Re: [Libreoffice] Make dev-install without running smoketests
On Wed, Aug 03, 2011 at 10:42:55PM +0200, Lionel Elie Mamane wrote: > On Wed, Aug 03, 2011 at 04:51:13PM +0100, Caolán McNamara wrote: >> What are your configure options for me to see if I can replicate this ? >> Looks like evilness which should be fixed. > I'm in a similar situation on master branch, but for me the smoketest > fails in officeconnection.cxx in function OfficeConnection::setUp(), > at that assertion: > for (;;) { > try { > context_ = > css::uno::Reference< css::uno::XComponentContext >( > resolver->resolve( > rtl::OUString(RTL_CONSTASCII_USTRINGPARAM("uno:")) > + > desc + > rtl::OUString( > RTL_CONSTASCII_USTRINGPARAM( > ";urp;StarOffice.ComponentContext"))), > css::uno::UNO_QUERY_THROW); > break; > } catch (css::connection::NoConnectException &) {} > if (process_ != 0) { > TimeValue delay = { 1, 0 }; // 1 sec > CPPUNIT_ASSERT_EQUAL( > osl_Process_E_TimedOut, > osl_joinProcessWithTimeout(process_, &delay)); > } > } > (..) the assertion succeeds several times, and then > finally fails, when the return value of the function call becomes > osl_Process_E_None (that is, 0). Forgot to write: I configured with '--prefix=/home/master/src/libreoffice/usr/' '--verbose' '--with-num-cpus=2' '--with-max-jobs=2' '--enable-debug' '--enable-dbgutil' '--enable-crashdump' '--enable-dependency-tracking' '--with-extension-integration' '--with-system-libs' '--without-system-mozilla' '--with-build-version=(LEM development build)' '--with-system-odbc' '--with-system-sane-header' '--with-system-xrender-headers' '--disable-zenity' '--enable-gcjaot' '--enable-ext-report-builder' '--without-system-jars' '--with-openldap' '--without-system-lpsolve' '--enable-ext-mysql-connector' '--with-system-mysql' Compiler: gcc (Debian 4.6.1-5) 4.6.1 Environment: Debian GNU/Linux squeeze (stable) and sid (unstable) mix -- Lionel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] link error in comhelper with undefined references from ucbhelper
Hi, building master (as of this evening with origin 23f430a7dea74ee4eeb797ae5874384e34da362f in libs-gui) unxlngi6 gave a linker error in comphelper complaining about undefined references that should be provided by ucbhelper, for example ucbhelper::Content::Content() ucbhelper is built, libucbhelper4gcc3.so exists, contains the symbols and is linked to comphelper with -lucbhelper4gcc3 I'm currently stuck. Ideas, anyone? Eike -- PGP/OpenPGP/GnuPG encrypted mail preferred in all private communication. Key ID: 0x293C05FD - 997A 4C60 CE41 0149 0DB3 9E96 2F1A D073 293C 05FD pgpdpfMKQq8HX.pgp Description: PGP signature ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [PATCH] Easy Hack - Get rid ofSvULongs in writer
Dnia 2011-08-03, śro o godzinie 00:23 -0400, Kohei Yoshida pisze: > > @@ -2017,7 +2011,7 @@ bool > WW8_WrPlcSubDoc::WriteGenericTxt( WW8Export& rWrt, sal_uInt8 nTTyp, > { > // textbox - content > WW8_CP nCP = rWrt.Fc2Cp( rWrt.Strm().Tell() ); > -aCps.Insert( nCP, i ); > +aCps.push_back( nCP ); > pTxtPos->Append( nCP ); > > Hmm are you sure that you can replace that with vector's push_back() > there? For this to work correctly the i has to always point to the > end > position (last element + 1), but from the surrounding code I'm not > sure > if that's guaranteed. You probably should use vector's insert() > method > with a specific insert position, instead of using push_back() in this > case. Ups, I will be more careful in the future, I have checked the function itself but as you said calling this function is not guaranteed with empty aCps. > > Also, I've seen a code like the following > > if (!vec.empty()) > vec.clear(); > > in several places. While this code is not wrong (since the original > code was written that way), calling clear() on an empty vector does > nothing & is harmless. So just calling clear() should be fine without > checking for empty-ness. Ok changed. > Other than that, the rest looks okay to me. I'd prefer someone more > experienced in Writer's code to sanity-check though. O, so by the way there is in wrtww8.hxx a VARARR WW8Bytes which should be migrated to ww::bytes. Is it okay to do? Best Regards, Maciej >From 80b069e455f1f6f166d279235f3b8c14da07c9b8 Mon Sep 17 00:00:00 2001 From: Maciej Rumianowski Date: Thu, 4 Aug 2011 00:38:26 +0200 Subject: [PATCH] Get rid of SvULongs in writer Replace SvULongs with std::vector. In SwHTMLWriter simultaneously replace aOutlineMarks with boost::ptr_vector. --- sw/source/filter/html/htmlfly.cxx | 10 ++-- sw/source/filter/html/wrthtml.cxx | 18 sw/source/filter/html/wrthtml.hxx |7 ++- sw/source/filter/ww8/wrtw8esh.cxx |4 +- sw/source/filter/ww8/wrtw8sty.cxx | 84 + sw/source/filter/ww8/wrtww8.cxx | 24 +-- sw/source/filter/ww8/wrtww8.hxx | 24 --- 7 files changed, 78 insertions(+), 93 deletions(-) diff --git a/sw/source/filter/html/htmlfly.cxx b/sw/source/filter/html/htmlfly.cxx index d278127..da7dcd0 100644 --- a/sw/source/filter/html/htmlfly.cxx +++ b/sw/source/filter/html/htmlfly.cxx @@ -1832,20 +1832,20 @@ void SwHTMLWriter::AddLinkTarget( const String& rURL ) SwPosition aPos( *pCurPam->GetPoint() ); if( pDoc->GotoOutline( aPos, aOutline ) ) { -sal_uLong nIdx = aPos.nNode.GetIndex(); +sal_uInt32 nIdx = aPos.nNode.GetIndex(); -sal_uInt16 nIns=0; -while( nIns < aOutlineMarkPoss.Count() && +sal_uInt32 nIns=0; +while( nIns < aOutlineMarkPoss.size() && aOutlineMarkPoss[nIns] < nIdx ) nIns++; -aOutlineMarkPoss.Insert( nIdx, nIns ); +aOutlineMarkPoss.insert( aOutlineMarkPoss.begin()+nIns, nIdx ); if( bEncoded ) { aURL.Erase( nPos, 2 ); aURL.SetChar( nPos-1, cMarkSeperator ); } -aOutlineMarks.Insert( new String( aURL ), nIns ); +aOutlineMarks.insert( aOutlineMarks.begin()+nIns, new String( aURL ) ); } } else if( sCmp.EqualsAscii( pMarkToText ) ) diff --git a/sw/source/filter/html/wrthtml.cxx b/sw/source/filter/html/wrthtml.cxx index 1d5975a..ae4fee6 100644 --- a/sw/source/filter/html/wrthtml.cxx +++ b/sw/source/filter/html/wrthtml.cxx @@ -402,11 +402,9 @@ sal_uLong SwHTMLWriter::WriteStream() if( aImplicitMarks.Count() ) aImplicitMarks.DeleteAndDestroy( sal_uInt16(0), aImplicitMarks.Count() ); -if( aOutlineMarks.Count() ) -aOutlineMarks.DeleteAndDestroy( sal_uInt16(0), aOutlineMarks.Count() ); +aOutlineMarks.clear(); -if( aOutlineMarkPoss.Count() ) -aOutlineMarkPoss.Remove( sal_uInt16(0), aOutlineMarkPoss.Count() ); +aOutlineMarkPoss.clear(); if( aNumRuleNames.Count() ) aNumRuleNames.DeleteAndDestroy( sal_uInt16(0), aNumRuleNames.Count() ); @@ -1066,18 +1064,18 @@ void SwHTMLWriter::OutBookmarks() pBookmark = (pMarkAccess->getMarksBegin() + nBkmkTabPos)->get(); } -sal_uInt16 nPos; -for( nPos = 0; nPos < aOutlineMarkPoss.Count() && +sal_uInt32 nPos; +for( nPos = 0; nPos < aOutlineMarkPoss.size() && aOutlineMarkPoss[nPos] < nNode; nPos++ ) ; -while( nPos < aOutlineMarkPoss.Count() && aOutlineMarkPoss[nPos] == nNode ) +while( nPos < aOutlineMarkPoss.size() && aOutlineMarkPoss[nPos] == nNode ) { -String sMark( *aOutlineMarks[nPos] ); +String sMark( aOutlineMarks[nPos] ); sMark.SearchAndReplaceAll( '?', '_' ); //
[Libreoffice] va_start without va_end
Hello, In svl/source/items/itemset.cxx b/svl/source/items/itemset.cxx, I noticed the use of va_start without va_end. I read it could create undefined behaviour, so I propose this simple patch. I propose this patch : diff --git a/svl/source/items/itemset.cxx b/svl/source/items/itemset.cxx index 3be762d..0ca7223 100644 --- a/svl/source/items/itemset.cxx +++ b/svl/source/items/itemset.cxx @@ -225,6 +225,7 @@ SfxItemSet::SfxItemSet( SfxItemPool& rPool, pArgs, sal::static_int_cast< sal_uInt16 >(nWh1), sal::static_int_cast< sal_uInt16 >(nWh2), sal::static_int_cast< sal_uInt16 >(nNull)); +va_end(pArgs); } } If it's ok, i can commit and push it on master. Julien. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Make dev-install without running smoketests
Lionel Elie Mamane wrote: > > On Wed, Aug 03, 2011 at 04:51:13PM +0100, Caolán McNamara wrote: > >> What are your configure options for me to see if I can replicate this ? >> Looks like evilness which should be fixed. > > I'm in a similar situation on master branch, but for me the smoketest > fails in officeconnection.cxx in function OfficeConnection::setUp(), > at that assertion: > > for (;;) { > try { > context_ = > css::uno::Reference< css::uno::XComponentContext >( > resolver->resolve( > rtl::OUString(RTL_CONSTASCII_USTRINGPARAM("uno:")) > + > desc + > rtl::OUString( > RTL_CONSTASCII_USTRINGPARAM( > ";urp;StarOffice.ComponentContext"))), > css::uno::UNO_QUERY_THROW); > break; > } catch (css::connection::NoConnectException &) {} > if (process_ != 0) { > TimeValue delay = { 1, 0 }; // 1 sec > CPPUNIT_ASSERT_EQUAL( > osl_Process_E_TimedOut, > osl_joinProcessWithTimeout(process_, &delay)); > } > } > > I changed the last lines to: > > oslProcessError > > if (process_ != 0) { > TimeValue delay = { 1, 0 }; // 1 sec > oslProcessError rv = osl_joinProcessWithTimeout(process_, &delay); > fprintf(stderr, "%d\n", rv); >CPPUNIT_ASSERT_EQUAL( > osl_Process_E_TimedOut, > rv); > } > > And then I see that the assertion succeeds several times, and then > finally fails, when the return value of the function call becomes > osl_Process_E_None (that is, 0). > > -- > Lionel > I noticed exactly the same thing but then I don't know how to go on the investigation. On IRC, Moggi told me this was a know issue but it doesn't happen on every machine. I'm curious to know if it happens on Mac or Windows. Julien. -- View this message in context: http://nabble.documentfoundation.org/Make-dev-install-without-running-smoketests-tp3221086p3223566.html Sent from the Dev mailing list archive at Nabble.com. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Vcl module compilation failed
Le 03/08/2011 14:10, Matúš Kukan a écrit : On 3 August 2011 14:03, serv serva wrote: De: Caolán McNamara Objet: Re: [Libreoffice] Vcl module compilation failed À: "Julien Nabet" Cc: libreoffice@lists.freedesktop.org Date: Mercredi 3 août 2011, 11h40 On Wed, 2011-08-03 at 08:06 +0200, Julien Nabet wrote: [ WARN ] !!! [ WARN ] !!! duplicate linked objects [ WARN ] !!! That's a warning, not an error. If your compilation failed, it didn't fail because of "salmain in linked twice", that happens for everyone. The pb is I've nothing else. "Make" in the "general directory" libo tells there's an error in vcl module and asks to go to vcl then make -r. I did this but had no error message. Is there a command I could use to have more info ? I've seen it also. Just run make clean and make -r again. But there is problem somewhere. It's not good that you can't see an error. Also in another module (don't remember which one) I ran make and nothing happened so I again have to run make clean&& make -r. Compilation is now ok ! So vcl module and automation seem ok now. (except the error doesn't appear clearly) I runned a "make check", I came back to my previous pb : officeconnection.cxx:140:Assertion Test name: N12_GLOBAL__N_14TestE::test setUp() failed - equality assertion failed - Expected: 2 - Actual : 0 So i can go on. Julien. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Make dev-install without running smoketests
On Wed, Aug 03, 2011 at 04:51:13PM +0100, Caolán McNamara wrote: > What are your configure options for me to see if I can replicate this ? > Looks like evilness which should be fixed. I'm in a similar situation on master branch, but for me the smoketest fails in officeconnection.cxx in function OfficeConnection::setUp(), at that assertion: for (;;) { try { context_ = css::uno::Reference< css::uno::XComponentContext >( resolver->resolve( rtl::OUString(RTL_CONSTASCII_USTRINGPARAM("uno:")) + desc + rtl::OUString( RTL_CONSTASCII_USTRINGPARAM( ";urp;StarOffice.ComponentContext"))), css::uno::UNO_QUERY_THROW); break; } catch (css::connection::NoConnectException &) {} if (process_ != 0) { TimeValue delay = { 1, 0 }; // 1 sec CPPUNIT_ASSERT_EQUAL( osl_Process_E_TimedOut, osl_joinProcessWithTimeout(process_, &delay)); } } I changed the last lines to: oslProcessError if (process_ != 0) { TimeValue delay = { 1, 0 }; // 1 sec oslProcessError rv = osl_joinProcessWithTimeout(process_, &delay); fprintf(stderr, "%d\n", rv); CPPUNIT_ASSERT_EQUAL( osl_Process_E_TimedOut, rv); } And then I see that the assertion succeeds several times, and then finally fails, when the return value of the function call becomes osl_Process_E_None (that is, 0). -- Lionel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] compiling libreoffice-3-4: copy of libcuuc as dependency of svtools fails (missing source operand)
Hi, A one-day-old clone of branch libreoffice-3-4 fails to build for me. Directory svtools fails with: R=/home/master/src/libreoffice && S=$R/libreoffice-3.4 && O=$S/solver/340/unxlngx6 && W=$O/workdir && mkdir -p $O/lib/ && /bin/cp --remove-destination --force --preserve=timestamps $O/lib/libicuuc.so /bin/cp: missing destination file operand after `/home/master/src/libreoffice/libreoffice-3.4/solver/340/unxlngx6/lib/libicuuc.so' Try `/bin/cp --help' for more information. make: *** [/home/master/src/libreoffice/libreoffice-3.4/solver/340/unxlngx6/lib/libicuuc.so] Error 1 So I did: cd ../icu build cd ../svtools R=/home/master/src/libreoffice && S=$R/libreoffice-3.4 && O=$S/solver/340/unxlngx6 && cp -vi --remove-destination --force --preserve=timestamps $S/clone/libs-extern-sys/icu/unxlngx6/lib/libicuuc.so $O/lib/libicuuc.so make -r and that worked, at least as far as directory svtools is concerned :) -- Lionel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Making a debian installation
On Wed, Aug 03, 2011 at 09:04:12PM +0530, swagat sharma wrote: > I want some details about the packaging of LO. Sorry if I am asking it in > the wrong channel but I did not get much information elsewhere. > I have created some macros and packed them as extensions. I have also > created some templates for company's purpose. How can I create standard > debian package? dh-make will create a template for a Debian package and get you started. As far as I understand, on the Debian package level, all you'll have to do is install the right files in the right place. With the template created for you by dh-make, you'd add the files that need to be installed in the file debian/PACKAGE_NAME.install (in the Debian source package), with format: source_file /directory/where/to/install To create the Debian binary package (the .deb), run "debuild" or "dpkg-buildpackage". General documentation about building Debian packages: http://www.debian.org/doc/maint-guide/ Other documents of use: http://www.debian.org/devel/ Mailing list for "beginners in Debian packaging": http://lists.debian.org/debian-mentors/ -- Lionel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [UNPUSHED][PATCH] Fix building bridge with -mavx
Ok so now I read the patch... Did it work for you ? 1/ it seems to me that it is missing a piece to put that HAVE_GCC_AVX in Env.Host.sh (see set_soenv.in) without that the second patch will never see it 2/ the makefile patches in the second patch are dmake makefile so the syntax is more something like .IF "$(HAVE_GCC_AVX)" == "YES" [...] .ENDIF take a look at HAVE_GCC_NO_LONG_DOUBLE and how it is set-up in configure.in/set_soenv.in and used later in the makefiles... Norbert ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] No Python scripts visible in UI - build 350m1 from master nightly 27/06/2011 for Mac OSX
Hi Caolán, *, On Wed, Aug 3, 2011 at 2:49 PM, Caolán McNamara wrote: > On Wed, 2011-06-29 at 00:04 +0200, Christian Lohmaier wrote: >> On Mon, Jun 27, 2011 at 5:14 PM, Alexander Thurgood >> wrote: >> > >> > MacOSX_10.6.7_Intel_no-moz/master/2011-06-27_08.26.21 >> > >> > There is no Python script menu entry in the Tools > Macros > Organise UI >> > menu, only Basic, Javascript and Beanshell. >> >> This is caused by... > > So, is this all ok now ? Yes, Internal python is now built on Mac. On both 3.4.2 as well as on master all the supplied samples work. ciao Christian ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [UNPUSHED][PATCH] Fix building bridge with -mavx
Dne 3.8.2011 17:43, Norbert Thiebaud napsal(a): I haven't read the patch, but you should use HAVE_GCC_AVX to stay similar to existing check like that. Ok, that is easy, I renamed the variable :) Tom >From 6ac456072b160f2786840b878c136408bac2559a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Tom=C3=A1=C5=A1=20Chv=C3=A1tal?= Date: Wed, 3 Aug 2011 13:52:13 +0200 Subject: [PATCH] Add check for avx support. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Tomáš Chvátal --- configure.in | 31 +++ 1 files changed, 31 insertions(+), 0 deletions(-) diff --git a/configure.in b/configure.in index cf308dc..3317d42 100755 --- a/configure.in +++ b/configure.in @@ -1885,6 +1885,37 @@ fi AC_SUBST(HAVE_LD_BSYMBOLIC_FUNCTIONS) dnl === +dnl Check for AVX +dnl === +AS_IF([test "x$AVX_CFLAGS" = "x"], [ + AS_IF([test "x$SUNCC" = "xyes" && test "x$AMD64_ABI" = "xno"], [ + AVX_CFLAGS="-xarch=avx" + ]) +], [ + AVX_CFLAGS="-mavx -Winline" +]) + +have_avx_intrinsics=no +AC_MSG_CHECKING(whether to use AVX intrinsics) +avx_save_CFLAGS=$CFLAGS +CFLAGS="$AVX_CFLAGS $CFLAGS" +AC_COMPILE_IFELSE([ +#if defined(__GNUC__) && (__GNUC__ < 4 || (__GNUC__ == 4 && __GNUC_MINOR__ < 5)) +# if !defined(__amd64__) && !defined(__x86_64__) +# error "Need GCC >= 4.5 for AVX intrinsics on x86" +# endif +#endif +#include +int main () { +__m256i a = _mm256_set1_epi32 (0), b = _mm256_set1_epi32 (0), c; +c = _mm256_permute2f128_si256 (a, b, 0); +return 0; +}], have_avx_intrinsics=yes) +CFLAGS=$avx_save_CFLAGS +AC_MSG_RESULT([$have_avx_intrinsics]) +AM_CONDITIONAL(HAVE_GCC_AVX, [test "x$have_avx_intrinsics" = "xyes"]) + +dnl === dnl Set up a different compiler to produce tools to run on the build dnl machine when doing cross-compilation dnl === -- 1.7.3.4 >From 735f0f2d175fcfd656468fef25596753bea9c926 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Tom=C3=A1=C5=A1=20Chv=C3=A1tal?= Date: Mon, 1 Aug 2011 17:41:29 +0200 Subject: [PATCH] Fix segmentation fault when linking -mavx on new CPUs is enabled by explicitly adding -mno-avx to CXXFLAGS. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Tomáš Chvátal --- .../source/cpp_uno/gcc3_linux_intel/makefile.mk|6 ++ .../source/cpp_uno/gcc3_linux_x86-64/makefile.mk |6 ++ bridges/source/cpp_uno/mingw_intel/makefile.mk |6 ++ 3 files changed, 18 insertions(+), 0 deletions(-) diff --git a/bridges/source/cpp_uno/gcc3_linux_intel/makefile.mk b/bridges/source/cpp_uno/gcc3_linux_intel/makefile.mk index beedf9d..d0acc23 100644 --- a/bridges/source/cpp_uno/gcc3_linux_intel/makefile.mk +++ b/bridges/source/cpp_uno/gcc3_linux_intel/makefile.mk @@ -52,6 +52,12 @@ CFLAGS += -DLEAK_STATIC_DATA # work with the .cxx sources in this directory: CFLAGSCXX += -fno-omit-frame-pointer -fno-strict-aliasing +# In case the compiler supports AVX this code segfaults so specifically turn +# it off. +ifdef HAVE_GCC_AVX +CFLAGSCXX+= -mno-avx +endif + CFLAGSNOOPT=-O0 SLOFILES= \ diff --git a/bridges/source/cpp_uno/gcc3_linux_x86-64/makefile.mk b/bridges/source/cpp_uno/gcc3_linux_x86-64/makefile.mk index dd35432..ee934e0 100644 --- a/bridges/source/cpp_uno/gcc3_linux_x86-64/makefile.mk +++ b/bridges/source/cpp_uno/gcc3_linux_x86-64/makefile.mk @@ -52,6 +52,12 @@ CFLAGS += -DLEAK_STATIC_DATA # work with the .cxx sources in this directory: CFLAGSCXX += -fno-omit-frame-pointer -fno-strict-aliasing +# In case the compiler supports AVX this code segfaults so specifically turn +# it off. +ifdef HAVE_GCC_AVX +CFLAGSCXX+= -mno-avx +endif + SLOFILES= \ $(SLO)$/abi.obj \ $(SLO)$/except.obj \ diff --git a/bridges/source/cpp_uno/mingw_intel/makefile.mk b/bridges/source/cpp_uno/mingw_intel/makefile.mk index 2c7a8da..189cb15 100644 --- a/bridges/source/cpp_uno/mingw_intel/makefile.mk +++ b/bridges/source/cpp_uno/mingw_intel/makefile.mk @@ -51,6 +51,12 @@ CFLAGS += -DBROKEN_ALLOCA # work with the .cxx sources in this directory: CFLAGSCXX += -fno-omit-frame-pointer +# In case the compiler supports AVX this code segfaults so specifically turn +# it off. +ifdef HAVE_GCC_AVX +CFLAGSCXX+= -mno-avx +endif + NOOPTFILES= \ $(SLO)$/uno2cpp.obj -- 1.7.3.4 ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] [PATCH] Consistent paragraph border handling for HTML export
Attached patches remove the special case where border settings tab disappears from paragraph settings dialog when Writer is in HTML editor mode and "HTML 3.2" is selected as format option for HTML export. The second patch enables exporting the borders in this mode as well. Although CSS styles were not in use when HTML 3.2 was standardized I think these days it is reasonable to export the styles anyway. Harri From 1ac00e10e3eac9341e9edcc3d249db6ecf35fc0e Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Harri=20Pitk=C3=A4nen?= Date: Mon, 1 Aug 2011 22:12:21 +0300 Subject: [PATCH 1/2] Enable paragraph border settings in HTML editor for all HTML variants Previously border settings were not available in paragraph format dialog when HTML export format was "HTML 3.2". This was confusing since documents with borders could still be imported and you could even access the border dialog with suitably configured shortcut key. --- sw/source/ui/chrdlg/pardlg.cxx | 12 sw/source/ui/config/viewopt.cxx |6 +++--- sw/source/ui/fmtui/tmpdlg.cxx |2 -- 3 files changed, 7 insertions(+), 13 deletions(-) diff --git a/sw/source/ui/chrdlg/pardlg.cxx b/sw/source/ui/chrdlg/pardlg.cxx index 5e85e3f..359e619 100644 --- a/sw/source/ui/chrdlg/pardlg.cxx +++ b/sw/source/ui/chrdlg/pardlg.cxx @@ -161,14 +161,10 @@ SwParaDlg::SwParaDlg(Window *pParent, { RemoveTabPage(TP_BACKGROUND); } -if(!bHtmlMode || (nHtmlMode & HTMLMODE_PARA_BORDER)) -{ -OSL_ENSURE(pFact->GetTabPageCreatorFunc( RID_SVXPAGE_BORDER ), "GetTabPageCreatorFunc fail!"); -OSL_ENSURE(pFact->GetTabPageRangesFunc( RID_SVXPAGE_BORDER ), "GetTabPageRangesFunc fail!"); -AddTabPage(TP_BORDER, pFact->GetTabPageCreatorFunc( RID_SVXPAGE_BORDER ), pFact->GetTabPageRangesFunc( RID_SVXPAGE_BORDER ) ); -} -else -RemoveTabPage(TP_BORDER); + +OSL_ENSURE(pFact->GetTabPageCreatorFunc( RID_SVXPAGE_BORDER ), "GetTabPageCreatorFunc fail!"); +OSL_ENSURE(pFact->GetTabPageRangesFunc( RID_SVXPAGE_BORDER ), "GetTabPageRangesFunc fail!"); +AddTabPage(TP_BORDER, pFact->GetTabPageCreatorFunc( RID_SVXPAGE_BORDER ), pFact->GetTabPageRangesFunc( RID_SVXPAGE_BORDER ) ); } if (nDefPage) diff --git a/sw/source/ui/config/viewopt.cxx b/sw/source/ui/config/viewopt.cxx index afe6d06..5ca8f19 100644 --- a/sw/source/ui/config/viewopt.cxx +++ b/sw/source/ui/config/viewopt.cxx @@ -338,18 +338,18 @@ sal_uInt16 GetHtmlMode(const SwDocShell* pShell) switch ( rHtmlOpt.GetExportMode() ) { case HTML_CFG_MSIE_40: -nRet |= HTMLMODE_PARA_BORDER|HTMLMODE_SMALL_CAPS| +nRet |= HTMLMODE_SMALL_CAPS| HTMLMODE_SOME_STYLES| HTMLMODE_FULL_STYLES|HTMLMODE_GRAPH_POS| HTMLMODE_FULL_ABS_POS|HTMLMODE_SOME_ABS_POS; break; case HTML_CFG_NS40: -nRet |= HTMLMODE_PARA_BORDER|HTMLMODE_SOME_STYLES| +nRet |= HTMLMODE_SOME_STYLES| HTMLMODE_FRM_COLUMNS|HTMLMODE_BLINK|HTMLMODE_GRAPH_POS| HTMLMODE_SOME_ABS_POS; break; case HTML_CFG_WRITER: -nRet |= HTMLMODE_PARA_BORDER|HTMLMODE_SMALL_CAPS| +nRet |= HTMLMODE_SMALL_CAPS| HTMLMODE_SOME_STYLES| HTMLMODE_FRM_COLUMNS|HTMLMODE_FULL_STYLES| HTMLMODE_BLINK|HTMLMODE_DROPCAPS|HTMLMODE_GRAPH_POS| diff --git a/sw/source/ui/fmtui/tmpdlg.cxx b/sw/source/ui/fmtui/tmpdlg.cxx index c802f1b..26a02f8 100644 --- a/sw/source/ui/fmtui/tmpdlg.cxx +++ b/sw/source/ui/fmtui/tmpdlg.cxx @@ -218,8 +218,6 @@ SwTemplateDlg::SwTemplateDlg(Window* pParent, RemoveTabPage(TP_BACKGROUND); RemoveTabPage(TP_DROPCAPS); } -if(!(nHtmlMode & HTMLMODE_PARA_BORDER)) -RemoveTabPage(TP_BORDER); } else { -- 1.7.5.4 From 6da9496d745f13973f3e7841ee7ed2cad2fe554e Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Harri=20Pitk=C3=A4nen?= Date: Wed, 3 Aug 2011 18:09:11 +0300 Subject: [PATCH 2/2] Enable export of basic CSS styles in HTML 3.2 mode Although "style" does not seem to be valid attribute in HTML 3.2 it is harmless to export it. This simplifies the code and makes the exported HTML more useful in modern browsers. --- sw/source/ui/config/viewopt.cxx |7 ++- 1 files changed, 2 insertions(+), 5 deletions(-) diff --git a/sw/source/ui/config/viewopt.cxx b/sw/source/ui/config/viewopt.cxx index 5ca8f19..7cdc31a 100644 --- a/sw/source/ui/config/viewopt.cxx +++ b/sw/source/ui/config/viewopt.cxx @@ -333,24 +333,21 @@ sal_uInt16 GetHtmlMode(const SwDocShell* pShell) sal_uInt16 nRet = 0; if(!pShell || PTR_CAST(SwWebDocShell, pShell)) { -nRet = HTMLMODE_
Re: [Libreoffice] Make dev-install without running smoketests
On Wed, 2011-08-03 at 14:49 +0200, Thomas Arnhold wrote: > > tools->options->security->macro security->low : open > > libo/smoketestdoc/PLATFORM/bin/smoketestdoc.sxw and click the "run" > > button. > > The error in smoketest is (write test results): > > assertion caught: Error: sentence indices out of range From File > /home/thomas/temp/bootstrap/editeng/source/editeng/impedit2.cxx at Line 1632 > write_status > assertion caught: Error: sentence indices out of range From File > /home/thomas/temp/bootstrap/editeng/source/editeng/impedit2.cxx at Line 1632 > > I have to disable Java and Extensions under Smoketest options then it > runs clean. What are your configure options for me to see if I can replicate this ? Looks like evilness which should be fixed. C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [UNPUSHED][PATCH] Fix building bridge with -mavx
2011/8/3 Tomáš Chvátal : > Dne 3.8.2011 13:27, Caolán McNamara napsal(a): >> >> On Wed, 2011-08-03 at 12:11 +0100, Caolán McNamara wrote: >>> >>> On Wed, 2011-08-03 at 12:47 +0200, Tomáš Chvátal wrote: Attaching the updated patch that adds it only where it should really be needed. >>> >>> seems reasonable, pushed. >> >> *bzzt*, x86_64 gcc 4.3.4 doesn't support it, so breaks the build on >> those buildbots. >> >> I guess need to instead add a configure check for it, stuff it into an >> env variable if supported, and add that env variable to the makefile.mks >> instead. >> >> C. >> > Darn, > Ok first part is the configure.in check, I think the attachment should do > the trick. > > But I am not entirely sure how to do the check in plain GNUMake. > I wrote it with ifdef USE_AVX and I hope this is the way it should work :) I haven't read the patch, but you should use HAVE_GCC_AVX to stay similar to existing check like that. Norbert ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] Making a debian installation
Hi, I want some details about the packaging of LO. Sorry if I am asking it in the wrong channel but I did not get much information elsewhere. I have created some macros and packed them as extensions. I have also created some templates for company's purpose. How can I create standard debian package? -- Swagat ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] How to set fillcolor and linecolor to a SdrObject?
Hi Caolán, Caolán McNamara schrieb: On Wed, 2011-08-03 at 00:23 +0200, Regina Henschel wrote: But I do not know how to make the object transparent. Especially, I do not know how to use these "itempool"-thing. Does pObject->SetMergedItem(XFillTransparenceItem(some_value)); do anything useful for you. opengroking around for stuff like that might help anyway Yes, thanks. That works perfectly. The pop-up list looks good now. Next step will be the preview. I will try the same trick there. Kind regards Regina ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] No Python scripts visible in UI - build 350m1 from master nightly 27/06/2011 for Mac OSX
On Wed, 2011-06-29 at 00:04 +0200, Christian Lohmaier wrote: > Hi Alexander, *, > > On Mon, Jun 27, 2011 at 5:14 PM, Alexander Thurgood > wrote: > > > > MacOSX_10.6.7_Intel_no-moz/master/2011-06-27_08.26.21 > > > > There is no Python script menu entry in the Tools > Macros > Organise UI > > menu, only Basic, Javascript and Beanshell. > > This is caused by... So, is this all ok now ? Under master Linux it all appears to works well, and I see f10464a1060139c84f9c1671de088560bf2b4010 fixed up some whitespace in the meantime. C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Make dev-install without running smoketests
tools->options->security->macro security->low : open libo/smoketestdoc/PLATFORM/bin/smoketestdoc.sxw and click the "run" button. The error in smoketest is (write test results): assertion caught: Error: sentence indices out of range From File /home/thomas/temp/bootstrap/editeng/source/editeng/impedit2.cxx at Line 1632 write_status assertion caught: Error: sentence indices out of range From File /home/thomas/temp/bootstrap/editeng/source/editeng/impedit2.cxx at Line 1632 I have to disable Java and Extensions under Smoketest options then it runs clean. Thomas ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [UNPUSHED][PATCH] Fix building bridge with -mavx
Dne 3.8.2011 13:27, Caolán McNamara napsal(a): On Wed, 2011-08-03 at 12:11 +0100, Caolán McNamara wrote: On Wed, 2011-08-03 at 12:47 +0200, Tomáš Chvátal wrote: Attaching the updated patch that adds it only where it should really be needed. seems reasonable, pushed. *bzzt*, x86_64 gcc 4.3.4 doesn't support it, so breaks the build on those buildbots. I guess need to instead add a configure check for it, stuff it into an env variable if supported, and add that env variable to the makefile.mks instead. C. Darn, Ok first part is the configure.in check, I think the attachment should do the trick. But I am not entirely sure how to do the check in plain GNUMake. I wrote it with ifdef USE_AVX and I hope this is the way it should work :) Tom >From 78ebd6077193a83f9be01fc5c2ee29558e840187 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Tom=C3=A1=C5=A1=20Chv=C3=A1tal?= Date: Wed, 3 Aug 2011 13:52:13 +0200 Subject: [PATCH] Add check for avx support. --- configure.in | 31 +++ 1 files changed, 31 insertions(+), 0 deletions(-) diff --git a/configure.in b/configure.in index cf308dc..7dc6fa3 100755 --- a/configure.in +++ b/configure.in @@ -1885,6 +1885,37 @@ fi AC_SUBST(HAVE_LD_BSYMBOLIC_FUNCTIONS) dnl === +dnl Check for AVX +dnl === +AS_IF([test "x$AVX_CFLAGS" = "x"], [ + AS_IF([test "x$SUNCC" = "xyes" && test "x$AMD64_ABI" = "xno"], [ + AVX_CFLAGS="-xarch=avx" + ]) +], [ + AVX_CFLAGS="-mavx -Winline" +]) + +have_avx_intrinsics=no +AC_MSG_CHECKING(whether to use AVX intrinsics) +avx_save_CFLAGS=$CFLAGS +CFLAGS="$AVX_CFLAGS $CFLAGS" +AC_COMPILE_IFELSE([ +#if defined(__GNUC__) && (__GNUC__ < 4 || (__GNUC__ == 4 && __GNUC_MINOR__ < 5)) +# if !defined(__amd64__) && !defined(__x86_64__) +# error "Need GCC >= 4.5 for AVX intrinsics on x86" +# endif +#endif +#include +int main () { +__m256i a = _mm256_set1_epi32 (0), b = _mm256_set1_epi32 (0), c; +c = _mm256_permute2f128_si256 (a, b, 0); +return 0; +}], have_avx_intrinsics=yes) +CFLAGS=$avx_save_CFLAGS +AC_MSG_RESULT([$have_avx_intrinsics]) +AM_CONDITIONAL(USE_AVX, [test "x$have_avx_intrinsics" = "xyes"]) + +dnl === dnl Set up a different compiler to produce tools to run on the build dnl machine when doing cross-compilation dnl === -- 1.7.3.4 >From 2f566e2041ddc71039362f1761e3f87cbe999c99 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Tom=C3=A1=C5=A1=20Chv=C3=A1tal?= Date: Mon, 1 Aug 2011 17:41:29 +0200 Subject: [PATCH] Fix segmentation fault when linking -mavx on new CPUs is enabled by explicitly adding -mno-avx to CXXFLAGS. --- .../source/cpp_uno/gcc3_linux_intel/makefile.mk|6 ++ .../source/cpp_uno/gcc3_linux_x86-64/makefile.mk |6 ++ bridges/source/cpp_uno/mingw_intel/makefile.mk |6 ++ 3 files changed, 18 insertions(+), 0 deletions(-) diff --git a/bridges/source/cpp_uno/gcc3_linux_intel/makefile.mk b/bridges/source/cpp_uno/gcc3_linux_intel/makefile.mk index beedf9d..919cc3f 100644 --- a/bridges/source/cpp_uno/gcc3_linux_intel/makefile.mk +++ b/bridges/source/cpp_uno/gcc3_linux_intel/makefile.mk @@ -52,6 +52,12 @@ CFLAGS += -DLEAK_STATIC_DATA # work with the .cxx sources in this directory: CFLAGSCXX += -fno-omit-frame-pointer -fno-strict-aliasing +# In case the compiler supports AVX this code segfaults so specifically turn +# it off. +ifdef USE_AVX +CFLAGSCXX+= -mno-avx +endif + CFLAGSNOOPT=-O0 SLOFILES= \ diff --git a/bridges/source/cpp_uno/gcc3_linux_x86-64/makefile.mk b/bridges/source/cpp_uno/gcc3_linux_x86-64/makefile.mk index dd35432..0bf5543 100644 --- a/bridges/source/cpp_uno/gcc3_linux_x86-64/makefile.mk +++ b/bridges/source/cpp_uno/gcc3_linux_x86-64/makefile.mk @@ -52,6 +52,12 @@ CFLAGS += -DLEAK_STATIC_DATA # work with the .cxx sources in this directory: CFLAGSCXX += -fno-omit-frame-pointer -fno-strict-aliasing +# In case the compiler supports AVX this code segfaults so specifically turn +# it off. +ifdef USE_AVX +CFLAGSCXX+= -mno-avx +endif + SLOFILES= \ $(SLO)$/abi.obj \ $(SLO)$/except.obj \ diff --git a/bridges/source/cpp_uno/mingw_intel/makefile.mk b/bridges/source/cpp_uno/mingw_intel/makefile.mk index 2c7a8da..05a2400 100644 --- a/bridges/source/cpp_uno/mingw_intel/makefile.mk +++ b/bridges/source/cpp_uno/mingw_intel/makefile.mk @@ -51,6 +51,12 @@ CFLAGS += -DBROKEN_ALLOCA # work with the .cxx sources in this directory: CFLAGSCXX += -fno-omit-frame-pointer +# In case the compiler supports AVX this code segfaults so specifically turn +# it off. +ifdef USE_AVX +CFLAGSCXX+= -mno-avx +endif + NOOPTFILES= \ $(SLO)$/uno2cpp.obj -- 1.7.3.4 ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org htt
Re: [Libreoffice] Vcl module compilation failed
On 3 August 2011 14:03, serv serva wrote: >> De: Caolán McNamara >> Objet: Re: [Libreoffice] Vcl module compilation failed >> À: "Julien Nabet" >> Cc: libreoffice@lists.freedesktop.org >> Date: Mercredi 3 août 2011, 11h40 >> On Wed, 2011-08-03 at 08:06 +0200, >> Julien Nabet wrote: >> > [ WARN ] !!! >> > [ WARN ] !!! duplicate linked >> objects >> > [ WARN ] !!! >> >> That's a warning, not an error. If your compilation failed, >> it didn't >> fail because of "salmain in linked twice", that happens for >> everyone. > The pb is I've nothing else. "Make" in the "general directory" libo tells > there's an error in vcl module and asks to go to vcl then make -r. > I did this but had no error message. > Is there a command I could use to have more info ? > I've seen it also. Just run make clean and make -r again. But there is problem somewhere. It's not good that you can't see an error. Also in another module (don't remember which one) I ran make and nothing happened so I again have to run make clean && make -r. Matus ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Vcl module compilation failed
> De: Caolán McNamara > Objet: Re: [Libreoffice] Vcl module compilation failed > À: "Julien Nabet" > Cc: libreoffice@lists.freedesktop.org > Date: Mercredi 3 août 2011, 11h40 > On Wed, 2011-08-03 at 08:06 +0200, > Julien Nabet wrote: > > Hello, > > > > Here is what I get if I try to compile vcl : > > maryline@maryline:~/compile-libreoffice/libo/vcl$ rm > -rf unxlngi6/ && > > make -r > > CxxObject/vcl/source/salmain/salmain.o is linked in by > > > Library/libvcllo.so StaticLibrary/libvclmain.a > > > /home/maryline/compile-libreoffice/libo/solenv/gbuild/extensions/final_ObjectOwner.mk:43: > > > > > [ WARN ] !!! > > [ WARN ] !!! duplicate linked > objects > > [ WARN ] !!! > > That's a warning, not an error. If your compilation failed, > it didn't > fail because of "salmain in linked twice", that happens for > everyone. The pb is I've nothing else. "Make" in the "general directory" libo tells there's an error in vcl module and asks to go to vcl then make -r. I did this but had no error message. Is there a command I could use to have more info ? Julien. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Reporting an error when loading of a component fails
On 03/08/11 11:08, Noel Power wrote: ( /me must remember to reply-all ) Hi Lubos On 02/08/11 13:58, Lubos Lunak wrote: Hello, last week .docx support was broken because of a mistake during the gbuild changes and I wasted quite some time finding out what was wrong. As something similar had happened to me already same for me ( too many times ) before I'd like to apply the attached patches. However I have some issues with it: - the error dialog that pops up just says 'General error. General input/output error.', which is completely unhelpful, however I don't feel like digging more for how to pass the information as far as the place where the dialog is shown, hmm I think it would be nicer to handle at least the CannotActivateFactoryException problems from the filter instatiations and pop a dialog ( debug output to terminal especially on weirdy windows is not the best way to see this problem ), I vaguely recall the twisted mechanism for getting some dialogs to pop with the interaction handler stuff, I will have a look maybe this patch is useful to you, it has the disadvantage of creating the needed error dialog but then it also pops up the "General Error" one after that. I contemplated throwing ( well I actually half implemented it but gave up ) the exception up to allow SfxBaseModel::load to handle it but it appears there is all sorts of stuff inbetween that runs ( not expecting an exception ). Not passing an exception but setting a new ( or reusing an existing ) ERROR_IO_xxx ( or something ) also has problems like you can't easily ( or at all ) pass the error message, also I suppose you could use such a new error type to suppress the second 'General Error' but in the end it all just seemed too much trouble for what probably is just something you want in debug ( I guess ) So.. anyway patch attached if you think it might be useful, feel free to use/disregard/change totally etc. Noel diff --git sfx2/source/doc/objstor.cxx sfx2/source/doc/objstor.cxx index 31ee8f0..0716a48 100644 --- sfx2/source/doc/objstor.cxx +++ sfx2/source/doc/objstor.cxx @@ -75,6 +75,8 @@ #include #include +#include + #include #include #include @@ -102,6 +104,7 @@ #include #include #include +#include #include #include #include @@ -153,6 +156,33 @@ using namespace ::cppu; namespace css = ::com::sun::star; //= +class FilterActivationHandlerHelper +{ +uno::Reference< XInteractionHandler > mxHandler; +public: +FilterActivationHandlerHelper( SfxMedium& rMedium) +{ +mxHandler = rMedium.GetInteractionHandler(); +} +void handleException( const loader::CannotActivateFactoryException& e ) +{ +// not sure if we want this in non debug builds +//#if OSL_DEBUG_LEVEL > 0 +if ( mxHandler.is() ) +{ +comphelper::OInteractionRequest* pReq = new comphelper::OInteractionRequest( uno::makeAny( e ) ); +pReq->addContinuation( new comphelper::OInteractionApprove() ); +uno::Reference< task::XInteractionRequest > xReq( pReq ); +mxHandler->handle( xReq ); +} +else +{ +OSL_TRACE("Caught CannotActivateFactoryException: %s", rtl::OUStringToOString( e.Message, RTL_TEXTENCODING_UTF8 ).getStr()); +} +//#endif +} +}; + void impl_addToModelCollection(const css::uno::Reference< css::frame::XModel >& xModel) { if (!xModel.is()) @@ -2186,7 +2216,13 @@ sal_Bool SfxObjectShell::ImportFrom( SfxMedium& rMedium, bool bInsert ) try{ xLoader = uno::Reference< document::XFilter > ( xFilterFact->createInstanceWithArguments( aFilterName, uno::Sequence < uno::Any >() ), uno::UNO_QUERY ); -}catch(const uno::Exception&) +} +catch(const loader::CannotActivateFactoryException& e ) +{ +FilterActivationHandlerHelper helper( rMedium ); +helper.handleException( e ); +} +catch(const uno::Exception& e) { xLoader.clear(); } } if ( xLoader.is() ) @@ -2281,6 +2317,11 @@ sal_Bool SfxObjectShell::ExportTo( SfxMedium& rMedium ) try{ xExporter = uno::Reference< document::XExporter > ( xFilterFact->createInstanceWithArguments( aFilterName, uno::Sequence < uno::Any >() ), uno::UNO_QUERY ); +} +catch(const loader::CannotActivateFactoryException& e ) +{ +FilterActivationHandlerHelper helper( rMedium ); +helper.handleException( e ); }catch(const uno::Exception&) { xExporter.clear(); } } ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [PATCH] MSWordExportBase::ExportDocument: check if we have a DocShell
Hi Miklos, On 03/08/11 11:47, Miklos Vajna wrote: On Tue, Aug 02, 2011 at 05:51:49PM +0100, Noel Power wrote: On 02/08/11 16:46, Miklos Vajna wrote: I guess the reason is that the doc shell points to the whole document, while you usually export only part of the doc during copy&paste export. 'copy&paste export' is it a new thing you can't normally do from writer ? ( not familiar with it anyway ) No, it's not new at all. The only recent change was that you changed SwMSConvertControls to require a valid SwDocShell. doh, yes of course I made a change that requires the doc shell to be used immediately (even for export ) [...] yes of course and no problem for me with the fix for that OK, I've pushed the patch. cool, I'd like to reorganize that at some stage soon so we can remove that check you put in. Noel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] [UNPUSHED][PATCH] Fix building bridge with -mavx
On Wed, 2011-08-03 at 12:11 +0100, Caolán McNamara wrote: > On Wed, 2011-08-03 at 12:47 +0200, Tomáš Chvátal wrote: > > Attaching the updated patch that adds it only where it should really be > > needed. > > seems reasonable, pushed. *bzzt*, x86_64 gcc 4.3.4 doesn't support it, so breaks the build on those buildbots. I guess need to instead add a configure check for it, stuff it into an env variable if supported, and add that env variable to the makefile.mks instead. C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] [PUSHED][PATCH] Fix building bridge with -mavx
On Wed, 2011-08-03 at 12:47 +0200, Tomáš Chvátal wrote: > Attaching the updated patch that adds it only where it should really be > needed. seems reasonable, pushed. C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] [PATCH] Add missing libdir path for qt4 detection
>From d1c187bd6005339ba260c166b2539791628e4702 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Tom=C3=A1=C5=A1=20Chv=C3=A1tal?= Date: Wed, 3 Aug 2011 13:02:38 +0200 Subject: [PATCH] Add one more libdir path for detection with qt4 on 32b platform to be same with 64b. Fixes gentoo bug #377377. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Tomáš Chvátal --- configure.in |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/configure.in b/configure.in index f4363db..cf308dc 100755 --- a/configure.in +++ b/configure.in @@ -7310,7 +7310,7 @@ KDE_GLIB_LIBS="" KDE_HAVE_GLIB="" if test "$test_kde4" = "yes" -a "$ENABLE_KDE4" = "TRUE" ; then qt_incdirs="$QT4INC /usr/include/qt4 /usr/include $x_includes" -qt_libdirs="$QT4LIB /usr/lib $x_libraries" +qt_libdirs="$QT4LIB /usr/lib/qt4 /usr/lib $x_libraries" kde_incdirs="/usr/include /usr/include/kde4 $x_includes" kde_libdirs="/usr/lib /usr/lib/kde4 /usr/lib/kde4/devel $x_libraries" -- 1.7.3.4 ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] [GSoC] [rtfimport] 10th week
Hi, Previous week starts here: http://cgit.freedesktop.org/~vmiklos/lo-gsoc/tree/README#n1128 In short, it was about: - fixed the importer to handle test documents from the existing unit test - fixed/implemented the following bugs: fdo#36877, fdo#36922, fdo#35985, fdo#36089 Miklos pgp6l0q1JzcaO.pgp Description: PGP signature ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [PATCH] Fix building bridge with -mavx
Dne 2.8.2011 14:50, Caolán McNamara napsal(a): On Tue, 2011-08-02 at 13:18 +0200, Tomas Chvatal wrote: Libreoffice bridge code won't compile on new CPUs which add -mavx to cxxflags with --march=native. As first idea for the fix is actual ban for the -mavx cxxflag by adding -nomavx. Isn't that a x86 (and x86_64 ?) only option, i.e. no need for it except in the _intel and _x86_64 bridges ? C. Yeah you are right, it is only for 64b or intel. I just threw it in with sed :) Attaching the updated patch that adds it only where it should really be needed. Tom >From 38f3459c8b9f24f0e41ae9eff79e5fff85115e35 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Tom=C3=A1=C5=A1=20Chv=C3=A1tal?= Date: Mon, 1 Aug 2011 17:41:29 +0200 Subject: [PATCH] Fix segmentation fault when linking -mavx on new CPUs is enabled by explicitly adding -mno-avx to CXXFLAGS. --- .../source/cpp_uno/gcc3_linux_intel/makefile.mk|2 +- .../source/cpp_uno/gcc3_linux_x86-64/makefile.mk |2 +- bridges/source/cpp_uno/mingw_intel/makefile.mk |2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/bridges/source/cpp_uno/gcc3_linux_intel/makefile.mk b/bridges/source/cpp_uno/gcc3_linux_intel/makefile.mk index beedf9d..eae6df4 100644 --- a/bridges/source/cpp_uno/gcc3_linux_intel/makefile.mk +++ b/bridges/source/cpp_uno/gcc3_linux_intel/makefile.mk @@ -50,7 +50,7 @@ CFLAGS += -DLEAK_STATIC_DATA # In case someone enabled the non-standard -fomit-frame-pointer which does not # work with the .cxx sources in this directory: -CFLAGSCXX += -fno-omit-frame-pointer -fno-strict-aliasing +CFLAGSCXX += -fno-omit-frame-pointer -fno-strict-aliasing -mno-avx CFLAGSNOOPT=-O0 diff --git a/bridges/source/cpp_uno/gcc3_linux_x86-64/makefile.mk b/bridges/source/cpp_uno/gcc3_linux_x86-64/makefile.mk index dd35432..ad5de1a 100644 --- a/bridges/source/cpp_uno/gcc3_linux_x86-64/makefile.mk +++ b/bridges/source/cpp_uno/gcc3_linux_x86-64/makefile.mk @@ -50,7 +50,7 @@ CFLAGS += -DLEAK_STATIC_DATA # In case someone enabled the non-standard -fomit-frame-pointer which does not # work with the .cxx sources in this directory: -CFLAGSCXX += -fno-omit-frame-pointer -fno-strict-aliasing +CFLAGSCXX += -fno-omit-frame-pointer -fno-strict-aliasing -mno-avx SLOFILES= \ $(SLO)$/abi.obj \ diff --git a/bridges/source/cpp_uno/mingw_intel/makefile.mk b/bridges/source/cpp_uno/mingw_intel/makefile.mk index 2c7a8da..a3c2b2a 100644 --- a/bridges/source/cpp_uno/mingw_intel/makefile.mk +++ b/bridges/source/cpp_uno/mingw_intel/makefile.mk @@ -49,7 +49,7 @@ CFLAGS += -DBROKEN_ALLOCA # In case someone enabled the non-standard -fomit-frame-pointer which does not # work with the .cxx sources in this directory: -CFLAGSCXX += -fno-omit-frame-pointer +CFLAGSCXX += -fno-omit-frame-pointer -mno-avx NOOPTFILES= \ $(SLO)$/uno2cpp.obj -- 1.7.3.4 ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [PATCH] MSWordExportBase::ExportDocument: check if we have a DocShell
On Tue, Aug 02, 2011 at 05:51:49PM +0100, Noel Power wrote: > On 02/08/11 16:46, Miklos Vajna wrote: > > I guess the reason is that the doc shell points to the whole document, > > while you usually export only part of the doc during copy&paste export. > 'copy&paste export' is it a new thing you can't normally do from writer > ? ( not familiar with it anyway ) No, it's not new at all. The only recent change was that you changed SwMSConvertControls to require a valid SwDocShell. Given that SwMSConvertControls is called by MSWordExportBase, and copy&paste export (when RTF format is selected and which uses MSWordExportBase) does not provide a SwDocShell (AFAIK it never did, but it wasn't a problem till now). > > And yes, I don't think missing the control export is a big issue (at > > least it's a smaller one thatn the current instant crash), > yes of course and no problem for me with the fix for that OK, I've pushed the patch. pgpXuANBFZqAV.pgp Description: PGP signature ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] No Edit entry in context menu of a chart: bug or intentional ?
On Tue, 2011-06-21 at 22:07 +0200, Jean-Baptiste Faure wrote: > Hi all, > > In LibO 3.4.1 rc1 the context menu of a chart inserted in a speadsheet > does not show the menu entry Edit as it does in LibO 3.3.x. In addition, > the menu Edit > Object is grayed when you select a chart. Seems the same for me with current 3.4.X and 3.3.X with an edit option available for both on right click of a freshly inserted chart. C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Reporting an error when loading of a component fails
On Tue, 2011-08-02 at 14:58 +0200, Lubos Lunak wrote: > - osl_loadModule*() functions are plain C functions from stable API, so no > overloading to add the extra argument for the error message, the best I've > come up with was simply a wrapper around dlerror(), which I admit is not very > nice API either, would there be any better approach? Has the obvious issue of an gap between error and retrieving the error and some other thread churning away in between, how about... A oslModule is a void* and we're currently just directly casting the return from dlopen (and the windows equivalent) to a oslModule Maybe one possibility is instead a private struct like struct unx_impl_Module { void *m_pRealHandle; const char *m_pLastError; }; in osl/unx/module.c, malloc one of those and return *that* as the oslModule from the osl_loadModule* family. Free it in osl_unloadModule. Do the fixup in the various methods to cast oslModule to unx_impl_Module and get the realhandle out of it for the dl* family. Now can add a new toplevel osl c-function to check for error. oslModule stays a void* and all should remain abi-groovy methinks ? C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Reporting an error when loading of a component fails
( /me must remember to reply-all ) Hi Lubos On 02/08/11 13:58, Lubos Lunak wrote: Hello, last week .docx support was broken because of a mistake during the gbuild changes and I wasted quite some time finding out what was wrong. As something similar had happened to me already same for me ( too many times ) before I'd like to apply the attached patches. However I have some issues with it: - the error dialog that pops up just says 'General error. General input/output error.', which is completely unhelpful, however I don't feel like digging more for how to pass the information as far as the place where the dialog is shown, hmm I think it would be nicer to handle at least the CannotActivateFactoryException problems from the filter instatiations and pop a dialog ( debug output to terminal especially on weirdy windows is not the best way to see this problem ), I vaguely recall the twisted mechanism for getting some dialogs to pop with the interaction handler stuff, I will have a look so I'd like to add at least debug output - the debug output in the patch is done just using fprintf(stderr), as I cannot find anything that would simply log a debug message and not do more - is there anything like that in LO? - osl_loadModule*() functions are plain C functions from stable API, so no overloading to add the extra argument for the error message, the best I've come up with was simply a wrapper around dlerror(), which I admit is not very nice API either, would there be any better approach? don't see one, don't see that it is so bad either Noel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] How to set fillcolor and linecolor to a SdrObject?
On Wed, 2011-08-03 at 00:23 +0200, Regina Henschel wrote: > But I do not know how to make the object transparent. Especially, I do > not know how to use these "itempool"-thing. Does pObject->SetMergedItem(XFillTransparenceItem(some_value)); do anything useful for you. opengroking around for stuff like that might help anyway C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [PUSHED] fix for crash in new ww8 field code
On 03/08/11 10:44, Caolán McNamara wrote: i.e. accessing top of an empty stack, noticed when doing regression tests on #i38880#. Proposing cherry-pick to 3-4 http://cgit.freedesktop.org/libreoffice/writer/commit/?id=156c5bd4b04402f8dd724da6c8f984c4d63e4171 look good to me ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] [REVIEW] fix for crash in new ww8 field code
i.e. accessing top of an empty stack, noticed when doing regression tests on #i38880#. Proposing cherry-pick to 3-4 http://cgit.freedesktop.org/libreoffice/writer/commit/?id=156c5bd4b04402f8dd724da6c8f984c4d63e4171 C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Vcl module compilation failed
On Wed, 2011-08-03 at 08:06 +0200, Julien Nabet wrote: > Hello, > > Here is what I get if I try to compile vcl : > maryline@maryline:~/compile-libreoffice/libo/vcl$ rm -rf unxlngi6/ && > make -r > CxxObject/vcl/source/salmain/salmain.o is linked in by > Library/libvcllo.so StaticLibrary/libvclmain.a > /home/maryline/compile-libreoffice/libo/solenv/gbuild/extensions/final_ObjectOwner.mk:43: > > > [ WARN ] !!! > [ WARN ] !!! duplicate linked objects > [ WARN ] !!! That's a warning, not an error. If your compilation failed, it didn't fail because of "salmain in linked twice", that happens for everyone. C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Make dev-install without running smoketests
On Wed, 2011-08-03 at 10:54 +0400, Dmitry A. Ashkadov wrote: > I have the same problem. I can run LibreOffice, but make dev-install > fails on cppunit tests. > > > smoketest.cxx:203:Assertion > > Test name: N12_GLOBAL__N_14TestE::test > > equality assertion failed > > - Expected: > > - Actual : Impress:XML Draw:XML An install set that fails the smoketest is still installed in solver/350/PLATFORM/installation/, e.g solver/350/unxlngx6.pro/installation/ It doesn't fail the smoketest for me, so the right thing to do is to run the smoke test manually and see what fails. If it fails even the very minimal smoketest its really a big useless pile of bits tools->options->security->macro security->low : open libo/smoketestdoc/PLATFORM/bin/smoketestdoc.sxw and click the "run" button. C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [sumu...@novell.com: Re: [PATCH] for Bug 32719 (jeffrey chang)]
Hi, Regarding the fact that I may have trimmed too many outline levels as well as leaving some extraneous comments, I will resubmit the patch. Sorry for the trouble, Jeffrey On Wed, Aug 3, 2011 at 4:04 PM, jeffrey chang wrote: > Hi Muthu, > > 1. Sorry, next time I will do that. > > 2. Yes I did adjust the multipliers because after several searches through > codes, I could not locate where Impress extracts data specific to each > template (eg. where does template Water get dimensions xx by xx for body, or > Glacier xx by xx for title), and I didn't want to change the template layout > files (.lyt), so I decided that the best place would be multipliers. Of > course, I made sure that this was a reasonable choice, because I understood > that all templates would be affected. From the looks of it, after reducing > the height and width of the body, it seemed to eliminakernel imagete the > oversized textbox leaking problems. Also, none of the other text boxes in > any of the other slides look uncomfortably small, so I find it as a viable > solution. (For those that do, consider the Master Page View, in which the > Template has to make accommodations for the footer, date, slide # > underneath). Lastly, if by blank template you mean the default template, > there is no problem. There has been no problems I can see of after making > the changes I did, apart from the bug I reported. > > 3. I realized that there were two sources for the observable problems. The > first was that some of the outline/body dimensions for the templates were > plain wrong and just too big. That was addressed in 2. But another problem > is that the slide layout is based on the Master Page View Outline, and for > some of them, the outline points were causing a problem. For example, if you > like at the buggy Wine template, the outline/body leaks way to far below the > slide. The reason is that when you go to Master Page view, the Ninth level > of bullet point is jammed and squeezed, resulting in over-extending the > outline text box to make room for it, hence the huge increase in height. If > you eliminate the ninth level, the problem is reduced after leaving Master > Page View. Originally I was going to remove the eighth and ninth level of > the outline to solve this problem, but I decided to with 5-9 for test and > forgot to reset it to remove 8-9. And since I think that rarely will some > have to go to the eight level of bullet points in a presentation, once again > this is a viable solution. > > Sorry for the confusion and ambiguity. I hope this clears things up and the > patch is ok. Please reply if you see any more problems. > > Thanks, > Jeffrey > > > > On Wed, Aug 3, 2011 at 3:56 PM, Yifan Jiang wrote: > >> fyi >> -- >> Yifan Jiang >> Libreoffice / SuSE >> Contact: yifan - irc.freenode.net/libreoffice >> = >> http://www.libreoffice.org/ >> http://www.documentfoundation.org/ >> >> >> >> -- Forwarded message -- >> From: Muthu Subramanian K >> To: jeffrey chang >> Date: Wed, 03 Aug 2011 12:38:03 +0530 >> Subject: Re: [Libreoffice] [PATCH] for Bug 32719 (jeffrey chang) >> Hi, >> >> Thank you for the patch. >> Assuming this is your first patch (or at-least one of the first few), I >> have a few suggestions: >> 1. It would nice if you could remove the test code commented - helps in >> reviewing quickly. >> 2. You have changed the multipliers in multiple locations - are you sure >> these don't affect the normal working? (say, having a blank template?) >> 3. It would also be nice if you could explain the reason for removing >> the content from STR_PRESOBJ_MPOUTLLAYER5 to STR_PRESOBJ_MPOUTLLAYER9 >> >> Will wait for your updates before pushing... >> Thanks again for your nice work... >> >> Regards, >> Muthu Subramanian >> >> >> -- >> Date: Mon, 1 Aug 2011 13:27:33 +0800 >> From: jeffrey chang >> Subject: [Libreoffice] [PATCH] for Bug 32719 >> To: libreoffice@lists.freedesktop.org >> >> Here is the patch for EasyHack 32719. I have also discovered another bug >> relating to Impress Templates, documented here: >> >> https://bugs.freedesktop.org/show_bug.cgi?id=39705 >> >> I am submitting this under the LGPLv3+/MPL 1.1 license. >> >> Thanks, >> -- --Jeffrey Chang >> ___ >> LibreOffice mailing list >> LibreOffice@lists.freedesktop.org >> http://lists.freedesktop.org/mailman/listinfo/libreoffice >> >> > > > -- > --Jeffrey Chang > > ___ > LibreOffice mailing list > LibreOffice@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/libreoffice > > -- --Jeffrey Chang ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [MHST] GDB doesn't show name of methods in backtrace
On 08/02/2011 05:06 PM, Ta Duc Tung wrote: On 08/01/2011 06:55 PM, Michael Meeks wrote: Hi there, On Thu, 2011-07-28 at 01:09 +0700, Ta Duc Tung wrote: I've tried to make with debug=t but no change. GDB log still the same :( Any one please help! You need to 'make clean' in sw and then 'make debug=true' - and -then- you need to ensure that your install is linked to the dev tree with solenv/bin/linkoo. Then you need to re-run gdb, (with ooenv sourced), and you should have more luck. Thanks ! Michael. Hi, Firstly, I want to thank Michael. I still don't know how to ensure whether my install is linked to the dev tree but follow your word, GDB is somehow working (even there are still some "??" things, but at least, I can "next" and "step" now). I've try to use DDD to debug the code and I get this Segmentation Fault http://pastebin.com/6ciQanTq The bug above happens when I try to add a radio button to print dialog of Writer. Like this: http://pastebin.com/mS2G3e8t Please, how to fix this fault. I think I miss something in adding new radiobutton, but I don't know where it is. Thanks and Best Regards, Ta Tung Hi, I attach the patch for sw. It's used to add a radio button to print dialog in Writer. Please apply and see the bug. In this patch, I add a new radio button named Current page to print dialog. It does appears but when I click to choose that radio button, Writer gets terminated. Thanks and Best Regards, Ta Tung PS: Thank you, Caolán. diff --git a/sw/source/core/view/printdata.cxx b/sw/source/core/view/printdata.cxx index db5313d..b59372d 100644 --- a/sw/source/core/view/printdata.cxx +++ b/sw/source/core/view/printdata.cxx @@ -195,8 +195,8 @@ SwPrintUIOptions::SwPrintUIOptions( { ResStringArray aLocalizedStrings( SW_RES( STR_PRINTOPTUI ) ); -OSL_ENSURE( aLocalizedStrings.Count() >= 30, "resource incomplete" ); -if( aLocalizedStrings.Count() < 30 ) // bad resource ? +OSL_ENSURE( aLocalizedStrings.Count() >= 31, "resource incomplete" ); +if( aLocalizedStrings.Count() < 31 ) // bad resource ? return; // printing HTML sources does not have any valid UI options. @@ -310,9 +310,9 @@ SwPrintUIOptions::SwPrintUIOptions( // create a choice for the content to create rtl::OUString aPrintRangeName( RTL_CONSTASCII_USTRINGPARAM( "PrintContent" ) ); -uno::Sequence< rtl::OUString > aChoices( 3 ); -uno::Sequence< sal_Bool > aChoicesDisabled( 3 ); -uno::Sequence< rtl::OUString > aHelpIds( 3 ); +uno::Sequence< rtl::OUString > aChoices( 4 ); +uno::Sequence< sal_Bool > aChoicesDisabled( 4 ); +uno::Sequence< rtl::OUString > aHelpIds( 4 ); aChoices[0] = aLocalizedStrings.GetString( 27 ); aChoicesDisabled[0] = sal_False; aHelpIds[0] = rtl::OUString( RTL_CONSTASCII_USTRINGPARAM( ".HelpID:vcl:PrintDialog:PrintContent:RadioButton:0" ) ); @@ -321,7 +321,11 @@ SwPrintUIOptions::SwPrintUIOptions( aHelpIds[1] = rtl::OUString( RTL_CONSTASCII_USTRINGPARAM( ".HelpID:vcl:PrintDialog:PrintContent:RadioButton:1" ) ); aChoices[2] = aLocalizedStrings.GetString( 29 ); aChoicesDisabled[2] = sal_Bool(! bHasSelection); -aHelpIds[2] = rtl::OUString( RTL_CONSTASCII_USTRINGPARAM( ".HelpID:vcl:PrintDialog:PrintContent:RadioButton:2" ) ); +aHelpIds[2] = rtl::OUString( RTL_CONSTASCII_USTRINGPARAM( ".HelpID:vcl:PrintDialog:PrintContent:RadioButton:2" ) ); +aChoices[3] = aLocalizedStrings.GetString( 30 ); +aChoicesDisabled[3] = sal_False; +aHelpIds[3] = rtl::OUString( RTL_CONSTASCII_USTRINGPARAM( ".HelpID:vcl:PrintDialog:PrintContent:RadioButton:3" ) ); + m_aUIProperties[nIdx++].Value = getChoiceControlOpt( rtl::OUString(), aHelpIds, aPrintRangeName, diff --git a/sw/source/ui/config/optdlg.src b/sw/source/ui/config/optdlg.src index 8a21f45..3264001 100644 --- a/sw/source/ui/config/optdlg.src +++ b/sw/source/ui/config/optdlg.src @@ -1069,6 +1069,7 @@ StringArray STR_PRINTOPTUI < "~All pages"; >; < "Pa~ges"; >; < "~Selection"; >; + < "Cu~rrent page"; >; }; }; ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [sumu...@novell.com: Re: [PATCH] for Bug 32719 (jeffrey chang)]
Hi Muthu, 1. Sorry, next time I will do that. 2. Yes I did adjust the multipliers because after several searches through codes, I could not locate where Impress extracts data specific to each template (eg. where does template Water get dimensions xx by xx for body, or Glacier xx by xx for title), and I didn't want to change the template layout files (.lyt), so I decided that the best place would be multipliers. Of course, I made sure that this was a reasonable choice, because I understood that all templates would be affected. From the looks of it, after reducing the height and width of the body, it seemed to eliminakernel imagete the oversized textbox leaking problems. Also, none of the other text boxes in any of the other slides look uncomfortably small, so I find it as a viable solution. (For those that do, consider the Master Page View, in which the Template has to make accommodations for the footer, date, slide # underneath). Lastly, if by blank template you mean the default template, there is no problem. There has been no problems I can see of after making the changes I did, apart from the bug I reported. 3. I realized that there were two sources for the observable problems. The first was that some of the outline/body dimensions for the templates were plain wrong and just too big. That was addressed in 2. But another problem is that the slide layout is based on the Master Page View Outline, and for some of them, the outline points were causing a problem. For example, if you like at the buggy Wine template, the outline/body leaks way to far below the slide. The reason is that when you go to Master Page view, the Ninth level of bullet point is jammed and squeezed, resulting in over-extending the outline text box to make room for it, hence the huge increase in height. If you eliminate the ninth level, the problem is reduced after leaving Master Page View. Originally I was going to remove the eighth and ninth level of the outline to solve this problem, but I decided to with 5-9 for test and forgot to reset it to remove 8-9. And since I think that rarely will some have to go to the eight level of bullet points in a presentation, once again this is a viable solution. Sorry for the confusion and ambiguity. I hope this clears things up and the patch is ok. Please reply if you see any more problems. Thanks, Jeffrey On Wed, Aug 3, 2011 at 3:56 PM, Yifan Jiang wrote: > fyi > -- > Yifan Jiang > Libreoffice / SuSE > Contact: yifan - irc.freenode.net/libreoffice > = > http://www.libreoffice.org/ > http://www.documentfoundation.org/ > > > > -- Forwarded message -- > From: Muthu Subramanian K > To: jeffrey chang > Date: Wed, 03 Aug 2011 12:38:03 +0530 > Subject: Re: [Libreoffice] [PATCH] for Bug 32719 (jeffrey chang) > Hi, > > Thank you for the patch. > Assuming this is your first patch (or at-least one of the first few), I > have a few suggestions: > 1. It would nice if you could remove the test code commented - helps in > reviewing quickly. > 2. You have changed the multipliers in multiple locations - are you sure > these don't affect the normal working? (say, having a blank template?) > 3. It would also be nice if you could explain the reason for removing > the content from STR_PRESOBJ_MPOUTLLAYER5 to STR_PRESOBJ_MPOUTLLAYER9 > > Will wait for your updates before pushing... > Thanks again for your nice work... > > Regards, > Muthu Subramanian > > > -- > Date: Mon, 1 Aug 2011 13:27:33 +0800 > From: jeffrey chang > Subject: [Libreoffice] [PATCH] for Bug 32719 > To: libreoffice@lists.freedesktop.org > > Here is the patch for EasyHack 32719. I have also discovered another bug > relating to Impress Templates, documented here: > > https://bugs.freedesktop.org/show_bug.cgi?id=39705 > > I am submitting this under the LGPLv3+/MPL 1.1 license. > > Thanks, > -- --Jeffrey Chang > ___ > LibreOffice mailing list > LibreOffice@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/libreoffice > > -- --Jeffrey Chang ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Make dev-install without running smoketests
I have the same problem. I can run LibreOffice, but make dev-install fails on cppunit tests. smoketest.cxx:203:Assertion Test name: N12_GLOBAL__N_14TestE::test equality assertion failed - Expected: - Actual : Impress:XML Draw:XML 03.08.2011 07:37, swagat sharma пишет: Hi, My build is successful but when I issue make dev-install, it crashes in the smoketest module. Is there any way by which I can skip smoketest? Regards -- Swagat -- Best regards, Dmitry <>___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice