* src/rtl/filebuf.c
! fixed file handle sharing in POSIX systems when writeonly mode is used
! fixed accessing files in writeonly mode in POSIX systems which use
fcntl() locks for DENY_* flag emulation.
! added for POSIX systems protection against potential race condition in
MT programs and file open and close operations. Now the whole low level
open and close operations are covered by mutex. It means that they are
fully serialized so in case of remote file access (i.e. NFS or SMBFS)
it may reduce the performance in applications which simultaneously open
and closes many different files in different threads but its necessary.
For local file access it should not create noticeable scalability
problems. In my Linux box Harbour can open and close in MT mode more
then 250000 files per second.
% use different mutexes for file open/close and file lock operations to
not reduce the lock performance by potentially slow open() operation
in POSIX systems.
; NOTE: please remember that DENY_* flag emulation in POSIX systems
emulates only shared and exclusive modes so if any of
FO_DENYREAD, FO_DENYWRITE or FO_EXCLUSIVE flag is used then
it effectively works like FO_EXCLUSIVE in POSIX system and
is not mandatory for other processes so it blocks only
other Harbour applications which use Harbour virtual handles
(RDD code, hb_vf*() functions)
* src/rtl/filesys.c
* do not convert exclusive lock to shared one when file is open in
readonly mode on POSIX system using flock() for DENY_* flag emulation.
Now HB_USE_BSDLOCKS is enabled only for Linux and it's documented that
this implementation allow to use shared and exclusive locks regardless of
file open mode.
% do not create file name copy in hb_fsExtOpen() when passed parameters
does not force file name modification
* include/hbapicls.h
* src/vm/classes.c
* minor modification in variable name
* src/rtl/filebuf.c
* src/rtl/fscopy.c
* src/rtl/fslink.c
* src/rtl/vfile.c
* cleaned FError() setting. It should not be changed by C file IO
functions because they can be activated from RDD code so accessing
work area field could change the FError() result, i.e. this code
does not work when FError() is changed by C level IO operations:
AnyFunc( FSeek( hFile, 0, FS_END ), field->MEMOVAL, FError() )
Even simple field access may cause file IO operations in RDD code
when relations are set because records in child area are repositioned
only when necessary (when user tries to make some operations on this
area). In the past RDD code was full of constructions like:
uiSaveError = hb_fsFError();
// make some IO operations
hb_fsSetFerror( uiSaveError );
to eliminate such overhead I separated PRG level FError() from
C level IOError(). Please keep it.
* ChangeLog.txt
! few typos
* include/hbapifs.h
* src/rtl/filebuf.c
* added new C function hb_fileLoadData() - it can be used to load data
from already opened regular files and streams
* include/harbour.hbx
* src/rtl/vfile.c
+ added new PRG function:
hb_vfLoad( <cFileName>, [ <nMaxSize> ] ) -> <cFileBody> | NIL
* src/vm/runner.c
* use hb_fileLoadData()
* contrib/hbwin/win_bmp.c
* use hb_fileLoad()
* include/hbapifs.h
* src/rtl/filebufd.c
* moved HB_FILE_ERR_UNSUPPORTED macro to header file
* include/hbapifs.h
* src/rtl/filebuf.c
* added new C function hb_fileLoad() - it can be used to load data
from regular files and streams
* src/rtl/memofile.c
* modified [hb_]Memo{Read|Writ}() to work with streams, i.e. now
this code works:
REQUEST HB_PIPEIO
cDir := hb_memoRead( "|ls -la" )
? upper( cDir )
or:
cData := hb_memoRead( "|xz -c data.gz" )
? hb_memoWrit( "data2.txt", cData )
? hb_memoWrit( "|sh -c 'xz -9 -e > data2.xz'", cData )
* src/rtl/fscopy.c
* removed not used in current code unix header files
* ChangeLog.txt
! fixed typo in ChangeLog entry. I used hb_socketNew() instead of
hb_socketSetFilter()
* include/hbapifs.h
* src/rtl/filebuf.c
+ added helper C function:
HB_SIZE hb_fileResult( HB_SIZE nSize );
It converts ( HB_SIZE ) FS_ERROR to 0 so it can be used to wrap
hb_fileRead()/hb_fileWrite() to force previous results.
* src/rdd/dbffpt/dbffpt1.c
* reduced variable scope
* src/rtl/copyfile.c
! fixed typo in last commit - thanks to Viktor.
2015-07-31 14:04 UTC+0200 Viktor Szakats (vszakats users.noreply.github.com)
* include/hbapifs.h
* src/rtl/filebuf.c
* src/rtl/vfile.c
+ add C-level hb_fileSizeGet() function
* update HB_VFSIZE() to use hb_fileSizeGet() when passed a filename
It means now HB_VFSIZE() works for non-virtual filenames
just like HB_FSIZE() did.
* include/hbapifs.h
* src/rtl/filebuf.c
* src/rtl/filebufd.c
* src/rtl/filesys.c
* src/rtl/vfile.c
* src/rtl/iousr.c
* src/rdd/dbfcdx/dbfcdx1.c
* src/rdd/dbffpt/dbffpt1.c
* src/rdd/dbfnsx/dbfnsx1.c
* src/rdd/dbfntx/dbfntx1.c
* src/rdd/hsx/hsx.c
* contrib/hbcomio/comio.c
* contrib/hbmemio/memio.c
* contrib/hbtcpio/tcpio.c
* contrib/hbnetio/netio.h
* contrib/hbnetio/netiocli.c
* contrib/hbnetio/netiosrv.c
* changed HB_USHORT uiAttr parameter to HB_FATTR nAttr
in hb_fsExtOpen() and hb_fileExtOpen()
* include/fileio.ch
* src/rtl/vfile.c
+ added support for FO_DEFAULTS in hb_vfOpen().
When this flag is used in <nModeAttr> parameters then new
_SET_DEFAULT is respected when new file is created and
_SET_DEFAULT + _SET_PATH when file is opened.
+ added support for checking file size by it's in hb_vfSize()
so it can work like hb_FSize() but for virtual FS not only
for local files. Now this function has the following syntax:
hb_vfSize( <pHandle> | <cFileName> [, <lUseDirEntry> ] ) -> <nSize>
The 2-nd parameter <lUseDirEntry> is significant only when
the 1-st one <cFileName> is character value.
* src/rtl/gttrm/gttrm.c
* replaced C++ comments with ANSI ones
* src/rtl/vfile.c
* updated casing of function names in comments
* return F_ERROR instead of NIL hb_vfHandle() when wrong parameter
is given (borrowed from Viktor's branch)
* added C function hb_fileParamGet() borrowed from Viktor's branch
* src/rtl/filesys.c
* variable localization (synced with Viktor's branch)
* src/rtl/disksphb.c
! fixed bug with potentially uninitialized buffer
(fix borrowed from Viktor's branch)
* src/common/hbfsapi.c
* src/rtl/filesys.c
* src/rtl/filebuf.c
* contrib/hbmzip/mzip.c
* contrib/xhb/filestat.c
! if available always use stat64() instead of stat() in *nixes builds.
On 32-bit platforms stat() fails for larger files (>2GB).
It should fix problem reported by Lorenzo.
* src/rtl/filebuf.c
! fixed bad typo inside hb_fileIsLocalName() - result was reverted.
Many thanks to Juan Gálvez for locating the problem.
* src/debug/dbgbrwsr.prg
! fixed typo in previous commit.
It should fix also problem with positioning in debugger source
window reported by Alexey Myronenko.
% minor simplification
* contrib/hbmisc/udpds.prg
! updated to send correct broadcast messages in * nixes and
respect broadcast addresses returned by system for IP interfaces
* src/rdd/hbdbsort.c
* src/rdd/hbsix/sxcompr.c
* pacified warnings
* src/rtl/filebuf.c
! unlock HVM stack before locking local mutex and calling hb_fs*()
functions which also unlocks HVM. It fixes possible deadlock condition
when hb_gcAll( .T. ) is executed.
* include/hbapi.h
* src/vm/garbage.c
! changed mark function semantic.
Adding support for user defined mark function I created race condition
in MT GC code. It happens because blocks marked as deleted were not
scanned by GC mark code so their subitems where not accessible.
To fix it we have to change mark function semantic. Now mark function
can be executed also for blocks currently deleted. It means that GC
block destructor should clean references to just removed items and
subblocks. The best place to make it is clearing pointers to GC blocks
just after hb_itemRelease() or hb_gcRefFree().
It is save to clean the reference just before hb_itemRelease() or
hb_gcRefFree() but only for the single block passed to these functions.
It is not save to clear reference to more then one block and then
execute above functions.
+ check reference count after destructor execution for all blocks
not only arrays - warning it may exploit some wrong C code.
- removed not longer used hb_gcRefCheck() function.
* include/hbvm.h
* src/vm/hvm.c
+ added new internal function hb_vmLockForce()
! fixed to mark GC blocks with active threads
* src/vm/thread.c
* include/hbthread.h
% modified sync mutexes used by xBase++ signal class emulation
to not use item array internally
* updated mutex destructor to new GC mark semantic
! mark GT items if GT is bound with thread item
! fixed mutex notify/subscribe code to not change GC items
when HVM stack is unlocked
! use hb_vmLockForce() to eliminate potential deadlock
+ added new C function hb_threadEnterCriticalSectionGC().
It should be used instead of hb_threadEnterCriticalSection()
in code which manipulates GC items inside critical section.
It's slower but eliminates possible deadlock condition.
Please remember that remember that both functions cannot
be used for the same mutex. So if it's necessary to use
hb_threadEnterCriticalSectionGC() in one place then it
has to be used in all others when the same mutex is locked.
* src/vm/arrays.c
* src/vm/codebloc.c
* src/vm/hashes.c
* src/rtl/hbgtcore.c
* contrib/hbcurl/core.c
* contrib/hbexpat/core.c
* updated destructors to new GC mark semantic
* contrib/hbxpp/dllx.c
! fixed destructor for pointer item created by DllPrepareCall()
Now it respects reference counter.
; added note about real behavior of library handle destructor,
I haven't changed existing code behavior but maybe it should
be done.
* contrib/xhb/hbserv.c
* src/debug/dbgentry.c
* src/rdd/wacore.c
! fixed possible deadlocks by using hb_threadEnterCriticalSectionGC()
instead of hb_threadEnterCriticalSection().
* contrib/hbmemio/memio.c
! slightly modified hb_memfsDirectory() code to avoid race condition
without using hb_threadEnterCriticalSectionGC()
* src/vm/fm.c
% reduce the lock range in HB_FM_STAT builds
* src/rtl/hbsocket.c
! added missing HVM stack unlocking in hb_socketSelect() function
* src/rtl/gtxwc/gtxwc.c
! added few missing locks for version compiled with HB_XWC_XLIB_NEEDLOCKS
* src/rdd/dbtotal.prg
* allow to use symbols instead of codeblocks
* modified Harbour extension which uses ordKey() as default group
signature to work also without index. In such case all records
are summarized into single one.
; Most of above modifications were critical for stability of MT programs.
They should allow to activate GC in any place.
* include/hbapifs.h
* src/rtl/filebuf.c
* src/rtl/filebufd.c
* contrib/hbcomio/comio.c
* contrib/hbmemio/memio.c
* contrib/hbnetio/netiocli.c
* contrib/hbtcpio/tcpio.c
* changed internal FILE IO API. Now all functions which do
not operate on PHB_FILE receive pointer to function table
(PHB_FILE_FUNCS) used by given redirector.
* src/rtl/vfile.c
* open wirtual files in READWRITE mode by default
* translate FO_* attributes to FXO_* ones
* src/vm/classes.c
! fixed typo in RTE message
* src/vm/garbage.c
* use different error code (1301/1302) for destructor RTE
* src/rtl/gtxwc/gtxwc.c
* src/rtl/gtxwc/gtxwc.h
+ added support for HB_GTI_FONTWEIGHT with HB_GTI_FONTW_* actions
* contrib/hbpgsql/rddcopy.c
! fixed memory leak
% encapsulate PQ calls in hb_vmUnlock()/hb_vmLock() to not stop
other threads when GC is activated
* include/hbapifs.h
* src/rtl/filebuf.c
* move HB_FILE_TYPE_MAX definition to header file
% include C stat() header only in *nix builds
* src/pp/ppcore.c
! force stringify when illegal characters are included inside
square brackets []
* src/rdd/hbdbsort.c
* casting
* src/rtl/filesys.c
+ added additional method to extract file time in MS-Windows version of
hb_fsGetFileTime()
* include/hbapifs.h
* src/rtl/vfile.c
+ added C functions to manipulate PRG level hb_vf*() file pointer items
* include/hbapifs.h
* src/rtl/filebuf.c
+ added C functions to create/extract file IO handle from raw file handle
+ added C functions to check if file IO handle points to local file
* contrib/hbmzip/mzip.c
* changed hb_fs*() API to hb_file*() API
! fixed file time setting in *nix builds
* src/vm/set.c
! fixed freed memory access exploited by interactions between
modifications in last few years.
! do not try to remove trailing ^Z char reading from the stream
if FS_END seek reports 0.
! if possible open printer/alternate/extra output in write only mode
to avoid problem on devices where read access is not available.
* src/rtl/langapi.c
* keep EN lang active after deinitialization
* src/rtl/filesys.c
! fixed condition for file time refresh in *nix builds
! fixed file time setting in *nix builds
* include/hbapifs.h
* src/rtl/Makefile
* src/rtl/filebuf.c
+ src/rtl/filebufd.c
* renamed hb_fileRegister2() to hb_fileRegisterFull()
+ added new function hb_fileRegisterPart() which simplifies writing
small FILE IO redirectors.
* contrib/hbmemio/memio.c
* contrib/hbnetio/netiocli.c
* use hb_fileRegisterFull()
* contrib/hbplist.txt
+ contrib/hbtcpio/hbtcpio.hbc
+ contrib/hbtcpio/hbtcpio.hbp
+ contrib/hbtcpio/hbtcpio.hbx
+ contrib/hbtcpio/tcpio.c
+ added new Harbour FILE IO redirector.
It recognizes and process names with "tcp:" prefix, in form like:
tcp:<host>:<port>[:<timeout>]
It can be used in different subsystems using Harbour FILE IO and
stream read/write operations, i.e.
REQUEST HB_TCPIO
SET PRINTER TO tcp:192.168.0.110:9100
can be used to connect to network printers using TCP direct printing.
* utils/hbmk2/po/hbmk2.pt_BR.po
! fixed typo in translation
* src/vm/set.c
! added missing FXO_APPEND flag.
It should fix#54.
* src/vm/asort.c
! respect milliseconds part when timestamp values are compared in ASORT()
Note: ASORT() does not use exact comparison so when mixed date and
timestamp values are used only date is significant. Just like in:
ASort( aValue, {|x, y| x < y } )
* src/rtl/filebuf.c
! fixed C&P typo.
It should fix problem with SET PRINTER TO reported by Rolf.
* src/rtl/hbgtcore.c
! added missing break after HB_GTI_VERSION
* ChangeLog.txt
* include/harbour.hbx
* src/rtl/Makefile
- src/rtl/xfile.c
+ src/rtl/vfile.c
* renamed PRG functions HB_X*() to HB_VF*()
* renamed PRG functions hb_vf[GS]etAttr() to hb_vfAttr[GS]et()
* renamed PRG functions hb_vf[GS]etDateTime() to hb_vfTime[GS]et()
* include/hbapifs.h
* src/rtl/filebuf.c
* src/rtl/fscopy.c
* contrib/hbmemio/memio.c
* contrib/hbnetio/netio.h
* contrib/hbnetio/netiocli.c
* contrib/hbnetio/netiosrv.c
* renamed C functions hb_file[GS]etAttr() to hb_fileAttr[GS]et()
* renamed C functions hb_file[GS]etFileTime() to hb_fileTime[GS]et()
* contrib/hbnetio/netio.h
* contrib/hbnetio/netiocli.c
* contrib/hbnetio/netiosrv.c
+ added support for redirecting hb_vfConfig() to the server.
NETIO does not use it anyhow NETIO works as full redirector
to server system and it may be used on the server side.
* include/hbapifs.h
* src/rtl/filebuf.c
+ added new C function hb_filePOpen()
* include/hbset.h
* src/vm/set.c
* src/rtl/console.c
* changed hb_fs*() API to hb_file*() API in HVM SETs
* src/vm/memvars.c
* changed hb_fs*() API to hb_file*() API in .mem files
* src/vm/runner.c
* changed hb_fs*() API to hb_file*() API in .hrb files
* include/hbapifs.h
* src/rtl/filesys.c
* renamed hb_fsPOpen() param
* contrib/hbhpdf/3rd/libhpdf/hpdffdfj.c
! removed executable attribute
* src/vm/thread.c
! fixed typo reported by Viktor - thanks.
* include/hbapifs.h
* src/rtl/hbproces.c
* src/rtl/spfiles.c
* src/rtl/filesys.c
* src/rtl/file.c
* synced parameter file and dir names used in declarations and
implementations
* include/hbrddcdx.h
+ added few fields used in ADI files
* src/codepage/cp_utf8.c
* src/rtl/cdpapi.c
! always return -1, 0 and 1 from hb_cdpcmp() function.
Workaround for problem exploited by optimized memcmp() function.
* src/rdd/dbfntx/dbfntx1.c
! fixed possible sort problem when optimized memcmp() is used
* added feew missing const declarations
* src/rdd/dbfntx/dbfntx1.c
* src/rdd/dbfnsx/dbfnsx1.c
* src/rdd/dbfntx/dbfntx1.c
! fixed OrdWildSeek() to work with optimized memcmp()
* include/hbapifs.h
* src/rtl/disksphb.c
+ added new C function:
double hb_fsDiskSpace( const char * pszPath, HB_USHORT uiType );
* include/hbapifs.h
* src/rtl/direct.c
+ added new C function:
PHB_ITEM hb_fsDirectory( const char * pszDirSpec,
const char * pszAttributes );
* src/rtl/filesys.c
! report EOF in hb_fsEof() also when file position is out of file size
* updated comment
* include/hbapifs.h
* src/rtl/filebuf.c
+ extended Harbour FILE IO API and added support for read/write filepos,
directory and link operations.
Read/write filepos support in some systems (i.e. *nixes) needs additional
emulation code. If it's not necessary then this functionality can be
disabled by FXO_NOSEEKPOS attribute. Now this attribute is used by our
RDD code.
* src/rtl/Makefile
+ src/rtl/xfile.c
+ added PRG interface to Harbour FILE IO API
The following PRG functions are available for Harbour programmers:
HB_XEXISTS( <cFileName>, [ @<cDestFileName> ] ) -> <lOK>
HB_XERASE( <cFileName> ) -> <nResult>
HB_XRENAME( <cFileSrc>, <cFileDst> ) -> <nResult>
HB_XCOPYFILE( <cFileSrc>, <cFileDst> ) -> <nResult>
HB_XDIREXISTS( <cDirName> ) -> <lExists>
HB_XDIRMAKE( <cDirName> ) -> <nSuccess>
HB_XDIRREMOVE( <cDirName> ) -> <nSuccess>
HB_XDIRECTORY( [ <cDirSpec> ], [ <cAttr> ] ) -> <aDirectory>
HB_XDIRSPACE( <cDirName>, [ <nInfoType> ] ) -> <nFreeSpace>
HB_XGETATTR( <cFileName>, @<nAttr> ) -> <lOK>
HB_XSETATTR( <cFileName>, <nAttr> ) -> <lOK>
HB_XGETDATETIME( <cFileName>, @<tsDateTime> ) -> <lOK>
HB_XSETDATETIME( <cFileName>, <tsDateTime> ) -> <lOK>
HB_XLINK( <cExistingFileName>, <cNewFileName> ) -> <nSuccess>
HB_XLINKSYM( <cTargetFileName>, <cNewFileName> ) -> <nSuccess>
HB_XLINKREAD( <cFileName> ) -> <cDestFileName> | ""
HB_XOPEN( <cFileName>, [ <nMode> ], [ <nAttr> ] ) -> <pHandle> | NIL
HB_XCLOSE( <pHandle> ) -> <lOK>
HB_XLOCK( <pHandle>, <nStart>, <nLen>, [ <nType> ] ) -> <lOK>
HB_XUNLOCK( <pHandle>, <nStart>, <nLen> ) -> <lOK>
HB_XLOCKTEST( <pHandle>, <nStart>, <nLen>, [ <nType> ] ) ;
-> <nPID> | 0 (nolock) | -1 (err)
HB_XREAD( <pHandle>, @<cBuff>, [ <nToRead> ], [ <nTimeOut> ] ) ;
-> <nRead>
HB_XWRITE( <pHandle>, <cBuff>, [ <nToWrite> ], [ <nTimeOut> ] ) ;
-> <nWritten>
HB_XREADAT( <pHandle>, @<cBuff>, [ <nToRead> ], [ <nAtOffset> ] ) ;
-> <nRead>
HB_XWRITEAT( <pHandle>, <cBuff>, [ <nToWrite> ], [ <nAtOffset> ] ) ;
-> <nWritten>
HB_XSEEK( <pHandle>, <nOffset>, [ <nWhence> ] ) -> <nOffset>
HB_XTRUNC( <pHandle>, [ <nAtOffset> ] ) -> <lOK>
HB_XSIZE( <pHandle> ) -> <nSize>
HB_XEOF( <pHandle> ) -> <lEOF>
HB_XFLUSH( <pHandle>, [ <lDirtyOnly> ] ) -> NIL
HB_XCOMMIT( <pHandle> ) -> NIL
HB_XCONFIG( <pHandle>, <nSet>, [ <nParam> ] ) -> <nResult>
HB_XHANDLE( <pHandle> ) -> <nOsHandle>
HB_XTEMPFILE( @<cFileName>, [ <cDir> ], [ <cPrefix> ], [ <cExt> ], ;
[ <nAttr> ] ) -> <pHandle> | NIL
Please remember that not all redirectors have to support all of them.
I.e. MEMIO allows to use directory names in files but does not allow to
create or remove directories. It also does not support file attributes
and times in current implementation though it can be easy added.
HB_XCONFIG() is general function which can be used by redirector
authors to make some special operations which are specific to given
device.
HB_XTEMPFILE() always creates temporary files on local system without
switching to FILE IO redirector. It's intentional behavior.
* src/rtl/fscopy.c
* used Harbour FILE IO API in hb_fsCopy()/HB_FCOPY()
Please remember that HB_FCOPY() always makes operation using current
system when HB_XCOPY() may optimize it moving the operation to low
level subsystem or even other computer i.e. NETIO can do that when
both files are located on the same server.
* src/rdd/delim1.c
* src/rdd/sdf1.c
* src/rdd/dbf1.c
* src/rdd/dbfcdx/dbfcdx1.c
* src/rdd/dbffpt/dbffpt1.c
* src/rdd/dbfnsx/dbfnsx1.c
* src/rdd/dbfntx/dbfntx1.c
* src/rdd/hsx/hsx.c
* open files with FXO_NOSEEKPOS attribute to inform low level API that
read/write filepos is not used (disables unnecessary filepos emulation).
* contrib/hbnetio/netio.h
* contrib/hbnetio/netiocli.c
* contrib/hbnetio/netiosrv.c
+ added support for extended Harbour FILE IO API.
* contrib/hbnetio/utils/hbnetio/netiomgm.hb
! renamed "quit" command to "shutdown" as suggested by Rolf.
* contrib/hbnetio/utils/hbnetio/hbnetio.prg
* allow to interrupt server process using ESC when run in console window.
* contrib/hbmemio/memio.c
+ added support for extended Harbour FILE IO API.
I left two minor TODO notes - maybe someone will try to implement it.
* (all files)
* stripped svn header
* minor cleanups
; use following command to find out the history of files:
git log
git log --follow
git blame
git annotate
