0
server stopped
---
Brgds,
Viktor
On Wed, Jul 29, 2009 at 7:21 PM, Viktor Szakáts wrote:
>
> On 2009.07.29., at 7:14, Przemyslaw Czerpak wrote:
>
>> On Wed, 29 Jul 2009, Szak�ts Viktor wrote:
>>>
>>> Probably some dead obvious reason, but I'm getting:
As it has been suggested, you should try building with hbmk2.
Brgds,
Viktor
On 2009.07.29., at 9:32, Moises Romero wrote:
Hi There,
I am migrating my clipper application to Harbour. I have
400 .prg’s and 1 MyFunctions.c which I migrated from C++ 2.0 to C++
5.51.
Ev
..@harbour-project.org
[mailto:harbour-boun...@harbour-project.org] On Behalf Of Viktor
Szakáts
Sent: Wednesday, July 29, 2009 3:46 PM
To: Harbour Project Main Developer List.
Subject: Re: [Harbour] Linking Harbour with Bcc 5.5.1
As it has been suggested, you should try building with hbmk2.
Brg
hbmk2 is part of harbour distribution
you can dowload here
http://sourceforge.net/projects/harbour-project/files/
click and opopen binaries-windows and 2.0.0beta1 or 1.0.1
1.0.1 doesn't have it.
unofficial beta2 is available here:
http://syenar.hu/harbour
Brgds,
Viktor
___
wxHarbour or xhgtk.
wxHarbour uses native GUI subsystem on Win/OSX platforms and GTK on
Linux.
Brgds,
Viktor
On 2009.07.29., at 9:57, Alejandro de Garate wrote:
I need to build a lightweight application using Harbour and a
graphical interface for linux and win
I would like to use GTK l
You should also check tm_isdst value more precisely:
tm_isdst > 0 => daylight saving time is in effect
tm_isdst == 0 => no daylight saving
tm_isdst < 0 => information is not available
Here is modified version hb_timeUTCOffset():
{
struct tm timeinfo;
time_t current, utc, l
I didn't remove it, and don't intend to do so until we
have the new high-level which is being used here too.
I also some differences between the two branches, but didn't
investigate.
Brgds,
Viktor
On 2009.07.30., at 0:37, Francesco Saverio Giudice wrote:
Hi Viktor,
Il 30/07/2009 12.01, vszak
Hi All,
I'd like to add Jose Luis Capel to our developer team.
Jose is the author of SMS/SIM interface code just recently
committed by me, and this way he will be able to add further
updates directly to SVN.
Brgds,
Viktor
___
Harbour mailing list
Har
Il 30/07/2009 14.32, Przemyslaw Czerpak ha scritto:
Francesco please make quick test with Harbour code after:
2009-07-29 05:19 UTC+0200 Przemyslaw Czerpak (druzus/at/
priv.onet.pl)
or better simply take today CVS.
I did this morning before writing email and the behavior is that I
described,
; TODO:
+ Add other Sun Pro CPU targets.
I do not understand above. Most of *nixes builds can be compiled
with many different CPUs. The list of supported architectures
by such systems like Linux or *BSD is really huge. So far we
didn't have to replicate build configurations for each of t
AFAIK SunPro has SPARC 32/64 and x86/x64 CPU targets. It can be
solved by simple few lines line long wrapper .cf files, so it won't
create any maintenance nightmare. For this reason I don't see this
is a problem at all in this case. It's however more convenient for
someone who wants to build for m
Sry i don't understand the problem. There is 1 to 1 relation between
error nums and desc and num is still stored per socket. If the other
is an issue it should be fixed elsewhere too
On 7/31/09, Przemyslaw Czerpak wrote:
> On Fri, 31 Jul 2009, vszak...@users.sourceforge.net wrote:
>
> Hi,
>
>> 20
Citing "personal reason" is totally wrong. Unless messing/having fun
with harbour counts as such. Please stop using this argument. Or i can
drop dealing with every component which i dont use, this means about
90% of them. Imo we should allow such configs if possible, but the
focus should be on scen
All this seems completely reasonable to me, i'll make some steps asap
(when i get near my devenv tomorrow)
On 7/31/09, Przemyslaw Czerpak wrote:
> On Fri, 31 Jul 2009, Szak�ts Viktor wrote:
>> Citing "personal reason" is totally wrong. Unless messing/having fun
>> with harbour counts as such. Ple
Pelles C 6 final is not yet released and has bugs, use 5, which
is the latest stable release.
It's also useful to report this bug on Pelles C forums to help
the author fixing it.
Brgds,
Viktor
On 2009.08.01., at 1:49, Manu wrote:
pocc.exe -I. -Ze -Go -MT -W1 -Ot -I../../../../../include-
-Messaggio Originale- Da: "Petr Chornyj"
A:
Data invio: domenica 2 agosto 2009 17.14
Oggetto: Re: [Harbour] Array access at C level
> HB_FUNC( TEST )
> {
> hb_retc( hb_parc( 1, 1 ) );
> }
HB_FUNC( TEST )
{
hb_retc( hb_parvc( 1, 1 ) );
}
Can I ask the reason for this Clipper co
Hi Przemek,
Wishing you a very nice holiday.
Brgds,
Viktor
On 2009.08.02., at 0:34, Przemyslaw Czerpak wrote:
Hi All,
I'll be offline for a week.
best regards,
Przemek
___
Harbour mailing list
Harbour@harbour-project.org
http://lists.harbour-proje
You have to #include the Clipper compatible headers.
Brgds,
Viktor
On 2009.08.03., at 2:52, Enrico Maria Giordano wrote:
-Messaggio Originale- Da: "Viktor Szakáts" >
A: "Harbour Project Main Developer List." project.org>
Data invio: lunedì 3 agosto 2009
Hi Pritpal,
First - a good news - I am been able to simulate
TBrowse() behavior as far as vertical navigation is concerned.
Horizontal navigation - freeze - is practically non-existent in Qt's
model/view architect and hence so far I have no clues how
to simulate this. Barring "FREEZE" and "FOOTE
Hi Toninho,
I'd guess it's some sort of local corruption of your original bcc libs.
(or local corruption of Harbour sources).
For sure it works on my system both with default, and your settings.
Brgds,
Viktor
On 2009.08.05., at 0:18, toni...@fwi wrote:
Hi Massimo.
afaik you can remove
set
Hi Viktor,
The problem is solved by adding ws2_32.lib in
Yes, this symbol is defined in this lib. But it should
be linked automatically. At least on my system it does
without any special tricks.
\harbour\utils\hbrun\makefile
What can be the problem? My environment?
That would be my guess,
Hi Viktor
1. If I add
-debug
to my .hbp file I get:
LINK : warning LNK4098: defaultlib 'LIBCMTD' conflicts with use of
other libs; use /NODEFAULTLIB:library
with MSVC. I don't see any other difference in the output, maybe I
don't understand it.
You're trying to mix msvc libs with diffe
I got in log-msvc.bat:
C:\Users\Alex\AppData\Local\Temp\hbmk_sj1o73.rc(4) : error RC2135 :
file not found: C:\Userslex\SourceForge\harbour-2.0.0dev\package
\harbour.ico
hbmk: Error: Running resource compiler. 1
rc.exe -I"C:\users\alex\harbour-2.0.0dev\include" /fo
hbmk_sj1o73.res C:\Users
Viktor Szakáts wrote:
You're trying to mix msvc libs with different build settings, maybe
even
built with different msvc versions. Try rebuilding every libs with
the same
settings. Look at /MT /MD switches.
I don't think so, if -debug is used I get the message, if it is not
I
1. Whenever I compile, it will call the compiler for a temporary .c
file as below.
Setting environment for using Microsoft Visual Studio 2008 x86 tools.
hbmk: Processing configuration: C:\users\alex\harbour-2.0.0dev\bin
\hbmk.cfg
hbmk: Compiling...
hbmk: C compiler command:
cl.exe -nologo
attached below.
I can see the problem now, and it was what I'd guessed.
"\a" in the path is interpreted as escape char. I'll fix this ASAP.
Thanks for the report.
Brgds,
Viktor
Thanks
Alex
Blocked e-mail:
Viktor Szakáts wrote:
> Sorry it's not clear where 0
Hi Francesco,
Have a nice time off!
Brgds,
Viktor
On 2009.08.06., at 1:05, Francesco Saverio Giudice wrote:
I will be offline (reading list when possible) until 25th of August.
Best Regards,
Francesco
___
Harbour mailing list
Harbour@harbour-projec
Hi Maurilio,
since I have a few more days of vacations I'm trying to rebuild
current harbour cvs code on OS/2
I had to make a couple of changes to .cf files, but now I still have
a problem I'd like to discuss with you.
The problem is that config/os2-rm is called like this
os2-rm -r -f o
On 2009.08.07., at 4:08, maurilio longo wrote:
Hi Viktor,
Hi Maurilio,
It cannot be removed as is, as it may hold files for other compilers,
and cleaning one target should touch the others.
Ok, I do understand.
If there is a command to remove parent dirs *if empty*, in a clean
way, we m
Viktor,
silly me! Yes, there is an option in SmartSVN which stops it from
considering unversioned files and/or directories :)
Great :)
Thanks.
Maurilio.
PS. What about writing such a delete if empty in .prg code and
execute it with hbrun? We only need to make hbrun the latest deleted
Przemek, Viktor, Maurilio:
As make381-os2 show problems (limits) and incompatibility with
make381-windows (and perhaps Linux), there are some way to use
ANOTHER make system in OS/2 ?
Reading docs of Watcom I found wmake.exe, based in GNU make but with
some differences specified in docs
I tr
Hi,
I've made an attempt to implement this, but can't test it.
Could you send a code example with some default XP OLE component
which with I can test it?
Brgds,
Viktor
On 2009.08.08., at 9:30, toni...@fwi wrote:
Hi,
I have problems with OLE and variables by reference, exemple:
---cut---
oO
Okay. I hope someone will post portable test code for this
feature and I can look at it further, this way it's only
shot in the dark.
Brgds,
Viktor
On 2009.08.09., at 1:49, toni...@fwi wrote:
I'm rather sending you the patched source for you to do the testing,
as I don't want to install any ne
Hi Toninho,
Okay. I hope someone will post portable test code for this
feature and I can look at it further, this way it's only
shot in the dark.
Hi Viktor,
if you want to test, please download this package
(http://www.fwi.com.br/nfe.zip) and don't install anything, only you
need is dotnetfx.
I was thinking that's intentional, in the early days it displayed
proper tree lines, and after an update it has changed to the current
colored style. Looks indeed a bit strange though.
And hereby congrats to Pritpal for the new browse functionality.
It's really nice job.
Brgds,
Viktor
On 2009.
Viktor Szakáts wrote:
I was thinking that's intentional, in the early days it displayed
proper tree lines, and after an update it has changed to the current
colored style. Looks indeed a bit strange though.
Yes, this behavior is intentional. I am expermenting with
CSS compatible
cking again to static
fully fixes the shapes there too.
Brgds,
Viktor
On 2009.08.09., at 6:11, Pritpal Bedi wrote:
Hello Viktor
Viktor Szakáts wrote:
I was thinking that's intentional, in the early days it displayed
proper tree lines, and after an update it has changed to the curre
This year you changed make process and in meantime some workarounds
for OS/2-os2make376 were left off. Now Harbour under OS/2 does not
build with os2make376 and partially build with os2make381
We are currently discarding os2make376 and forcing to use
os2make381, and trying to find why and how
After this change the browse became extremely slow.
It comes down to the change in HBQT_BRW_CELLVALUE case.
If I restore this section, it's fast again.
Brgds,
Viktor
On 2009.08.10., at 1:19, vouch...@users.sourceforge.net wrote:
Revision: 12016
http://harbour-project.svn.sourceforge
Viktor Szakáts wrote:
After this change the browse became extremely slow.
It comes down to the change in HBQT_BRW_CELLVALUE case.
If I restore this section, it's fast again.
Please test again after:
2009-08-09 17:35 UTC-0800 Pritpal Bedi (prit...@vouchcac.com)
* contrib
Hi Pritpal,
On 2009.08.10., at 3:00, Pritpal Bedi wrote:
Viktor Szakáts wrote:
Thanks, did it now. It's at least that slow or even more so.
It's as if the keyboard speed would be throttled to the slowest
position (but it's not), while before it was just very fast.
What
Hi Przemek,
I hope vacation was good. Good to see you back.
2009-08-04 03:32 UTC+0200 Viktor Szakats (harbour.01 syenar.hu)
* source/vm/dynlibhb.c
! Fixed to not use dl*() functions on linux/sunpro.
Above modification simply broke working code by disabling all calls to
dl*() functions ins
Hi Viktor,
when hbmk2 is used with GCC without support for
-Wl,--start-group/-Wl,--end-group then the order
of linked libraries is not sufficient to resolve
cross references.
Just try to compile this code (for test you can
disable -Wl,--start-group in mingw builds):
proc main()
browse()
Hi Maurilio,
Yes, the CRC16 is different indeed, but the CRC32 was the same, in
fact (reading the old comments) the CRC32 code was taken from a generic
unicode library, so it's nothing tpathy specific, before deletion
I've verified the code table and formula to be the same as in core.
I've now c
../../../../../bin/darwin/gcc/harbour ../../../xbpbrowse.prg -n1 -
i../../../../../include -q0 -w3 -es2 -kmo -l -gc0 -I../../../../../
contrib/hbqt
../../../xbpbrowse.prg(1205) Warning W0001 Ambiguous reference
'QHEADERVIEW_INTERACTIVE'
../../../xbpbrowse.prg(1205) Warning W0001 Ambiguous re
ribs).
Brgds,
Viktor
On 2009.08.10., at 4:53, Pritpal Bedi wrote:
Hi
Viktor Szakáts wrote:
../../../../../bin/darwin/gcc/harbour ../../../xbpbrowse.prg -n1 -
i../../../../../include -q0 -w3 -es2 -kmo -l -gc0 -I../../../../../
contrib/hbqt
../../../xbpbrowse.prg(1205) Warning W
The error message is missing from your post.
You try to build with HB_REBUILD_PARSER=yes, and in this
case bison needs to be properly configured, which can
be tricky on Windows. I can't help you further as it's
been years I've tried this scenario.
Brgds,
Viktor
On 2009.08.10., at 5:38, Manu wro
Thank you Francek.
Brgds,
Viktor
On 2009.08.11., at 7:51, FRANČEK PRIJATELJ wrote:
Hi
In global.cf, line 210 is error (HB_LIB_INSTAL: ... , instead off
HB_INC_INSTALL
ifneq ($(HB_INC_INSTALL),)
$(info ! HB_LIB_INSTALL: $(HB_INC_INSTALL))
endif
BRGS
Franček
Hi Przemek,
I've recently fixed hb_fsCreate*() to honor FO_EXCL flag
on win platforms (it was ignored before). AFAICS this error
could be caused by interaction of temp file creation routines
in fstemp.c and above change.
I'm investigating, but I'd appreciate your help. For me its
end endless str
- Use FO_TRUNC on win platforms.
Maybe the second, but both needs win-only tweak, so maybe there is
a more elegant method.
Brgds,
Viktor
On 2009.08.11., at 11:05, Viktor Szakáts wrote:
Hi Przemek,
I've recently fixed hb_fsCreate*() to honor FO_EXCL flag
on win platforms (it was ignored befor
Hi Przemek,
On 2009.08.11., at 0:02, Przemyslaw Czerpak wrote:
On Mon, 10 Aug 2009, vszak...@users.sourceforge.net wrote:
Hi Viktor,
2009-08-10 22:34 UTC+0200 Viktor Szakats (harbour.01 syenar.hu)
* contrib/hbtpathy/telepath.prg
* contrib/hbtpathy/tpcommon.c
+ TP_CRC16() reimplemented i
Here is darwin 'man mkstemp':
---
MKTEMP(3)BSD Library Functions Manual
MKTEMP(3)
NAME
mkdtemp, mkstemp, mktemp, mktemps -- make temporary file name
(unique)
LIBRARY
Standard C Library (libc, -lc)
SYNOPSIS
#include
char *
mkdtemp(c
| The implementation of these functions calls arc4random(3), which is
| not reentrant. You must provide your own locking around this and
other
| consumers of the arc4random(3) API.
I do not like it. It means that in MacOSX and probably in BSD builds
we will have to add mutex protection for M
On Tue, 11 Aug 2009, Szak�ts Viktor wrote:
Maybe we should add to RTL set of functions to swap bytes in
numbers?
I.e. hb_swapI(), hb_swapW(), hb_swapL(), hb_swapU(), hb_swapLL()
The suffixes should correspond to BIN2?() functions but maybe you
have
better proposition for names.
With such fun
Hi Toninho,
You're doing nothing wrong, just discovered some non-ported
dependencies of these xhb functions.
I've checked and these are for me too problematic functions
with several deeper dependencies so I wouldn't dare trying
to port them to Harbour.
I'd suggest to look around to use some alt
Hi Itamar,
Yes, that's right, you need to take care of building
components in correct order:
hbmk2 rddleto.hbp letodb.hbp
(see readme.txt with these instructions)
Brgds,
Viktor
On 2009.08.12., at 1:46, Itamar Lins wrote:
Hi!
To Guru Mr. Viktor Szakats!
file letodb.hbp not make rddleto.lib f
There is also bin/hb-mkimp.bat inside the Harbour source
tree which solves the problem for most C compilers supported
by Harbour. Should be quite easy to adapt the logic for any
.dll file.
Maybe in the future this process will be built inside hbmk2,
I'll have to think about it. For sure on some c
Hi Lorenzo,
Thanks, it's caused by some formatting yesterday.
Seems like blocks inside define/endef cannot be
indented without side effect, I'll revise them.
Brgds,
Viktor
On 2009.08.12., at 7:26, Lorenzo Fiorini wrote:
Today under Fedora 11 I get:
make all
! MAKE: make
! MAKE: make
! HB_INS
After svn update 12079, under Windows XP msys I get:
$ make clean
! MAKE: make
! MAKE: make
! HB_INSTALL_PREFIX: /opt/harbour
! Autodetected C compiler: mingwarm
! HB_BIN_COMPILE not specified. Automatically set to: ./bin/win/mingw/
! HB_BIN_COMPILE not specified. Automatically set to: ../bin/win
Hi Przemek,
Please do these modification.
Brgds,
Viktor
On 2009.08.12., at 10:11, Przemyslaw Czerpak wrote:
On Mon, 10 Aug 2009, Tamas TEVESZ wrote:
Hi,
HB_CMP = suncc
instead of CC and cc. With above version you do not have keep
path to
Sun C compiler at the beginning of PATH envvar
It may be better to address letodb issues directly to the
letodb mailing list. This problem isn't a Harbour one.
This error shows the problem where 3rd party code uses
Harbour namespace (hb_*) for it's own public symbols. There
is no guarantee that Harbour won't eventually use any of
those. In th
Ops, thank you. I've removed a few of those.
I'll recheck, but I wonder how it didn't come
up when doing lots of local tests. I also
couldn't find the guiding logic regarding this
matter before removing. In some places it
was used, in some other it wasn't. Also same
trick isn't needed for C_SOURC
inclusion rules. ]
I'd guess the issue is somewhere else, maybe side effect
of other changes. But what?
Brgds,
Viktor
On 2009.08.12., at 11:41, Viktor Szakáts wrote:
Ops, thank you. I've removed a few of those.
I'll recheck, but I wonder how it didn't come
up when doing lots of lo
On Wed, Aug 12, 2009 at 11:54 AM, Viktor Szakáts0...@syenar.hu> wrote:
I'd guess the issue is somewhere else, maybe side effect
of other changes. But what?
I'm also trying to understand what could be.
Unfortunately I don't build frequently under Windows and there are
tons of diffs between act
Good finding. And it's even the fault of the generator,
the last empty lines were removed by my editor.
Anyhow I'll add a trick to not let this happen again.
Brgds,
Viktor
On 2009.08.12., at 0:53, Lorenzo Fiorini wrote:
On Wed, Aug 12, 2009 at 12:40 PM, Viktor Szakáts0...@syenar.hu> wrote:
I
>; QUESTION: Could an OS/2 user tell me whether $(CMDPREF) is
>*really* needed? This is the only place it's used,
>so I wonder. If not, we should delete it.
I do not know what $(CMDPREF) mean or what it does
In current builds I do not see any difference
Thank
; TODO: Streamline macro usage regarding external dependencies.
There is currently HAS*, HAVE*, HB_HAVE*, HB_WITHOUT*
plus some other variations.
We should probably stick to HB_BUILD_WITH_* for user
level control, and HB_HAS_* for internal detection l
2009-08-11 13:24 UTC+0200 Viktor Szakats (harbour.01 syenar.hu)
* Cleaned LDLIBS usage:
One of recent modifications enabled some unnecessary libraries
in linking phase, i.e. now harbour compiler in Linux builds is linked
with -lgpm, i.e.:
ldd harbour
linux-gate.so.1 => (0xe000)
Thank you. It was explicitly invoking the shell command processor
to launch the shell commands. Probably much better to use $(shell)
command in case such forced shell invocation is to be used, but
these weren't such cases.
I only guess but IMO it was workaround for older GNU make versions
which
Hi Pritpal.
Okay, so the regression I've signaled just kicked in.
I'd appreciate if someone could tell some good advice
how to order libs to make it work without lib grouping
in this case.
Brgds,
Viktor
On 2009.08.12., at 6:46, Pritpal Bedi wrote:
Hello Viktor
Today I updated from SVN and
Hi,
external/libhpdf/Makefile
39 ifeq ($(HB_XBUILD),)
40 HB_INC_LIBPNG = /usr/include
is there any reason for explicitly specifying /usr/include here? it'll
be picked up by the compiler anyway, but now i've found a case where
it _might_ cause some level of harm - probably not in hpdf, though,
Hi Pritpal,
Trouble isn't that big fortunately.
Please modify contrib/hbqt/hbqt.hbc so that the
line ending with supc++ gets moved after the
last libs= line. This should solve the problem.
I'll commit it in a while, currently I'm in the middle
of other changes.
Brgds,
Viktor
On 2009.08.12., a
i see it's to check for the presence of png.h, and such techniques
are
used elsewhere too. would filtering /usr/include out of the include
list be too cumbersome?
I'm not sure I understand. Could you tell some more about this?
what i mean that after all the magic is done, it ends up with a
c
what i mean that after all the magic is done, it ends up with a
command line like that:
cc -I. -I../../../../../include -K udk -KPIC \
-I../../../../../source/hbzlib -I/usr/include -c \
../../../_hbhbpdf.c -o_hbhbpdf.o
(wrapped for clarity). would it be too cumbersome to by some means
filt
Hi Pritpal,
A windows.h slipped into the source.
Brgds,
Viktor
On 2009.08.13., at 7:50, vouch...@users.sourceforge.net wrote:
Revision: 12091
http://harbour-project.svn.sourceforge.net/harbour-project/?rev=12091&view=rev
Author: vouchcac
Date: 2009-08-13 05:50:56 + (Thu, 1
Hi All, especially OS/2 users,
We now have config/os2-mkdir.exe config/os2-rm.exe
hosted in SVN, with links provided in config/readme.txt.
There is one problem though, the website which I used
to get the binaries doesn't provide the sources.
I'd like to ask OS/2 users to look for different bina
Hi David,
>[ or, if someone can find the source for currently
>hosted binaries, it's also a solution. ]
hobbes.nmsu.edu/pub/os2/util/disk/gnufutil.zip
Thank you. I can see the list of files via FTP, but
can't find this specific file there.
Update: Found it in ../file !
The only problem with
Yes, that'd be good, although the one we have already
has these properties: okay dependencies, no source.
Maybe I'll delete the tools and look for mkdir/rm at
build time, issue error if missing, here my only concern
is that mkdir has the same name as built-in OS/2 mkdir,
so I don't know how these
I don't agree with all the verbose output, but as I said
you can enable it locally if you need them, please test if
it works, if not we need to fix it.
As for .tds, I intentionally didn't create difference
in logic for different shells, specifically to help cross
build scenarios. f.e. bcc can be
>I don't agree with all the verbose output, but as I said
>you can enable it locally if you need them, please test if
>it works, if not we need to fix it.
I do not found use of HB_USER_MKFLAGS, so I used
set HB_USER_MAKEFLAGS=-w
Yes, my mistake. It's best to refer to docs (INSTALL) rather
than
I am waiting for any guru corrections in letodb source.
Unfortunately I do
not program in C/C++.
But I change the name hb_socketConnect to hbleto_socketConnect file
hbip.c
and my project link fine, but not works more ... :-( GPF show in
letodb.exe.
You need to change all occurrences, not j
POCC 6 final release is out.
Unfortunately its bugs hitting Harbour weren't fixed,
so it's broken in both 32-bit mode (internal error),
and 64-bit mode (link error - reported on pocc forum a
month ago).
FYI and don't use this compiler with Harbour.
Brgds,
Viktor
___
Hi Przemek,
Probably the OS CP conversion should still be retained in this case.
Brgds,
Viktor
On 2009.08.14., at 3:03, dru...@users.sourceforge.net wrote:
Revision: 12106
http://harbour-project.svn.sourceforge.net/harbour-project/?rev=12106&view=rev
Author: druzus
Date: 2009-
Much better this way, thank you.
Brgds,
Viktor
On 2009.08.14., at 3:13, Przemyslaw Czerpak wrote:
On Fri, 14 Aug 2009, Szak�ts Viktor wrote:
Hi Przemek,
Probably the OS CP conversion should still be retained in this case.
Yes but not here.
See my second commit.
best regards,
Przemek
__
I think this is a new one:
---
! MAKE: mingw32-make.exe 3.81 clean install sh.exe
! HB_INSTALL_PREFIX: C:\devl\hb20
! HB_BUILD_PKG: yes
! HB_BUILD_DLL: yes
! HB_BUILD_DEBUG: no
! HB_BUILD_OPTIM: yes
! HB_HOST_ARCH: win HB_SHELL: nt
! HB_ARCHITECTURE: wce (autodetected)
! HB_COMPILER: poccarm
! HB_
Thank you, that solved it nicely.
It's not over yet though :( Here's the next new one:
---
pocc.exe -I. -Ze -Go -W1 -Ot -Tarm-coff -D_M_ARM -D_WINCE
-I../../../../../include-DUNICODE -Foolecore.obj -c
../../../olecore.c
../../../olecore.c(239): error #2152: Unknown field 'llVal' of 'union
(no
---
../../../hbsocket.c:179:1: warning: "socklen_t" redefined
In file included from /usr/include/sys/socket.h:15,
from ../../../hbsocket.c:167:
/usr/include/cygwin/socket.h:24:1: warning: this is the location of
the previous definition
../../../hbsocket.c: In function `hb_socketIni
On Fri, 14 Aug 2009, Szak�ts Viktor wrote:
It's not over yet though :( Here's the next new one:
---
pocc.exe -I. -Ze -Go -W1 -Ot -Tarm-coff -D_M_ARM -D_WINCE
-I../../../../../include-DUNICODE -Foolecore.obj -c
../../../olecore.c
../../../olecore.c(239): error #2152: Unknown field 'llVal' of
Hi Toninho,
Please don't use make_gnu.bat anymore.
Instead, either run mingw32-make.exe from your path (see INSTALL how to
get this file), or use the still provided config/mingw32-make.exe (you can even
copy that to your PATH).
[ config/mingw32-make.exe may be deleted in the near future from rep
This is a very very old problem, but only Harbour gives a
warning when redefining the same macro twice with same content,
I'll try to do something about it and I just now realized that
Harbour make uses a trick to relaunch itself even when making
a single target. I don't know the reason yet or whe
Please check any examples in INSTALL:
F.e.:
---
set PATH=C:\Borland\BCC55\Bin;%PATH%
rem
mingw32-make %1 %2 > log.txt 2>&1
---
Brgds,
Viktor
On Sat, Aug 15, 2009 at 11:47 AM, Enrico Maria
Giordano wrote:
> I tried
>
> mingw32-make
>
> but I'm not sure if it worked fine (it look
I've patched it to build on Harbour (using BCC as the most sloppy
compiler, the rest still have hundreds of warnings and errors), but
I have to say this component isn't the cleanest code I've ever seen.
I couldn't create one example which would work, but there is GPF,
various link errors due to mis
Just to report that using the prefix in the subj the libs are put in
/opt/harbour/lib/harbour.
make_gnu.sh worked since it tested only for /opt|/usr.
I've tried to replicate it in global.cf but I didn't find any clean
solution so to fix it
I've removed from config/global.cf:
# Not perfect, please
I didn't fix it yet. I'll do it tomorrow most probably,
I already decided on the way to fix this. Until then just
temporarily delete your HB_USER_PRGFLAGS and it will be
just fine.
Brgds,
Viktor
On 2009.08.15., at 5:37, FRANČEK PRIJATELJ wrote:
Viktor Szakáts wrote:
This is a
Hi,
Please read recent (and not so recent) ChangeLog entries,
these files are not needed anymore and they got deleted.
Brgds,
Viktor
On 2009.08.15., at 3:23, Rossine wrote:
Hello,
I deleted my directory "c:\harbour" and updated it again and all
these files
are not downloaded:
make_gnu.b
o ,Christian Azambuja and Marson)
Will gtwvw & wvwtools start a new open source project?
2009/8/15 Viktor Szakáts :
I've patched it to build on Harbour (using BCC as the most sloppy
compiler, the rest still have hundreds of warnings and errors), but
I have to say this component isn't
Nice :(
I've asked on the list whether COMSPEC is really needed,
got no answer, just this silent restore now, after David
confirmed it's working even without it.
So: Is it necessary?
Brgds,
Viktor
On 2009.08.15., at 11:40, mauriliolo...@users.sourceforge.net wrote:
Revision: 12125
Hi Reynaldo,
Now, using xHb+MSVC, gtwvw can be compiled and link and so the
gtwvw.lib
genarated without any warnings and errors, at all. And the
executable file,
created from "\xharbour\contrib\gtwvw\samples\wvwtest9.prg" runs fine.
Since now, I would like to try to test gtwvw with Harbour,
BTW it would be also nice to hear what was the problem
with previous commit (which was working for David), I've
copied the DOS shell method, where the command line limit
is even more restricted, yet it works fine. Maybe it'd be
enough to use GNU cp (like with DOS) for OS/2 too, since
that was also
Thanks for the input. Could you try the just removed
method with readded $(COMSPEC) /C? Maybe the logic was
good (after all it even works in MS-DOS), just this bit
was causing freeze.
Brgds,
Viktor
On 2009.08.16., at 0:37, maurilio longo wrote:
Viktor,
without the comspec trick a make install
1 - 100 of 4712 matches
Mail list logo
