered .prg.
This also replaces former 'pos=' line.
This can be useful to put some central sources into all
project without using a lib (or when a lib isn't an option
for some reason).
- man
- Dir deleted.
* source/common/hbhash.c
! Fixed error in prev commit.
(strange that MSVC didn't complain)
* mpkg_win.nsi
+ Added comment that this file shouldn't be used directly.
* source/rtl/hbmd5.c
+ HB_MD5(), HB_MD5FILE(): New logical parameter added, which
controls whether to return binary data or hexadecimal string.
The default is .T. which will return hexa string, to stay
compatible. Later we should switch the default to .F..
* mpkg_gnu.bat
* Updated requirement section.
* contrib/hbct/ascpos.c
* contrib/hbct/ctwfunc.c
* contrib/hbct/ctcrypt.c
* contrib/xhb/filestat.c
* contrib/hbsqlit3/hbsqlit3.c
* contrib/hbnf/fttext.c
* contrib/hbpgsql/postgres.c
* contrib/hbfimage/fi_wrp.c
* contrib/hbgd/gdwrp.c
* source/rtl/mouse53.c
* 'const' cleanup.
; NOTE: I rather didn't commit BYTE to char in CTWIN.
; TOFIX: Some of these will result in warnings when passing values
to external API calls. I'd like to ask developers to review
what the intent was. Sometimes external API declarations
are missing 'const', so it's safe to pass them these buffers,
but it should be checked in documentation.
---
../../fi_wrp.c: In function 'HB_FUN_FI_LOADFROMMEM':
../../fi_wrp.c:372: warning: passing argument 1 of 'FreeImage_OpenMemory' discards qualifiers from pointer target type
C:\devl\FreeImage-3.12.0\Dist/FreeImage.h:771: note: expected 'BYTE *' but argument is of type 'const BYTE *'
../../fi_wrp.c: In function 'HB_FUN_FI_GETFILETYPEFROMMEM':
../../fi_wrp.c:887: warning: passing argument 1 of 'FreeImage_OpenMemory' discards qualifiers from pointer target type
C:\devl\FreeImage-3.12.0\Dist/FreeImage.h:771: note: expected 'BYTE *' but argument is of type 'const BYTE *'
../../gdwrp.c: In function 'HB_FUN_GDIMAGESTRINGFTEX':
../../gdwrp.c:2718: warning: passing argument 4 of 'gdImageStringFTEx' discards qualifiers from pointer target type
C:\devl\gd-2.0.34\include/gd.h:459: note: expected 'char *' but argument is of type 'const char *'
../../gdwrp.c:2718: warning: passing argument 9 of 'gdImageStringFTEx' discards qualifiers from pointer target type
C:\devl\gd-2.0.34\include/gd.h:459: note: expected 'char *' but argument is of type 'const char *'
../../gdwrp.c: In function 'HB_FUN_GDIMAGESTRINGFTCIRCLE':
../../gdwrp.c:2809: warning: passing argument 7 of 'gdImageStringFTCircle' discards qualifiers from pointer target type
C:\devl\gd-2.0.34\include/gdfx.h:41: note: expected 'char *' but argument is of type 'const char *'
../../gdwrp.c:2809: warning: passing argument 9 of 'gdImageStringFTCircle' discards qualifiers from pointer target type
C:\devl\gd-2.0.34\include/gdfx.h:41: note: expected 'char *' but argument is of type 'const char *'
../../gdwrp.c:2809: warning: passing argument 10 of 'gdImageStringFTCircle' discards qualifiers from pointer target type
C:\devl\gd-2.0.34\include/gdfx.h:41: note: expected 'char *' but argument is of type 'const char *'
---
* harbour/source/rtl/dates.c
! fixed CTOD() to be CA-Cl*pper compatible in decoding dates like:
SET DATE FORMAT TO "YYYY-MM-DD"
? CTOD("000002009.000000012.0000000000000027.1")
? CTOD(" 0002009. 000012. 000028 1")
? CTOD("xx0002009.xxx000012.xxxxxxxxxx000028 1")
? CTOD("000 02009.00 000012.000 000028 1"), "*"
? CTOD("2007 11 29")
? CTOD(" 2007 11 29 ")
? CTOD("2007-1 1-29")
? CTOD("2007- 1 1 - 2 9 ")
! fixed HB_CTOT() to use the same formatting rules as CTOD() in
CA-Cl*pper - now delimiters between values are strictly required.
* INSTALL
* make_gnu.bat
+ Added support for Cygwin builds (passes call to make_gnu.sh).
+ Added Cygwin build starter example.
+ Cygwin mentioned as a requirement for mingwarm builds,
also added to PATH.
! Fixed cygwin autodetection.
! Fixed wrong PATH order in 'MinGW GCC + MSYS' example.
* Merged mingw+msys and cygwin bash based build methods,
and titled as "alternate" method.
Now all supported compilers work by starting up make_gnu.bat.
* utils/hbmk2/hbmk2.prg
! Change compiler autodetection to work for cygwin.
* mpkg_dos.bat
* mpkg_win.bat
* make_gnu.bat
! Fixed build process to work in paths containing spaces.
I didn't test the batch in pure DOS, so I may have broken it.
* config/win/install.cf
* config/wce/install.cf
! Fixed 'install' to work when there is a space in install path.
* make_gnu.bat
* INSTALL
+ Added support for HB_BUILD_LOG envvar. If set to 'yes',
it will automatically set log output filename to:
'log-<arch>-<comp>.txt'
if set to 'no' it will no make any extra redirection,
if set to any other value, it will be used for output
(both std and err). The batch will set HB_BUILD_LOGFILE
to the filename used for output for processing.
This feature requires Windows NT or upper host.
I didn't enable this feature yet by default to not break
existing user setups, but maybe I'll do it before the
final release.
* Modified text shown (+ beep thrown) when build failed.
* Messages now all start with 'Harbour'.
* Do not do the second build pass if the first one failed.
* contrib/hbwin/legacy.prg
! Fixed CreateObject() not working due to typo.
Thanks to Itamar Lins for the report.
* mpkg_dos.bat
* mpkg_win.bat
* mpkg_win.nsi
+ Enabled HB_BUILD_LOG=yes.
* Silenced copy, zip, makensis commands.
+ Added message about failed build process.
* Changed to use 'copy' instead of 'xcopy'
* Changed internals so now they don't require HB_COMPILER
and HB_ARCHITECTURE to be set. This means that now
we can rely on autodetection even for mpkg_*.bat files.
Creating an installer is very easy now for most cases.
Just type:
mpkg_gnu.bat
* source/vm/set.c
+ Added file commit support to __MVSAVE(), if Set( _SET_HARDCOMMIT )
is set to .T.. This is the default which means by default we aren't
CA-Cl*pper compatible. Addition protected by HB_C52_STRICT.
* source/rdd/dbf1.c
! Minor typo in comment.
* doc/en-EN/hb_apiit.txt
* include/hbapi.h
* include/hbapiitm.h
* source/vm/itemapi.c
* source/vm/extend.c
* source/rtl/gtwvt/gtwvt.c
* source/rdd/dbf1.c
* contrib/gtqtc/gtqtc.cpp
* contrib/gtwvg/gtwvg.c
* contrib/gtwvg/wvggui.c
* contrib/hbwin/olecore.c
* examples/hboleold/win_ole.c
* Finishing long planned function renames to clear up
a very (10 years) old naming inconsistency.
To adapt you code to this, please do changes below:
1) hb_itemPutCPtr() (with three parameters) -> hb_itemPutCLPtr()
2) hb_itemPutCPtr2() -> hb_itemPutCPtr()
Unmodified code will generate compile time errors, so it's easy
to spot where the modifications need to be done. For most users,
only case 1) will need to be dealt with.
; INCOMPATIBLE. Please update your code.
* source/rtl/hbi18n1.c
* source/rdd/dbfnsx/dbfnsx1.c
* hb_itemPutCPtr() (with three parameters) -> hb_itemPutCLPtr()
* contrib/gtwvg/wvggui.c
* contrib/gtqtc/gtqtc.cpp
! Fixed GPF in hb_gtInfo( HB_GTI_SCREENSIZE ) call with wrong
parameter.
* include/hbapi.h
* include/hbapiitm.h
* source/vm/itemapi.c
* source/vm/arrays.c
* source/vm/extend.c
; Started addition of 'const' keywords where applicable.
* First pass:
extern HB_EXPORT const char * hb_parc( int iParam );
extern HB_EXPORT const char * hb_parcx( int iParam );
extern HB_EXPORT const char * hb_itemGetCPtr( PHB_ITEM pItem );
extern HB_EXPORT const char * hb_arrayGetCPtr( PHB_ITEM pArray, ULONG ulIndex );
; NOTE: This modification will break code in most compilers,
especially in C++ mode.
As a next steps source code will have to be modified
according to this change. Be patient, don't report all
errors to the developers. You can join the code cleanup.
* source/vm/dynsym.c
* source/vm/itemapi.c
* source/vm/arrays.c
* source/vm/memvars.c
* source/vm/extend.c
* source/vm/set.c
* source/vm/classes.c
* 'const's added to caller code.
* source/vm/memvars.c
- Deleted hb_fsCommit() as non CA-Cl*pper compatible.
I'll let others decide what to do with this patch.
* contrib/hbclipsm/date.c
! Changed hb_parnl() calls used to extract dates to hb_pardl().
* source/vm/memvars.c
+ Added hb_fsCommit() to __MVSAVE().
(Change copied from xhb, created by Peter Rees)
* contrib/gtwvg/wvgsink.c
! Fixed BCC warning after recent change.
(BTW, does anyone know a switch to enable such warning:
"'x' is assigned a value that is never used" with *MinGW* ?
It's annoying that BCC catches it, while other compilers
don't. I'm now only using BCC when creating builds.)
