* harbour/include/hbapi.h
* harbour/include/hbcomp.h
+ harbour/include/hbcompdf.h
* harbour/include/hberrors.h
* harbour/include/hbexpra.c
* harbour/include/hbexprb.c
* harbour/include/hbexprc.c
* harbour/include/hbexprop.h
* harbour/include/hbmacro.h
* harbour/include/hbpp.h
* harbour/include/hbvm.h
* harbour/source/common/expropt1.c
* harbour/source/common/expropt2.c
* harbour/source/compiler/Makefile
* harbour/source/compiler/cmdcheck.c
* harbour/source/compiler/complex.c
* harbour/source/compiler/genc.c
* harbour/source/compiler/gencc.c
* harbour/source/compiler/gencli.c
* harbour/source/compiler/gencobj.c
* harbour/source/compiler/genhrb.c
* harbour/source/compiler/genjava.c
* harbour/source/compiler/genobj32.c
* harbour/source/compiler/harbour.c
* harbour/source/compiler/harbour.l
* harbour/source/compiler/harbour.y
* harbour/source/compiler/hbcomp.c
* harbour/source/compiler/hbdead.c
* harbour/source/compiler/hbfix.c
* harbour/source/compiler/hbfunchk.c
* harbour/source/compiler/hbgenerr.c
* harbour/source/compiler/hbident.c
* harbour/source/compiler/hbpcode.c
* harbour/source/compiler/ppcomp.c
* harbour/source/macro/macro.y
* harbour/source/pp/ppcore.c
* harbour/source/pp/ppgen.c
* harbour/source/pp/pplib.c
* harbour/source/vm/cmdarg.c
* harbour/source/vm/estack.c
* harbour/source/vm/fm.c
* harbour/source/vm/macro.c
* harbour/utils/hbpp/hbpp.c
* harbour/utils/hbpp/hbpp.h
* harbour/utils/hbpp/hbppcomp.c
* harbour/utils/hbpp/hbppcore.c
* harbour/utils/hbpp/hbpplib.c
* harbour/utils/hbpp/pragma.c
* changed the internal compiler API to be MT safe.
All global and static non constant variables replaced by
HB_COMP structure which have all compiler context settings.
It's possible to allocate simultaneously many compiler contexts
and compile code. Only constant/read only variables are shared.
In macro compiler HB_COMP is replaced by HB_MACRO.
In source code I everywhere used to macros: HB_COMP_DECL and
HB_COMP_PARAM which are equivalents of old HB_MACRO_DECL and
HB_MACRO_PARAM definitions.
We will only have to change compiler FATAL errors support to
not execute exit() for non batch systems.
* make macro compiler fully MT safe
* removed not longer used definitions and variables
! add direct accessing to PP line number information - it fixes missing
line numbers reported recently
! added generating .ppo files when tokens are teken directly by compiler
- it fixes empty .ppo file problem
! clear hb_stack internall variables after removing hb_stack to avoid
possible GPF if application still works and try to access unexsiting
hb_stack
! do not allocate new memory block when final FM statistic report is
generated - it fixes problem with GPF when application compiled with
FM statistic exits and CLIPPER envvar is set.
! fixed memory leak hb_cmdargCheck()
! added to harbour.y symbol destructors - it should fixes memory leaks
in syntax errors but it uses quite new bison feature which is not
fully supported yet. Unfortunately it's also not MT safe and the
destructors implementation in 1.875c does not respect %parse-param.
Bison documentation says that it should so I hope it will be fixed
soon (or maybe even already is in the newest bison versions) before
I'll add multi context compilation support. If not then I will have
to add some workaround. I can pass compiler context pointer inside
YYSTYPE using hack in a lexer but it will force really huge number
modifications in existing bison rules so probably it will be much
easier to fix bison or write a small tool to update generated parser.
* other modifications - the total size of patch is nearly 1MB and
detail description will have to take me few days.
* harbour/common.mak
* harbour/source/compiler/Makefile
* harbour/include/hbcomp.h
* harbour/include/hbexprb.c
* harbour/include/hbexprc.c
* harbour/source/compiler/cmdcheck.c
* harbour/source/compiler/genc.c
* harbour/source/compiler/harbour.c
* harbour/source/compiler/harbour.l
* harbour/source/compiler/harbour.slx
* harbour/source/compiler/harbour.y
* harbour/source/compiler/hbgenerr.c
* harbour/source/compiler/hbident.c
* harbour/source/compiler/ppcomp.c
+ harbour/source/compiler/complex.c
+ added new PP based compiler lexer - it's smaller, MT safe and a
little bit faster then then the FLEX version.
+ added HB_COMP structure to hold compiler data in future MT version
+ added global variable HB_COMP_PTR hb_comp_data to make conversion
to MT easier - now it holds only PP and lexer data.
* update PP related code in compiler to be MT safe
+ added %pure-parser, %parse-param and %lex-param for bison to generate
MT safe grammar parser.
* updated FLEX to work with recent compiler modifications and pure-parser
bison API
* harbour/makefile.bc
* harbour/makefile.vc
* harbour/source/macro/Makefile
* harbour/source/macro/macro.l
* harbour/source/macro/macro.y
* harbour/source/macro/macrolex.c
* use hb_macro prefix instead of hb_comp in bison/flex parser/lexer
used in macro compiler to avoid possible conflicts in the future
* separated lexer data
* harbour/include/hbapi.h
* harbour/include/hbpp.h
* harbour/source/pp/ppcore.c
* harbour/source/pp/ppgen.c
* harbour/source/pp/pplib.c
* harbour/source/vm/macro.c
* removed not used members from HB_MACRO structure to make it
cleaner before creating common to compiler and macro compiler
structure
+ added new token HB_PP_TOKEN_EPSILON
+ added void * cargo parameters passed to executed user functions
+ hb_pp_tokenGet(), hb_pp_tokenToString(), hb_pp_tokenBlockString()
functions for new PP based compiler lexer
* harbour/utils/hbpp/hbpp.c
* harbour/utils/hbpp/hbpp.h
* harbour/utils/hbpp/hbppcomp.c
* harbour/utils/hbpp/hbppcore.c
* harbour/utils/hbpp/hbpplib.c
* harbour/utils/hbpp/pragma.c
* updated to compile with recent compiler header file modifications
PP, new lexer and most of grammar parser should be MT safe. Now we should
update all compiler functions to pass pointer to HB_COMP data structure
where we should all current global variables. This structure as first
member should have HB_CMPCOMMON structure which will hold common to
compiler and macro compiler data. Ryszard I think you are the best person
to define this structure.
We have new lexer which is MT safe but please note that it has to be
extensively tested so I would like to ask everybody to compile as much
as possible different code and check if the final programs work as
expected. Working on new code I removed some limitations existing in
FLEX though not all. At the beginning I tried to replicate the exact
FLEX behavior but I've found that in few places it does not work as
it should so I begin to encode rules in a way which remove some
limitations. In fact now it's much easier to control some things.
I kept the FLEX code working and made all necessary modifications
so it still can be used but keeping FLEX working cost us IMHO too
much. It's not possible to introduce some improvements to grammar
parser. All identifiers, keyword and macros returned by new lexer
are converted to upper letters, do not have to be freed by hb_xfree()
and is guarantied that will be always accessible. So from grammar file
we can remove all hb_compIdentifierNew( hb_strupr($1), TRUE ) what
should give noticeable speed improvement but will break the FLEX code.
Ryszard and other you will have to decide if we will support FLEX in
the future. We can also clean the code and remove most of other
redundant hb_strupr() and hb_strdup() used in many places. BTW only
one terminal symbol can be returned with lower letters: DOIDENT
and I make it intentionally so it's possible to use:
DO prog1 WITH "sth"
on case sensitive file systems so this symbol should be cloned in
upper cases as function symbol but used without modification as
file name. It's current behavior but I'm not sure you will want
to keep it. Maybe compiler switch to always convert file names
created from
DO <id> [WITH <params,...>]
to lower cases will be better. Please think about it.
* harbour/makefile.bc
* harbour/makefile.vc
* added -Iobj\$(CC_DIRNAME) to CFLAGS as temporary workaround
for recent modifications in core files
This modification can be done in cleaner way but I intentionally
wanted to minimize modifications. I hope Windows developers will
make it better.
* harbour/include/hbexpra.c
* harbour/include/hbexprb.c
* harbour/include/hbexprc.c
* harbour/source/macro/macro.y
* harbour/source/macro/macrolex.c
* do not allocate separate memory in macro compiler for terminal
symbols which needs string representation - it fixes some set
of memory leaks and increase a little bit macro compilation
! fixed accessing freed memory in:
private cMacro := "cEarly", cEarly := {"Early"}
@ 12,10 GET &cMacro[1]
! fixed GPF in:
p:=@left()
and similar expressions
* harbour/source/common/expropt2.c
% minor modification
* harbour/source/compiler/harbour.c
! fixed HB_TRACE message
* harbour/source/pp/ppcore.c
! fixed casting - it resolves problem reported by Alexander with
preprocessing chr(255) from static buffer (PP used by .prg code)
* harbour/source/macro/macro.y
+ added some comments
* harbour/source/macro/Makefile
* commented out compilation of FLEX based lexer for macro compiler and
added compilation of new small lexer written in pure .c
* harbour/common.mak
* replaced macrol.obj by macrolex.obj
+ harbour/source/macro/macrolex.c
+ new small lexer written in pure .c. The code is much smaller,
MT safe and more friendly for grammar parser. I think it should
be also faster but I haven't made any tests yet. When we update
grammar parser to not free string representation of terminal
symbols then we should expect some farther speed improvement,
it will also resolve some set of memory leaks like in:
type("user input")
and will allow to make some parts of grammar parser common to
compiler and macro compiler by removing #if[n]def HB_MACRO_SUPPORT
* harbour/source/common/hbstr.c
! fixed decoding strings without trailing ASCII NUL character in
hb_strRemEscSeq()
* harbour/source/pp/ppcore.c
+ added decoding C escaped strings (e"...") to PP so they can be
preprocessed like a normal strings, it's not enabled when
HB_C52_STRICT macro is set
* harbour/include/hbapi.h
* harbour/source/common/hbstr.c
* harbour/source/compiler/harbour.l
* harbour/source/macro/macro.l
* harbour/source/macro/macro.slx
* harbour/source/compiler/harbour.slx
* changed parameters in function: hb_compStrToNum() now the string
size is passed explicitly so it can work with strings which does
not have trailing ASCII NUL character.
* harbour/source/macro/macro.y
* indenting
* harbour/include/hbmacro.h
* harbour/source/common/hbstr.c
* harbour/source/macro/macro.l
* harbour/source/macro/macro.y
* added two new functions for macro compiler: hb_macroLexNew() and
hb_macroLexDelete() to clearly separate lexer from grammar parser.
* harbour/source/compiler/harbour.l
* harbour/source/common/hbstr.c
* harbour/source/pp/ppcore.c
* added error generation for wrong e"..." strings and fixed \" quoting
* removed \q quoting
* harbour/source/pp/ppcore.c
* harbour/source/pp/ppgen.c
* added small description in headers
* harbour/common.mak
* harbour/source/pp/Makefile
* harbour/source/pp/pplib.c
+ harbour/source/pp/pplib2.c
* harbour/include/hbextern.ch
+ added auto destructor for allocated PP context, it's not longer
necessary to execute __PP_FREE() so this function was moved from
the standard .prg API to file with backward compatible functions
* changed __PREPROCESS() to __PP_PROCESS()
* moved setting std rules to separate file so now if user uses PP
with only his own rules we are not forcing linking static tables
which are quite huge (over 100kb)
To include stdandard rules in static binaries user should add
to his code:
REQUEST __PP_STDRULES
+ harbour/source/pp/pplib3.c
* old PP functions for backward compatibility:
__PPADDRULE(), __PREPROCESS(), __PP_FREE()
Please note that using this function is not MT safe. They should
work like with old PP code. Using any of this function automatically
forces linking __PP_STDRULES
* harbour/include/hbapi.h
+ added
void * pLex;
member to HB_MACRO structure - it will be necessary for MT safe
macro parser
* harbour/include/set.ch
* harbour/include/hbexprb.c
* harbour/source/macro/macro.l
* harbour/source/macro/macro.y
* indenting, formatting and converting tabs to spaces
* harbour/include/hbpp.h
* harbour/source/compiler/ppcomp.c
* harbour/source/pp/ppcore.c
* harbour/source/pp/ppgen.c
* harbour/source/pp/pplib.c
* make HB_PP_TOKEN structure public for other code and changed
public PP API to not force including static rules by programs
which does not need them.
+ added basic API for using PP as lexer
* divided HB_PP_TOKEN_MACRO to HB_PP_TOKEN_MACROVAR and
HB_PP_TOKEN_MACROTEXT and mark
[<keyword>][&<keyword>[.[<nextidchars>]]]+
as HB_PP_TOKEN_MACROTEXT not pair HB_PP_TOKEN_KEYWORD and
HB_PP_TOKEN_MACRO
It allows easier integration PP as (macro)compiler lexer and
fixes preprocessing [<keyword>][&<keyword>[.[<nextidchars>]]]+
in some patterns.
* harbour/utils/hbtest/Makefile
+ added PP library to linked library list
* harbour/source/macro/macro.y
+ added seaport to use PP as lexer. It's disabled by default and
can be enabled using by defining HB_PP_MACROLEX. It eliminates
using FLEX or SIMPLEX lexer in macro compiler. Now it emulates
the behavior of FLEX lexer because I wanted to minimize the
modification but in fact because parsed token are freed when
macro compiler finish its job then we can resolve RT memory
leaks which apear during parsing wrong expressions. It will be
enough to drop allocating memory for new terminal symbols and
using them as is like in compiler. It will also quite nice
simplify include/hbexpr*.c code and will allow to move some
functions to COMMON library. I want to make also pure compiler
code MT safe so I will have to add passing pointer to a structure
with compiler context just like in macro compiler. It means that
if we define a structure which will be common to compiler and macro
compiler with local (not common) data stored in other structure for
which will will keep only a pointer then we will be able to move most
of include/hbexpr*.c code to common library. But I do not want to
make farther modification without consultation without Ryszard.
Also to make PP default lexer some farther modifications should
be done in core PP code and include/hbexpr*.c files to not reduce
the speed. Now using HB_PP_MACROLEX for sure will have to cause
speed reduction in macro compiler. But if we make necessary
modifications then we should keep at least the same speed or
probably reach noticeable better results.
* harbour/TODO
* removed to resolved problems from TODO/TOFIX list
* harbour/include/hbexprb.c
* indenting and tab to spaces conversion
* harbour/source/compiler/ppcomp.c
! fixed typo which cased that dump buffer was not ended
with \0 character
* harbour/include/hbpp.h
* harbour/source/pp/ppcore.c
! fixed the problem reported by Enrico in code like:
IF ["] $ var
Unfortuantelly this fix also breaks some valid expressions like:
x := a[ f("]") ] $ "test"
x := a[ f( "'" ) ] $ "test"
what is also Clipper compatible anyhow I think that in the future
we may want to break Clipper compatibility and allow compilation
of such code.
* harbour/doc/pp.txt
* updated documentation for the above modification
* harbour/bin/hb-func.sh
+ pass GTK paths to GCC in hb* scripts when -xbgtk switch is used
* harbour/contrib/rdd_ads/rddads.h
* harbour/contrib/rdd_ads/adsfunc.c
* indenting and casting
* harbour/source/rdd/dbfcdx/dbfcdx1.c
* cleaned typos in comments
+ harbour/tests/speedtst.prg
+ added speed test
* harbour/include/hbpp.h
* harbour/source/pp/ppcore.c
+ added HB_PP_TOKEN_ISNEUTRAL() macro to make code a little bit more
readable
* harbour/include/std.ch
+ added some missing old PP commands
* harbour/source/Makefile
* minor modification in order of compiled libraries
* harbour/source/pp/pplib.c
* changed the PP interface for .prg code.
Now following functions are supported:
__PP_INIT( [<cIncludePath>] [, <cStdChFile> ] ) -> <pPP>
initialize new PP context and return pointer to it.
when <cStdChFile> is empty string ("") then no default rules are
used only the dynamically created #defines like __HARBOUR__,
__DATE__, __TIME__, __PLATFORM__*
__PP_FREE( <pPP> )
free PP context.
__PP_PATH( <pPP>, <cPath> [, <lClearPrev>] )
add new (or replace previous) include paths.
__PP_RESET( <pPP> )
reset the PP context (remove all rules added by user or
preprocessed code)
__PP_ADDRULE( <pPP>, <cDirective> )
preprocess and execute new preprocessor directive
__PREPROCESS( <pPP>, <cCode> ) -> <cPreprocessedCode>
preprocess given code and return result
User can create more then one PP context and then use each of them
separately. Any modification in one context have no effect in other.
* harbour/contrib/rdd_ads/rddads.h
+ added #define WIN32 when HB_OS_WIN_32 is set to not force -DWIN32
compiler switch
* harbour/source/pp/ppcore.c
* cleaned BCC warnings
* harbour/source/rdd/dbfcdx/dbfcdx1.c
! fixed ordKeyGoto() positioning when active filter is set and no
controlling order
* harbour/include/hbcomp.h
* harbour/include/hbpp.h
* harbour/source/compiler/cmdcheck.c
* harbour/source/compiler/hbusage.c
* harbour/source/compiler/ppcomp.c
* harbour/source/pp/ppcore.c
* harbour/source/pp/ppgen.c
* harbour/source/pp/pplib.c
* restored support for old hb_inLine() syntax
+ added support for hb_inLine() to new PP. It can be enabled by
-kI compiler switch. It's disabled by default.
The new hb_inLine{} syntax is:
[ <exp> ] hb_inLine [ ( [<params,...>] ) ] { [ <c_code> ] } [ <exp> ]
<c_code> can contain new line characters.
Nested hb_inLine{} in <params,...> is not supported. If you think
it's important I can add it in few lines.
This version does not break any valid Clipper syntax, hb_inLine is
not reserved word and hb_inLine can be repeated many times in the
same line. F.e. this code can be properly preprocessed and compied if
you disable in compiler old hb_inLine() syntax and enable the new one
by -kcI Harbour compiler switch:
proc main()
local hb_inLine := " (var) "
? hb_inLine{ hb_retc("inLine"); } + hb_inLine(" parameter ") {
hb_retc( hb_parc( 1 ) );
} + "!" + hb_inLine + hb_inLine() { hb_retc( ":-)" ); } + ;
hb_inLine() + "{}"
return
function hb_inLine()
return " func() "
So from user point of view this version have real "inline" syntax.
To be clear: I'm not a fun of any C inline extensions. They works
only when we use .c code as compiler backend so the code which uses
them cannot be used in .hrb files and any other format we will add
in the future which do not support later C code compilation. F.e it
will not work in compiler integrated with HVM or with .NET
Anyhow if someone finds it useful then I think that it should be
properly implemented. This is the reason I added the new version.
The old one I left only for backward compatibility. It breaks any
code which uses hb_inLine keyword can be used only once in a line
and it does not have "inline" so IMHO it should not be enabled by
default.
I would like to hear your opinion about default compiler switches:
1. should we disable both: hb_inLine() and hb_inLine {}
2. should we disable hb_inLine() and enable hb_inLine {}
3. should we disable hb_inLine {} and enable hb_inLine()
4. should we enable both: hb_inLine() and hb_inLine {}
+ harbour/tests/hbinline.prg
+ added test code for new hb_inLine{} syntax
* harbour/tests/inline_c.prg
! replaced direct access to HB_ITEM body by API function calls
* harbour/makefile.bc
* Adapted to the new Preprocessor utility ppgen.exe
* Cleaned hbpp.exe dependecies due to recent changes
* harbour/makefile.vc
* harbour/common.mak
* Cleaned hbpp.exe dependecies due to recent changes
* harbour/include/hbapi.h
* Added HB_EXPORT to hb_gcAlloc()
* harbour/include/hbstack.h
* Added HB_EXPORT to hb_stackSelfItem()
* harbour/source/vm/fm.c
* Added dipslaying contents of unreleased blocks at
program exit, together with logging unreleased memory
blocks report to fm.log file. Borrowed from xharbour.
* harbour/bin/hb-func.sh
+ added linker parametrs to hbcc
* harbour/include/hbapi.h
+ added hb_retclenAdoptRaw()
* harbour/include/hbapifs.h
* harbour/source/common/hbfsapi.c
* changed first parameters of hb_fsAddSearchPath() to const char *
* harbour/include/hbclass.ch
* update for new PP. I think that now we can try to create final
version of our OOP rules.
* harbour/include/hbdefs.h
+ added UCHAR and SCHAR typedefs
* harbour/include/hbrddcdx.h
* indenting
* harbour/source/rdd/workarea.c
! fixed minor typo
* harbour/source/common/hbdate.c
* harbour/source/rtl/dates.c
* moved hb_dateToday() and hb_dateTimeStr() from RTL to COMMON library
- harbour/include/hbpp.h
- harbour/source/pp/ppcomp.c
- harbour/source/pp/ppcore.c
- harbour/source/pp/pplib.c
- harbour/source/pp/pptable.c
- harbour/source/pp/pragma.c
* harbour/utils/hbpp/Makefile
* harbour/utils/hbpp/hbpp.c
+ harbour/utils/hbpp/hbpp.h
+ harbour/utils/hbpp/ppcomp.c
+ harbour/utils/hbpp/ppcore.c
+ harbour/utils/hbpp/pplib.c
+ harbour/utils/hbpp/pptable.c
+ harbour/utils/hbpp/pragma.c
* moved all PP code to harbour/utils/hbpp
It's interesting and working preprocessor and the code can be
usable for some other things so I do not want to remove it.
Probablly we should move it to contrib/hbpptext
* harbour/include/hbcomp.h
* harbour/include/hberrors.h
* harbour/include/hbsetup.ch
+ harbour/include/hbpp.h
+ harbour/include/hbstdgen.ch
* harbour/source/common/hbstr.c
* harbour/source/compiler/Makefile
* harbour/source/compiler/cmdcheck.c
* harbour/source/compiler/harbour.c
* harbour/source/compiler/harbour.l
* harbour/source/compiler/hbgenerr.c
+ harbour/source/compiler/ppcomp.c
* harbour/source/pp/Makefile
+ harbour/source/pp/ppcore.c
+ harbour/source/pp/pplib.c
+ harbour/source/pp/ppgen.c
+ New PP code written from scratch. It works in similar way to
Clipper PP even the error codes are replicated. The code is MT
safe does not have any limitation on size of preprocessed code,
line, etc. It's also Clipper compatible lexer. It means that
we do not longer need FLEX or SIMPLEX which can be replaced
by new PP after some small modifications. Anyhow I haven't
decided to make it myself. I would like to agree with with the
rest of developers. I will be very happy if such modifications
will be done by someone else, Ryszard?
Meanwhile I current PP join on output the line tokens and give
the string line to FLEX/SIMPLEX. It does not have any sense and
all FLEX/SIMPLEX limitations are still existing. Ryszard, even
if we keep it then I hope you can remove at least FLEX line buffer
and use the one returned from PP.
Because string tokens in parsed line are converted to text which is
later once again decoded to tokens by FLEX I had to introduce new
string format which supports embedded string delimiters. I chose
modified version of xHarbour extension with escaped strings e"<sting>"
so now FLEX understand such strings and decode them like C escaped
strings. It means that you can use them also in the .prg code f.e.:
outstd( e"Hello\n\rWorld" )
The new PP is also noticeable faster. You should see the difference
compiling long files. The build in PP rules are generated automatically
by ppgen program created from source/pp/ppgen.c
I had to add to GNU source/pp/Makefile these two lines:
pptable.c : ppgen$(EXE_EXT)
./ppgen$(EXE_EXT) $(TOP)$(ROOT)include/hbstdgen.ch -opptable.c -q
Sth like that will have to be done also in non GNU make system.
Now Harbour can be compiled only using GNU make.
Marek can you update non GNU make files? I would like to leave this
modification to you or other developers who can test it.
It was quite big modification and I do not believe that I haven't
make any mistakes but I hope that in few weeks I'll fix any reported
bugs and it will resolve any PP problems.
TODO:
* error messages
create one common list of errors and warnings and keep it
in common library. PP and compiler can still generate different
errors with the same number. It can be confusing for the users
and hard to document and add i18n translations.
If possible we should also try to keep Clipper error numbers.
In new PP code I added Clipper error numbers but I cannot use
them until compiler code is not updated.
We should aslo remove the ctype passed to error functions and
hack with first character in warning messages and use only
error number. The codes from 1000 to 1999 should be warnings
where range 1000:1099 is activated by -w, 1100:1199 by -w1,
1200:1299 by -w2, etc. 2000:2999 are errors and 3000:3999
fatal errors. All compiler functions which generate an error
should expect that error function will not stop the compiler
but return and cleanly finished their job. It's necessary for
MT support in compiler and making compiler part of some other
programs which may still work and compile different source code.
* FLEX/SIMPLEX
remove them at all and add some final pass to PP to create
more precise tokens for grammar parser or at least add better
integration to remove some redundant code and existing limits.
* hb_inLine() support - it's broken in new PP but as I can see
it was never working correctly. I can add a hack to PP to support
hb_inLine() but I'm not sure it's worth to do. Maybe in few days.
+ harbour/include/std.ch
+ added new std.ch. It was created without using Clipper's std.ch.
It's quite possible that some rules are wrong and should be fixed
so please help. Anyhow I created a set of programs based on new PP
code generating all possible combinations of different commands I
collected from different source code, documentation and match patterns
of rules I was adding to new std.ch and then I was comparing .ppo
files generated by Clipper and Harbour so I do not expect any bigger
problems then some minor typos. I had to introduce some modifications
in spacing as workaround for FLEX/SIMPLEX which cannot properly decode
text preprocessed by Clipper after stringify.
* harbour/utils/hbpptest/pretest.prg
* updated for new PP
It reports 16 wrong translations but some of them are valid
and some others are caused by escaped string e"..." (one of hack
for FLEX/SIMPLEX support)
I also had to remove all spaces in comparison the results
because they were not Clipper ocmpatible at all. Ryszard please
look at it.
* harbour/makefile.bc
* Updated comment section about environment variables used by makefile
* harbour/makefile.vc
* Updated comment section about environment variables used by makefile
* Cleaned compilation under Visual C++ 8.0 (VS2005 compiler), including
64 bit platforms. To properly compile under Visual C++ 8.0 you have
set a new environment variable HB_VISUALC_VER to 80 before running
make_vc.bat.
* harbour/include/hbclass.ch
+ added validation for class data names. As additional preprocessor
rule to not use <!marker!>
* harbour/include/hboo.ch
* harbour/source/rtl/tclass.prg
* harbour/source/vm/classes.c
+ added delegate messages
* harbour/include/hbstack.h
* harbour/source/vm/arrays.c
* harbour/source/vm/estack.c
* harbour/source/vm/hvm.c
* harbour/source/vm/proc.c
+ added startup symbol to hb_stack. It gives very good stop
condition for all procedures which trace stack calls and
resolves the problem with hb_stack.pBase which was never
pointing to valid function/procedure symbol when no symbol
was put on HVM stack. Now after hb_stackInit() the first
item is allocated for "hb_stackInit()" symbol so hb_stack.pBase
is always a pointer to valid function/procedure symbol.
* changed the guard condition for buggy code in hb_stackPop()
and similar code from:
hb_stack.pPos < hb_stack.pItems
to:
hb_stack.pPos <= hb_stack.pBase
The old condition was generating usable error message only in the
startup function. In deeply called functions it was only waste of
CPU time on one of the most often call functions. Before it was
activated internal stack structures were corrupted.
If we were leaving for many years without really working stack
underflow protection then maybe we should think about disabling
it at when HB_STACK_MACROS are used and keep it only in real
stack functions for HVM developers. Single protection in
hb_stackOldFrame() I've just added is enough for basic C code
validation.
I created the second version of stack macros without it in
hbstack.h but I would like to hear other developers opinion.
* harbour/include/hbexprc.c
* Temporary disabled optimization with references to object variables
until we will not have extended reference items in our HVM.
It can be enabled by compiling with HB_USE_OBJMSG_REF
* harbour/include/hbclass.h
* restored previous CLASSDATA behavior which ignores SHARED clause
and always create shared class variables
* harbour/include/hbexprc.c
* translate HB_P_INC / HB_P_DEC in (pre|post)(inc|dec)rementation
to HB_P_[PLUS|MINUS]EQ
* harbour/source/vm/hvm.c
- removed not necessary now hb_itemUnRef() in hb_vmInc()/hb_vmDec()
* harbour/include/hbclass.ch
* most of the rules rewritten
! fixed some wrong rules and general cleanup
+ added additional code validation
! fixed compilation of more then one class in single file.
Now it's even possible to declare all classes at beginning of
single file and then implementing their methods in any order
! fixed using static classes and classes
! fixed compilation without <ClassName>_ prefix in method names
+ added support for HB_CLS_NO_DECORATION macro which disable
adding <ClassName>_ prefix to method names - this macro is
set by default when HB_SHORTNAMES is set.
+ added support for declared parameters validation - it can be
disabled with HB_CLS_NO_PARAMS_ERR and I had to disable it
by default due to problems with our preprocessor.
Ryszard seems that our PP has serious problems with decoding
directives when there is no space between symbol and some other
non symbol character. I had to add some workarounds and even
introduce buggy rules to make it working. Please look at it.
You can remove #define HB_CLS_NO_PARAMS_ERR from hbclass.ch
and try to rebuild Harbour core code to see the problem.
* harbour/include/hboo.ch
* harbour/source/vm/classes.c
+ added support for new primitive message: HB_OO_MSG_PERFORM
* harbour/source/rtl/tclass.prg
- removed <lPersistent> parameter from HBClass messages and
internals data. Persistent is supported as scope bit and
separate variable was redundant.
- removed stripping of () from message names. Here is not a place
to fix wrong preprocessor rules.
* harbour/utils/hbtest/rt_class.prg
* use: METHOD PROCEDURE ... CALSS ...
instead of: PROCEDURE ... CALSS ...
The first version is preferable syntax.
* harbour/source/debug/dbgtmenu.prg
* harbour/source/rtl/checkbox.prg
! fixed some parameters in method declaration - global cleanup
will have to wait for preprocessor fixes
Hi all,
Please make test with current hbclass.ch code.
I hope that I haven't broken too much things ;-) but I rewrite
from scratch most rules and it's possible that I missed sth or
made some stupid typos. Current version is much shorter and should
be easier to updated. For sure I've intentionally changed one thing.
CLASSDATA was ignoring SHARED attribute and always created shared
class variables. Seems that it was long existing typo but the fix
may interact with already existing code which needs SHARED class
variables but does not use SHARED clause in CLASSDATA declaration.
In such case please update it and add missing SHARED.
Also in the end of CLASS declaration we have:
[ ; #translate Super( <SuperClassN> ): => ::<SuperClassN>: ] ;
[ ; #translate Super( <SuperClass1> ): => ::<SuperClass1>: ] ;
[ ; #translate Super(): => ::<SuperClass1>: ] ;
[ ; #translate Super: => ::<SuperClass1>: ] ;
[ ; #translate ::Super : => ::<SuperClass1>: ]
These rules introduce very serious bug - they are breaking supercasting
in code which makes sth like:
::super:super:super:msg
or in any other code which sends SUPER message to some other class
objects. I will have to remove them. At least the last three ones.
There were some other things I wanted to write about but it's too late
and I'm to tired - sorry. If you will have any question please ask. if
you will notice some problems with current rules please inform me.
* harbour/source/vm/classes.c
* allow to access object variables using assign messages without
parameter. I do not want to add symmetric behavior for access messages
though it's technically possible but such things should be controlled
by programmer because access messages can have less restricted scoping
then assign ones (READONLY) so I'll add new rule for it in hbclass.ch
* harbour/source/vm/hvm.c
* fixed warning message when object variable reference is
generated using macro compiler without '_' as prefix.
* harbour/contrib/rdd_ads/ads1.c
* harbour/contrib/rdd_ads/adsfunc.c
* synced with Brian fixes in xHarbour
* harbour/source/rdd/workarea.c
* minor modification
* harbour/source/rdd/dbfcdx/dbfcdx1.c
* minor Clipper compatibility update
eg:
o:&send()
o:&send.end()
o:&(expr)()
o:&var++
o:&var := 0
o:&(expr)++
WITH OBJECT obj
++:&var
:&var += :&(var2+"oo")
END
NOTICE:
For simple assigments (=,:=), compound assignments (+=,-=,*=,/=)
and for pre/post increment operators( ++,--) the macro should
evaluate to a symbol that starts with underscore '_'.
To access a variable the macro should evaluate to a symbol
with no '_' char.
* harbour/include/common.ch
- removed ISSYMBOL()
* harbour/source/rtl/tclass.prg
* use valtype() instead of IS*()
* some minor modifications
* harbour/source/vm/hvm.c
* minor modification
* harbour/include/hbclass.ch
* updated some not enabled by default rules to use function pointers
instead of function names which I forgot to update in previous
commit
* harbour/include/common.ch
+ added ISSYMBOL()
* harbour/include/hbextern.ch
- removed __CLS_PARAM()
* harbour/include/hbclass.ch
* harbour/source/rtl/tclass.prg
* harbour/source/vm/classes.c
* replaced __CLS_PAR00() and __CLS_PARAM() functions by preprocessor
rules. These functions are not longer necessary though I left them
in classes.c for backward binary compatibility. Probably they will
be removed in the future.
! use function pointers (@<funcname>()) instead of function names
with REQUEST for super classes ID. It fixes using STATIC class
functions.
* harbour/make_rpm.sh
* check for rpmbuild also when BUGGY_RPM is set
* harbour/make_tgz.sh
* removed rebuild hbpptest with shared libraries
* harbour/include/hbapicls.h
+ added definitions for hb_clsCreate(), hb_clsAdd(), hb_clsAssociate()
* harbour/include/hbapierr.h
* harbour/source/rtl/errorapi.c
+ added hb_errFuncName - pseudo function name which can be used as
operation description in new error object. It should help to simplify
some code and fix function names when RT error is generated from C
function which can be executed from different .prg functions.
* harbour/include/hbclass.ch
+ added some missing Classy(y) compatible object variable type
descriptions: Int, Integer, Block
* harbour/source/rdd/dbfcdx/dbfcdx1.c
! disabled early retrieving from the cache key position and key count
* harbour/include/hbvm.h
* harbour/source/vm/hvm.c
+ added: void hb_vmPushEvalSym( void ) for the code which do not
want to access any [P]HB_SYMB structures
* changed symbol name in hb_symEval from __EVAL to EVAL and modified
hb_symEval registration method to not register hb_vmDoBlock as global
function. Having PHB_DYNS address for "EVAL" in hb_symEval.pDynSym
allows to eliminate some strcmp() when we are looking for EVAL public
symbol.
! fixed symbol table updating in PCODE functions for future multi module
.hrb files
* harbour/source/vm/proc.c
* use hb_symEval.pDynSym instead of strcmp()
* harbour/source/vm/arrays.c
* harbour/source/vm/debug.c
* minor code cleanup
* harbour/source/rtl/tclass.prg
* harbour/source/vm/classes.c
% separate message hash table from methods table - it will cause farther
reduction of allocated memory - many thanks to all people who send
me statistics about number of used classes and symbols from their code.
% use hb_symEval.pDynSym
+ added support for object variable type scope checking like in Class(y).
! fixed typo in pushing object class variables by reference
% s_pClasses is now indexed directly by class handle
% some other optimiztions
! make hidden and non virtual invisible in child classes if it overloaded
some other non hidden method. In such case the overloaded method is used
and can be accessed.
! make friend functions real method class oriented not (super)object for
hidden messages
* include/hbexprb.c
* source/compiler/exproptb.c
* source/compiler/harbour.y
*added support for passing the reference to object's variable
e.g.
foo( @obj:var )
* harbour/source/vm/itemapi.c
* minor modification in function order to group together the ones
which can operate on reference counters
+ added additional stop condition in hb_itemUnRef() - we will have
to make some modifications for the code which execute hb_itemUnRef()
to resolve some possible conflicts with operators overloading and
compiler optimizations
* harbour/source/vm/classes.c
* do not inherit friends - seems that our classes code was too
friendly ;-)
It also effectively resolves logical problem with accessing
overloaded non virtual methods from friends.
+ added __clsLock( <hClass> ) - it locks any farther class
modifications and can be used if programmer wants to be sure
that no one will try to hack his class definition. It's very
important when programmers work in a team on bigger project.
Please do not add to CVS any __clsUnLock() functions - if someone
needs it then he can create his own custom HVM build.
* code cleanup
* harbour/source/vm/eval.c
+ added hb_execFromArray() - idea and syntax borrowed from xHarbour's
HB_ExecFromArray() by Giancarlo Niccolai though it's independent
implementation. Now when we have support for functions with variable
number of parameters then we need sth like this function to fully
benefit from new feature. The following syntax is supported:
hb_execFromArray( <cFuncName> [, <aParams> ] )
hb_execFromArray( @<funcName>() [, <aParams> ] )
hb_execFromArray( <bCodeBlock> [, <aParams> ] )
hb_execFromArray( <oObject> , <cMethodName> [, <aParams> ] )
hb_execFromArray( <oObject> , @<msgName>() [, <aParams> ] )
or:
hb_execFromArray( <aExecArray> )
where <aExecArray> is in one of the following format:
{ <cFuncName> [, <params,...>] }
{ @<funcName>() [, <params,...>] }
{ <bCodeBlock> [, <params,...>] }
{ <oObject> , <cMethodName> [, <params,...>] }
{ <oObject> , @<msgName>() [, <params,...>] }
* harbour/source/rtl/symbol.prg
* removed C code inside #pragma BEGINDUMP/ENDDUMP
now it's only .prg code which uses variable parameters function
and hb_execFromArray()
