* contrib/hbwhat32/whthead.c
* contrib/hbwhat32/whtmmcap.c
* contrib/hbwhat32/whtshell.c
* contrib/hbwhat32/whtdir.c
! Fixed a few problems when compiling with MinGW.
* contrib/Makefile
* contrib/make_b32_all.bat
* contrib/make_gcc_all.sh
* contrib/make_vc_all.bat
+ Readded hbwhat32 to default build lists.
; This doesn't mean this lib works, but the amount of
warnings are now on a manageable level on both 32 and
64 bit on BCC/MSVC/MINGW, we can gradually fix the
remaining ones, and maybe if we have this lib
built, someone will try to use it and maybe fix it.
[ We may temply as well disable it again, if it turns
out to be a showstopper for 1.0.1 release. ]
; If there is anything against the above, I'd suggest
to remove the whole contrib altogether.
+ contrib/hbwhat32/hbwhat.h
* contrib/hbwhat32/whtbmp.c
* contrib/hbwhat32/whtbrsh.c
* contrib/hbwhat32/whtcall.c
* contrib/hbwhat32/whtcdlg.c
* contrib/hbwhat32/whtclpb.c
* contrib/hbwhat32/whtcomm.c
* contrib/hbwhat32/whtcret.c
* contrib/hbwhat32/whtdate.c
* contrib/hbwhat32/whtdc.c
* contrib/hbwhat32/whtdir.c
* contrib/hbwhat32/whtdlg.c
* contrib/hbwhat32/whtdll.c
* contrib/hbwhat32/whtdraw.c
* contrib/hbwhat32/whtfont.c
* contrib/hbwhat32/whtgdi.c
* contrib/hbwhat32/whthead.c
* contrib/hbwhat32/whticon.c
* contrib/hbwhat32/whtilst.c
* contrib/hbwhat32/whtinet.c
* contrib/hbwhat32/whtini.c
* contrib/hbwhat32/whtkbrd.c
* contrib/hbwhat32/whtlv.c
* contrib/hbwhat32/whtmain.c
* contrib/hbwhat32/whtmapi.c
* contrib/hbwhat32/whtmem.c
* contrib/hbwhat32/whtmenu.c
* contrib/hbwhat32/whtmeta.c
* contrib/hbwhat32/whtmisc.c
* contrib/hbwhat32/whtmmcap.c
* contrib/hbwhat32/whtmous.c
* contrib/hbwhat32/whtmsg.c
* contrib/hbwhat32/whtpen.c
* contrib/hbwhat32/whtprn.c
* contrib/hbwhat32/whtrect.c
* contrib/hbwhat32/whtreg.c
* contrib/hbwhat32/whtrgn.c
* contrib/hbwhat32/whtscrlb.c
* contrib/hbwhat32/whtseria.c
* contrib/hbwhat32/whtshell.c
* contrib/hbwhat32/whtsock.c
* contrib/hbwhat32/whtsys.c
* contrib/hbwhat32/whttab.c
* contrib/hbwhat32/whttbar.c
* contrib/hbwhat32/whttext.c
* contrib/hbwhat32/whttree.c
* contrib/hbwhat32/whtview.c
* contrib/hbwhat32/whtwnd.c
* contrib/hbwhat32/wincorec.c
! Changed pointer related hb_ret, hb_par and hb_stor calls
to ptr versions. This might create incompatibilities and
since there is no test package for this lib, it's difficult
to test if it works. I relied on MSVC warnings.
; There are still some calls to change. One other problem is
there is no public equivalent of hb_vmPushLong() for pointers,
so these VM specific parts should be changed to use some more
official Harbour APIs.
* bin/bld.bat
* Some changes to make it more up-to-date.
* utils/hbtest/hbtest.prg
* utils/hbtest/rt_array.prg
+ Added comment to 'TAEVSM' test case saying this is
a CA-Cl*pper bug.
+ Added two additional tests doing resizes in ASORT()
blocks. They fail very badly in Harbour. (TOFIX)
; Made my indenting of ChangeLog entries compatible with
everyone else's.
* harbour/source/vm/hashes.c
! fixed typo in hash cloning code which caused that references
was copied to cloned array. Thanks to Mindaugas for example.
* contrib/mtpl_b32.mak
* contrib/mtpl_vc.mak
* contrib/mtpl_gcc.mak
* Changed the location of intermediate obj and c files from
/obj/<arch>/ to local [/contrib/<contribname>/]obj/<arch>/
dir so that similar filenames across different contribs
would never interfere with each other anymore when using
non-GNU make. This was a difficult to detect problem,
especially when someone is trying to use custom contribs,
and even more so if these custom contribs happen to have
3rd party code with non Harbour controlled filenames. This
change is also an important last step to make contribs
fully self-containing.
; TODO: mtpl_gcc.mak support has to be finished by someone
familiar with bash.
; NOTE: This change shouldn't go to 1.0.1 until well tested
and reviewed.
; NOTE: As a consequence of the above, from now on a
'make_*.bat clean' won't anymore clean the objects/.c
file of all contribs. If you hadn't so so far, now you'll
need to explicitly call 'contrib/make_*.all.bat clean'
to do that.
* contrib/mtpl_gcc.mak
* make_gcc.mak
% HB_ARCHITECTURE hack from cyg to w32 removed,
as cyg is not a valid public architecture name
in Harbour. cyg is used in make_gcc.sh internally, and
it's automatically converted to w32 before using
make_gcc.mak, so it would seem that this is solved
at the right place already, and the hack is not needed.
* contrib/hbbmcdx/bmdbfcdx.c
! fixed to not respect bitmap filters when structural order
is scanned, f.e. in OrdListAdd() with active bitmap filter.
* source/rtl/hbinet.c
! fixed return value of HB_INETGETSNDBUFSIZE
* source/common/hbver.c
! Version detection fixed for __DMC__
[TOMERGE 1.0.1]
* contrib/mtpl_gcc.mak
% One unnecessary C -I option deleted.
[TOMERGE 1.0.1]
- contrib/hbsqlit2
+ contrib/examples/hbsqlit2
* doc/whatsnew.txt
* Moved hbsqlit2 to examples.
[TOMERGE 1.0.1]
* contrib/examples/hbsqlit2/Makefile
* contrib/examples/hbsqlit2/make_b32.bat
* contrib/examples/hbsqlit2/make_gcc.sh
* contrib/examples/hbsqlit2/make_vc.bat
* contrib/examples/hbsqlit2/common.mak
* contrib/examples/hbsqlit2/tests/bld_b32.bat
* contrib/examples/hbsqlit2/tests/bld_vc.bat
* Modified make files to work from the new location.
* Location of sqlite2 headers can be specified by HB_INC_SQLITE2.
; Notice that sqlite2 lib itself should now be build separately.
[TOMERGE 1.0.1]
* harbour/source/rdd/dbfntx/dbfntx1.c
* harbour/source/rdd/dbfcdx/dbfcdx1.c
! fixed dirty index reading, thx Miguel for bug report
* harbour/source/compiler/hbmain.c
* removed old comment and hack
* make_vc.mak
* contrib/mtpl_vc.mak
+ -w34701 VC flag added to enabled warning 4701 (uninitialized vars) on -W3 level.
% -FD VC flag removed. This is only needed for IDEs.
% -Gs VC flag moved to the common section.
% VC flags synced in contrib with core.
+ TOFIX added for WinCE command line. It's full of stuff which is _probably_ not needed.
+ NOTE added about MSDN page explaining VC switches.
; NOTE: -W4 and -Wp64 are recommended for MSVC.
; TOFIX: There are still a few uninit vars warnings:
harboury.c(7226) : warning C4701: potentially uninitialized local variable 'hb_complval' used
macroy.c(3095) : warning C4701: potentially uninitialized local variable 'hb_macrolval' used
dbffpt1.c(2671) : warning C4701: potentially uninitialized local variable 'ulType' used
dbffpt1.c(2614) : warning C4701: potentially uninitialized local variable 'ulSize' used
dbffpt1.c(3073) : warning C4701: potentially uninitialized local variable 'ulOldSize' used
dbffpt1.c(3083) : warning C4701: potentially uninitialized local variable 'ulBlock' used
dbffpt1.c(3652) : warning C4701: potentially uninitialized local variable 'pBlock' used
* source/rdd/workarea.c
! Fixed DBI_POSITIONED value returned.
* include/hbsetup.h
* source/vm/pvalue.c
* source/rtl/fserror.c
* source/rtl/accept.c
* source/rtl/descend.c
* source/rtl/inkey.c
* source/rtl/dates.c
* source/rtl/fkmax.c
* source/rtl/natmsg.c
* source/rtl/hbinet.c
* source/rtl/dates.c
! Fixed some MSVS2008 -W4 warnings (we use -W3 as the default).
; NOTE: There is a lot more of these.
* source/rtl/dates.c
* Formatting.
* contrib/hbziparc/hbziparc.prg
! RTE fixed when not specifying some optional params.
! Fixed to parse the right array, and thus not exluding
all filenames to be zipped.
; Disclaimer: I'm not a user of these functions, and
I've admittedly not tested them besides tests/unzipa.prg
and tests/zipa.prg (which are pretty basic cases), so
bumps are expected. If you find some bugs, please try
to provide patches to fix them. Thanks.
* contrib/hbziparc/make_b32.bat
* contrib/hbziparc/make_vc.bat
% Special C flags not needed anymore.
- contrib/hbziparc/hbzipold.cpp
+ contrib/hbziparc/hbzipcpp.old
* Renamed to not look like something live.
* harbour-ce-spec
* make_deb.sh
* harbour-w32-spec
* harbour.spec
* contrib/Makefile
* contrib/make_b32_all.bat
* contrib/make_gcc_all.sh
* contrib/make_vc_all.bat
+ contrib/hbziparc
- contrib/hbziparch
+ hbziparc readded to default contrib list, as it
creates no more harm now.
; NOTE: And now that this is about ready, we might
step forward to develop a .7z interface ;)
* contrib/hbziparch/Makefile
* contrib/hbziparch/common.mak
- contrib/hbziparch/hbziparc.c
- contrib/hbziparch/hbxdirec.c
* contrib/hbziparch/hbzipnew.cpp
* Merged Harbour C parts to .cpp file.
* Made full low-level interface static.
* contrib/hbziparch/hbziparc.c
* contrib/hbziparch/hbzipnew.cpp
% Merged slightly different internal functions, thus
removing significant redundancy.
Higher level hb_unzipFile(), hb_unzipFileIndex(), hb_unzipAllFile()
are now functionally the same, since the only difference was
how the file list parameter was handled.
! Removed any remaining Win32 dependency.
* contrib/hbziparch/hbziparc.c
* contrib/hbziparch/hbzipnew.cpp
* Complete cleanup/refactor/fixes.
(original functions and internal logic was kept,
except some shady path/curdir manipulations, which
were removed, along with any directory changes)
The code got significantly easier to oversee, so if
someone wants to find out how this interface works,
now the code is much easier to read. If we want to
keep this in Harbour, maybe we could just swap the
low-level interface to utilize hbmzip calls instead
of ZipArchive ones. We would lose the segmented
archive feature though, but I'm not sure how
important this is these days, when there is no
floppy disk anymore.
Code size is now 40% smaller than in 1.0.0.
It's still far from being perfect, but now it's at
least possible to fix.
* utils/hbmake/hbmake.prg
; Applied patch by Bill Robertson:
! Fixed previously introduced (by me actually) bug. Thanks.
% Removed "non-extended" codepath, since it was never used.
! UI elements fixed to properly accomodate to any
screen dimensions.
* utils/hbmake/hbmake.prg
! Fixed to not change casing of files, plus a few other minor fixes.
! Made one filename comparison multiplatform by using hb_FileMatch()
; Added TOFIXes where filename comparison is not multiplatform.
% Removed some dead or commented code.
- contrib/hbw32ddr
* contrib/Makefile
* contrib/make_b32_all.bat
* contrib/make_gcc_all.sh
* contrib/make_vc_all.bat
- Removed contrib hbw32ddr.
For those wanting to peek into this code, or
pick up development, the exact source can be
found in Harbour 1.0.0 release.
