On Fri, Sep 15, 2006 at 09:27:36PM -0500, Bo Peng wrote:
> > I don't think
> > that scons let you choose these locations independently of each other,
> > though. Bo?
>
> It is not difficult at all to add these options, but will lyx know
> where to find these files?
I think so. Autotools let you
I don't think
that scons let you choose these locations independently of each other,
though. Bo?
It is not difficult at all to add these options, but will lyx know
where to find these files?
Bo
On Fri, Sep 15, 2006 at 12:52:46PM -0500, Bo Peng wrote:
> > Thanks, it compiles now and run for me.
> >
> > Is there a special scons command to install lyx and create the sysdir
> > somewhere ?
>
> What is sysdir for? I do not have this option yet. You can do
>
> scons DESTDIR=/path/to/de
Thanks, it compiles now and run for me.
Is there a special scons command to install lyx and create the sysdir
somewhere ?
What is sysdir for? I do not have this option yet. You can do
scons DESTDIR=/path/to/dest install
to install everything to /path/to/dest. Let me know what is under
Bo Peng wrote:
>> Maybe lyx::char_type is not wchar_t because you forgot to define
>> HAVE_WCHAR_T? Or SIZEOF_WCHAR_T is not 4. Did you double check?
>
> Well, I was bitten by the separation of config.h... HAVE_WCHAR_T is
> defined only in boost/config.h, not in other config files. A patch
> that
Maybe lyx::char_type is not wchar_t because you forgot to define
HAVE_WCHAR_T? Or SIZEOF_WCHAR_T is not 4. Did you double check?
Well, I was bitten by the separation of config.h... HAVE_WCHAR_T is
defined only in boost/config.h, not in other config files. A patch
that fixes the crash will be app
> "Georg" == Georg Baum <[EMAIL PROTECTED]> writes:
Georg> Peter Kümmel wrote:
>> Indeeed, there are two definitions of floatname(...): insetwrap.C
>> and insetfloat. One should be removed.
Georg> There is some duplicate code in insetwrap and insetfloat. One
Georg> should think about subclass
Bo Peng wrote:
> terminate called after throwing an instance of 'std::bad_cast'
> what(): St8bad_cast
>
> Program received signal SIGABRT, Aborted.
> [Switching to Thread 182896208448 (LWP 31188)]
> 0x003c8962e21d in raise () from /lib64/tls/libc.so.6
> (gdb) bt
> #0 0x003c8962e21d in
> The scons fix is in. Lyx can start, but crashes when I open a help
> document.
Which OS? backtrace?
RHEL4, x86-64. SIZEOF_WCHAR_T is 4.
Same as entered returned
Same as entered returned
Same as entered returned
Same as entered returned
Same as entered returned
Drawing string
Drawing lyx::cha
Bo Peng wrote:
> The scons fix is in. Lyx can start, but crashes when I open a help
> document.
Which OS? backtrace?
Georg
Peter Kümmel wrote:
> Indeeed, there are two definitions of floatname(...): insetwrap.C and
> insetfloat. One should be removed.
There is some duplicate code in insetwrap and insetfloat. One should think
about subclassing or something like that, but nevertheless the linker
should not complain bec
[EMAIL PROTECTED] wrote:
> insets/.libs/libinsets.a(insetwrap.o): In function
> `__gnu_cxx::new_allocator::deallocate(CursorSlice*, un
> signed int)':
> /usr/local/src/lyx-cvs/lyx-devel/src/insets/insetwrap.C:49: multiple
> definition of `(anonymous namespace)::floatname(s
> td::basic_string, std:
Missing scons support for SIZEOF_WCHAR_T. It will probably work with
autotools or if you add
The scons fix is in. Lyx can start, but crashes when I open a help document.
Bo
Bo Peng wrote:
> The biggest change to your patch is that I define SIZEOF_WCHAR_T
> directly,
That's better.
--
Peter Kümmel
[EMAIL PROTECTED] wrote:
> Hello,
>
> I have the same error executing LyX after applying Peter's patch and trying
> Georg's workaround.
>
> Trying to go the autotools way leads to an error in the ld phase :
>
Indeeed, there are two definitions of floatname(...): insetwrap.C and
insetfloat.
One
You could drop all size-2 relevant code.
I've added it only for completeness.
# define SIZEOF_WCHAR_T 2
The biggest change to your patch is that I define SIZEOF_WCHAR_T
directly, without SIZE_OF_WCHAR_T_IS_2 etc since I do not see them in
the source.
Anyway, I am applying the patch.
Bo
Bo Peng wrote:
> We do not really need SIZEOF_WCHAR_T_IS_2 and 4, right?
No.
> Then, I
> propose the following simplified version. It will be applied if there
> is no objection.
The only thing what matters is whether it sets SIZEOF_WCHAR_T correctly. If
it does, put it in.
Georg
Bo Peng wrote:
>> +'SIZEOF_WCHAR_T_IS_2',
>> +'SIZEOF_WCHAR_T_IS_4',
>> +#ifdef SIZEOF_WCHAR_T_IS_2
>> +# define SIZEOF_WCHAR_T 2
>> +#else
>> +# ifdef SIZEOF_WCHAR_T_IS_4
>> +#define SIZEOF_WCHAR_T 4
>> +# endif
>> +#endif
>
> We do not really need SIZEOF_WC
+'SIZEOF_WCHAR_T_IS_2',
+'SIZEOF_WCHAR_T_IS_4',
+#ifdef SIZEOF_WCHAR_T_IS_2
+# define SIZEOF_WCHAR_T 2
+#else
+# ifdef SIZEOF_WCHAR_T_IS_4
+#define SIZEOF_WCHAR_T 4
+# endif
+#endif
We do not really need SIZEOF_WCHAR_T_IS_2 and 4, right? Then, I
propose the
Hello,
I have the same error executing LyX after applying Peter's patch and trying
Georg's workaround.
Trying to go the autotools way leads to an error in the ld phase :
g++ -g -O -o lyx-qt4 main.o Bidi.o BufferView.o BufferView_pimpl.o Bullet.o
BranchList.o Chktex.o Color.o CutAndPaste.
o DepTa
On 9/14/06, Peter Kümmel <[EMAIL PROTECTED]> wrote:
Peter Kümmel wrote:
> + conf.Message('Check if size of whar_t is 4 ... ')
> +ret = conf.TryLink(check_whar_t_4, '.cpp')
Will this not give an error?
Peter, thanks for your patch. I will have a look and apply if it is
not yet applied
Peter Kümmel wrote:
> + conf.Message('Check if size of whar_t is 4 ... ')
> +ret = conf.TryLink(check_whar_t_4, '.cpp')
Will this not give an error?
--
Peter Kümmel
Index: SConstruct
===
--- SConstruct (revision 14991)
Georg Baum wrote:
> Missing scons support for SIZEOF_WCHAR_T.
here an untested patch which maybe needs some small fixes.
--
Peter Kümmel
Index: SConstruct
===
--- SConstruct (revision 14991)
+++ SConstruct (working copy)
@@ -653,
[EMAIL PROTECTED] wrote:
> [EMAIL PROTECTED]:/usr/local/src/lyx-cvs/lyx-devel/development/scons/debug$
> ./lyx-qt4 Same as entered returned
> terminate called after throwing an instance of 'std::bad_cast'
> what(): St8bad_cast
> Abandon
>
> I'm again clueless...
Missing scons support for SIZE
On Thursday 14 September 2006 10:01, [EMAIL PROTECTED] wrote:
> I'm again clueless...
Now you are on par with the rest of us. ;-)
The issue is being handled by Georg, so it is a know issue.
> Cheers,
> Charles
--
José Abílio
Hello,
I'm making progress :
- installing libqt4-debug and libqt4-dev-debug
- Running update-alternatives --config moc and update-alternatives --config
uic to point to qt4 version
- changing QTDIR to /usr/share/qt4 (don't know if its necessary)
I'm able to compile lyx-qt4
But when I execute it,
On Wed, Sep 13, 2006 at 07:23:55PM -0500, Bo Peng wrote:
> >I'm wondering if it does not pick the wrong moc. On Debian, moc is called
> >moc-qt4
>
> Why is it so? Is there a moc-qt3?
Yes and a symlink "moc" managed via update-alternatives pointing to the
current Debian default.
Yes sometimes Debi
I'm wondering if it does not pick the wrong moc. On Debian, moc is called
moc-qt4
Why is it so? Is there a moc-qt3?
Bo
With your patch, it compiles but stops with the qt4 frontend :
scons frontend=qt4 qt_dir=/usr/share/qt4 qt_lib_path=/usr/lib/qt4 -j3 lyx
g++ -o
debug/common/frontends/qt4/Action.o -c -g -O -DHAVE_CONFIG_H
-DQT_CLEAN_NAMESPACE -DQT_GENUINE_STR -DQT_NO_STL -DQT_NO_KEYWORDS -DQT_GUI_LIB
-Idebug
Bo Peng wrote:
>> I need to investigate this problem more.
>
> Can you compile the following .c code without libiconv installed, by
> using glibc or something else?
>
> #include
> #include
> int main() {
> iconv_t cd = iconv_open("","");
> iconv(cd,NULL,NULL,NULL,NULL);
> iconv_clo
On 9/13/06, Bo Peng <[EMAIL PROTECTED]> wrote:
> I need to investigate this problem more.
Could you please test the attached patch? I have added a check for
iconv function, without those libs.
Index: development/scons/SConstruct
=
I need to investigate this problem more.
Can you compile the following .c code without libiconv installed, by
using glibc or something else?
#include
#include
int main() {
iconv_t cd = iconv_open("","");
iconv(cd,NULL,NULL,NULL,NULL);
iconv_close(cd);
}
You can modify the code, add
> I have on my system /usr/local/lib/libiconv.so ...
You have downloaded and installed manually libiconv. Is it necessary ?
I am not 100% sure. If libc or someone else provides iconv() function,
then I should not have required it in scons
I need to investigate this problem more.
Bo
There is no libiconv.so
I thought iconv functionality was covered by glibc in Linux.
I have on my system /usr/local/lib/libiconv.so ...
Bo
Bo Peng wrote:
>> Trying scons, I'm quickly stopped :
>>
>> tombouctou:/usr/local/src/lyx-devel# scons -f
>> development/scons/SConstruct qt_lib_path=/usr/include/qt4
>> qt_dir=/usr/include/qt4 extra_lib_path=/usr/include all
>>
>> scons: Reading SConscript files ...
>> Checking for pkg-config...y
Trying scons, I'm quickly stopped :
tombouctou:/usr/local/src/lyx-devel# scons -f development/scons/SConstruct
qt_lib_path=/usr/include/qt4 qt_dir=/usr/include/qt4
extra_lib_path=/usr/include all
scons: Reading SConscript files ...
Checking for pkg-config...yes
Checking for main() in C library z
36 matches
Mail list logo
