[Libreoffice] Autogen switch --with-agfa-monotype-fonts fails to download fonts tarball, Win32 build-only switch ?
Hi *, Whilst trying to build from master on Mac OSX with the autogen switch --with-agfa-monotype-fonts, I get the following error : log for /Users/alex/DevHack/git/libo/more_fonts/fonts/ttf_amt dmake: Error: -- `../../unxmacxi.pro/misc/9bad41ce5c3fd069b87f28aeac70314d-agfa-monotype-fonts.unpack' not found, and can't be made If I've understood this commit : http://lists.freedesktop.org/archives/libreoffice-commits/2011-April/010653.html then the switch should only be available to Win32 builds ? Alex ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] Building from master on MacOSX with autogen --enable-ext-wiki-publisher fails in apache-commons
Hi *, Just a heads-up, building from master on Mac OSX with --enable-ext-wiki-publisher fails in apache-commons. The log file shows : compile: [javac] /Users/alex/DevHack/git/libo/clone/extensions/apache-commons/unxmacxi.pro/misc/build/commons-httpclient-3.1/build.xml:184: warning: 'includeantruntime' was not set, defaulting to build.sysclasspath=last; set to false for repeatable builds [javac] Compiling 139 source files to /Users/alex/DevHack/git/libo/clone/extensions/apache-commons/unxmacxi.pro/misc/build/commons-httpclient-3.1/target/classes [javac] /Users/alex/DevHack/git/libo/clone/extensions/apache-commons/unxmacxi.pro/misc/build/commons-httpclient-3.1/src/java/org/apache/commons/httpclient/ChunkedInputStream.java:39: package org.apache.commons.logging does not exist Alex ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] [GSOC][Report] MultiLine InputBar
Hello Kohei, Noel This is the brief of my work for this week. As I was supposed to create a new container class which will contain ScTextWnd obejct + ScrollBar + Button object, I've done so. Also I've tunneled the calls to ScTextWnd from ScInputWindow via the new methods created into the ScInputBarGroup(new class which I've created). But the problem I'm facing is now when I removed the aTextWnd.Show() from constructor of ScInputWindow and added it into the ScInputBarGroup's constructor the inputbar will not appear. I've passed the ScInputBarGroup's object in InsertWindow() function though. But Still not able to figure out what might be the problem. I'm working on it. Hope I'll figure it out soon and I'm not able to come online because of the internet connection problem; but it will be resolved soon. Thanks and regards -- Anurag Jain Final yr B.Tech CSE SASTRA University Thanjavur(T.N.)-613402 ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] QA-hints ? - question for devs !
Hi John, LeMoyne Castle wrote (15-06-11 13:01) On Wed, Jun 15, 2011 at 1:02 AM, Cor Nouws oo...@nouenoff.nl mailto:oo...@nouenoff.nl wrote: ... What I am more looking for, is sometimes the hints related to (some of) the vast amount of code clean-ups/improvements, that possibly have influence on area A or B. (In the commit-logs for master) Maybe maybe it is possible for the involved devs, when they think it is possibly relevant, to add a few words to the description so that it is indeed able to read that in the summaries. ... Sounds reasonable? It is quite reasonable to try to get the dev work result to mesh with qa testing. Sadly, git requests that the short commit message be 41 characters. Personally, I struggle with that at each commit ;-)... Git does allow more lines in the commit header but those usually describe changes to the code, are often technical and some of the comments about the removed/replaced code do *not* belong on the wiki. Thanks for this explanation. In the wiki of the future, the summary pages' bug numbers will link to the actual bug reports (at least for fdo) so that the summary indirectly provides the intended functional changes in LibreOffice. The bug Also with the present list it is not too difficult to go to the related issue. reports often contain steps to reproduce the bug that make it clear where to test the fix. Perhaps the bug-fix section of the summary could eventually include the bug summary (short desc) next to the link as well. What I am not looking for, is exact steps to reproduce / verify. When I know that some work has been done with function/area X, it gives the opportunity to work a bit around that area, according to my own habits and knowledge of the suite. The addition of a bug number to the short commit message is the most efficient way for the developers to pass testing info to qa. Thanks for raising and clarifying the issue of qa's need for clues about how to find and test the latest changes in any of master, 3.4 or the release branches, And thanks to you and others for pointing to the current summaries as a good starting point for that, anyway as far as I am concerned. I understand the limitations now too. Since there is much on the route in our QA-process/work, we can see if there shows up a natural, not too complicated, possibility to enhance it. Kind regards, Cor -- - Cor - http://nl.libreoffice.org ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] Debug compilation fails in sal module
Hello,
I recreated from scratch my local repository and here is what I get :
--
- start unit test #2 on library ../../../unxlngi6/lib/libosl_process.so
--
:
LD_LIBRARY_PATH=/home/maryline/compile-libreoffice/libo/clone/ure/sal/unxlngi6/lib:/home/maryline/compile-libreoffice/libo/solver/350/unxlngi6/lib${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
../../../unxlngi6/bin/cppunittester ../../../unxlngi6/lib/libosl_process.so
Error: File
/home/maryline/compile-libreoffice/libo/clone/ure/sal/cpprt/operators_new_delete.cxx,
Line 96: operator delete mismatch
Error: File
/home/maryline/compile-libreoffice/libo/clone/ure/sal/cpprt/operators_new_delete.cxx,
Line 96: operator delete mismatch
OK (2)
/bin/bash: line 1: 9574 Segmentation fault
LD_LIBRARY_PATH=/home/maryline/compile-libreoffice/libo/clone/ure/sal/unxlngi6/lib:/home/maryline/compile-libreoffice/libo/solver/350/unxlngi6/lib${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
../../../unxlngi6/bin/cppunittester ../../../unxlngi6/lib/libosl_process.so
dmake: Error code 139, while making 'test2'
Here is the autogen.lastrun :
--enable-symbols
--enable-binfilters
--enable-ext-barcode
--enable-ext-diagram
--enable-ext-google-docs
--enable-ext-hunart
--enable-ext-lightproof
--enable-ext-nlpsolver
--enable-ext-ct2n
--enable-ext-numbertext
--enable-ext-oooblogger
--enable-ext-pdfimport
--enable-ext-presenter-console
--enable-ext-presenter-minimizer
--enable-ext-report-builder
--enable-ext-scripting-beanshell
--enable-ext-scripting-javascript
--enable-ext-scripting-python
--enable-ext-typo
--enable-ext-validator
--enable-ext-watch-window
--enable-ext-wiki-publisher
--enable-dbus
--enable-graphite
--enable-evolution2
--enable-werror
--enable-debug
--enable-dbgutil
--enable-crashdump
--enable-kde4
--enable-dependency-tracking
I took a look at the git history of sal module and found nothing special.
(gcc (Debian 4.6.0-10) 4.6.1 20110526 (prerelease), on Debian testing x86)
Any ideas ?
Julien.
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [GSoC] [rtfimport] 4th week
Hi, Fourth week starts here: http://cgit.freedesktop.org/~vmiklos/lo-gsoc/tree/README#n517 In short, it was about: - finishing pictures - starting tables: borders, merges, cell properties (background, alignment, etc.) I plan to finish tables (nested ones, other missing table bits) next week. Miklos pgpG839wL7Iuj.pgp Description: PGP signature ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] Building from master on MacOSX with autogen --enable-ext-wiki-publisher fails in apache-commons
Hi Alexander, *, On Sun, Jun 19, 2011 at 10:49 AM, Alexander Thurgood alex.thurg...@gmail.com wrote: Just a heads-up, building from master on Mac OSX with --enable-ext-wiki-publisher fails in apache-commons. No it does not, both my tinderbox as well as the MacIntel tinderbox use the --with-distro=LibreOfficeMacOSX switch, and that enables the extension. So when it breaks, then it breaks once again because of poisoned environment. (or it was a very recent change and the bots did not compile it yet) ciao Christian ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] Did you hear the news? LibreOffice is going to have a conference
Greetings my hacker friends. Well I know you all heard about this thing called the LibreOffice Conference taking place in Paris France this year. Conferences tend to work better if people present at the them. The first step in presenting at a conference is usually to submit a paper proposal. *smile*... OK - all kidding aside - the call for papers is open, you know that so I'm just here to...hmm...nudge you all a little bit.. I know it is a PITA to stop and put together a proposal but a necessary one IMO. Thanks Drew Jensen ps - the number of nuisance reminders will be in an inverse relationship to the number of proposals received..hint, hint. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] autogen.sh fails to build Libre Office
This is a very noobish ask for help but when I tried to build LibreOffice, the script autogen fails to detect 'fontconfig' here's the output checking for FONTCONFIG... no configure: error: Package requirements (fontconfig = 2.2.0) were not met: No package 'fontconfig' found Consider adjusting the PKG_CONFIG_PATH environment variable if you installed software in a non-standard prefix. Alternatively, you may set the environment variables FONTCONFIG_CFLAGS and FONTCONFIG_LIBS to avoid the need to call pkg-config. See the pkg-config man page for more details. I have the fontconfig 2.8.0, on Ubuntu 11.04 32 bits, I've followed the steps on http://www.documentfoundation.org/develop/ and I searched google but all I could find was bug reports having the same problem but with other builds. Is there any workaround for this? ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] autogen.sh fails to build Libre Office
On Sun, Jun 19, 2011 at 3:04 PM, Ibrahim Beicker ibrabeic...@gmail.com wrote: This is a very noobish ask for help but when I tried to build LibreOffice, the script autogen fails to detect 'fontconfig' here's the output checking for FONTCONFIG... no configure: error: Package requirements (fontconfig = 2.2.0) were not met: No package 'fontconfig' found Consider adjusting the PKG_CONFIG_PATH environment variable if you installed software in a non-standard prefix. Alternatively, you may set the environment variables FONTCONFIG_CFLAGS and FONTCONFIG_LIBS to avoid the need to call pkg-config. See the pkg-config man page for more details. I have the fontconfig 2.8.0, on Ubuntu 11.04 32 bits, I've followed the steps on http://www.documentfoundation.org/develop/ and I searched google but all I could find was bug reports having the same problem but with other builds. Is there any workaround for this? See http://lists.freedesktop.org/archives/libreoffice/2011-April/010054.html Since you use a recent mainstream distribution, you can take advantage of the packaging facilities to select and install all necessary development dependencies. This single command, sudo apt-get build-dep libreoffice will install the necessary dependencies (including the fontconfig development package) in one go. Indeed, the wiki page at http://wiki.documentfoundation.org/Development/Native_Build should get a reference to the additional information that can be found at http://www.libreoffice.org/get-involved/developers/ Simos ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] [SOLVED] Building from master on MacOSX with autogen --enable-ext-wiki-publisher fails in apache-commons
Hi Christian, Le 19/06/11 13:39, Christian Lohmaier a écrit : Hi Alexander, *, On Sun, Jun 19, 2011 at 10:49 AM, Alexander Thurgood alex.thurg...@gmail.com wrote: Just a heads-up, building from master on Mac OSX with --enable-ext-wiki-publisher fails in apache-commons. No it does not, both my tinderbox as well as the MacIntel tinderbox use the --with-distro=LibreOfficeMacOSX switch, and that enables the extension. So when it breaks, then it breaks once again because of poisoned environment. (or it was a very recent change and the bots did not compile it yet) Further investigation revealed that I had a folder logging and associated make file missing from my apache-commons in my repo, which I have managed to restore from backup (although I would have thought a git pull would have sorted that out, it didn't). It had nothing to do with the installation of my ports environment. Alex ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] [PATCH] fix quickstart on platforms using GTK and use the correct plugin name
After switching to gbuild ENABLE_SYSTRAY_GTK was not defined in the
sfx target and the plugin name for libqstart was also hardcoded for
linux.
---
sfx2/Library_qstart.mk|7 +++
sfx2/Library_sfx.mk |9 +
sfx2/source/appl/shutdownicon.cxx |2 +-
3 files changed, 17 insertions(+), 1 deletions(-)
diff --git a/sfx2/Library_qstart.mk b/sfx2/Library_qstart.mk
index 0709f60..e46a581 100755
--- a/sfx2/Library_qstart.mk
+++ b/sfx2/Library_qstart.mk
@@ -44,6 +44,13 @@ $(eval $(call gb_Library_set_defs,qstart_gtk,\
-DENABLE_QUICKSTART_APPLET \
))
+ifeq ($(ENABLE_SYSTRAY_GTK),TRUE)
+$(eval $(call gb_Library_set_defs,qstart_gtk,\
+$$(DEFS) \
+-DENABLE_SYSTRAY_GTK \
+))
+endif
+
$(eval $(call gb_Library_set_cflags,qstart_gtk,\
$$(CFLAGS) \
$(filter-out -I%,$(GTK_CFLAGS)) \
diff --git a/sfx2/Library_sfx.mk b/sfx2/Library_sfx.mk
index c93efac..52bce60 100755
--- a/sfx2/Library_sfx.mk
+++ b/sfx2/Library_sfx.mk
@@ -56,6 +56,15 @@ $(eval $(call gb_Library_set_defs,sfx,\
-DSFX2_DLLIMPLEMENTATION \
))
+ifeq ($(ENABLE_SYSTRAY_GTK),TRUE)
+$(eval $(call gb_Library_set_defs,sfx,\
+$$(DEFS) \
+-DENABLE_QUICKSTART_APPLET \
+-DENABLE_SYSTRAY_GTK \
+-DPLUGIN_NAME=libqstart_gtk$(gb_Library_OOOEXT) \
+))
+endif
+
$(eval $(call gb_Library_add_linked_libs,sfx,\
comphelper \
cppu \
diff --git a/sfx2/source/appl/shutdownicon.cxx
b/sfx2/source/appl/shutdownicon.cxx
index 0a6569b..dbce847 100644
--- a/sfx2/source/appl/shutdownicon.cxx
+++ b/sfx2/source/appl/shutdownicon.cxx
@@ -90,7 +90,7 @@ extern C { static void SAL_CALL thisModule() {} }
# endif
#endif
-#if defined(UNX) defined(ENABLE_SYSTRAY_GTK)
+#if defined(UNX) defined(ENABLE_SYSTRAY_GTK) !defined(PLUGIN_NAME)
#define PLUGIN_NAME libqstart_gtkli.so
#endif
--
1.7.5
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] [PATCH] fdo#37341 EasyHack: Goal Seek hangs indefinitely for too many calculation steps (Formula Cell $F$110)
Hi, By poking the sample case of fdo#37341 with gdb, I found loosing the maximum limit of recursion of recalculation is enough to solve it, like https://bugs.freedesktop.org/attachment.cgi?id=48162 But its modified value of limit is still arbitrarily selected. Can we estimate the growth of memory comsumption according to the value? Any suggestion would be appreciated. Cheers, -- Takeshi Abe ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] [PATCH] Improvements to import/export of character attributes in HTML filter
Attached patches implement import of CSS text-transform property and extend
the import and export to work with all HTML browser compatibility options.
Contributed under LGPLv3+ and MPL.
Harri
From 8eea1bf955f6f2cb7c0e0622b0b2fa6975cee3db Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Harri=20Pitk=C3=A4nen?= hatap...@iki.fi
Date: Sat, 18 Jun 2011 15:57:14 +0300
Subject: [PATCH 1/2] Import and export character case transformations
unconditionally to/from HTML
Import small caps and export small caps, lowecase, uppercase and titlecase
transformations unconditionally to/from HTML. Previously import and export
was limited by browser compatibility settings but all mainstream browsers
released within last 5 years or so should support the relevant CSS attributes.
---
sw/source/filter/html/css1atr.cxx |3 ---
sw/source/filter/html/htmlatr.cxx |4 ++--
2 files changed, 2 insertions(+), 5 deletions(-)
diff --git a/sw/source/filter/html/css1atr.cxx b/sw/source/filter/html/css1atr.cxx
index 92a149b..d980b55 100644
--- a/sw/source/filter/html/css1atr.cxx
+++ b/sw/source/filter/html/css1atr.cxx
@@ -2561,9 +2561,6 @@ static Writer OutCSS1_SvxCaseMap( Writer rWrt, const SfxPoolItem rHt )
{
SwHTMLWriter rHTMLWrt = (SwHTMLWriter)rWrt;
-if( !rHTMLWrt.IsHTMLMode(HTMLMODE_SMALL_CAPS) )
-return rWrt;
-
switch( ((const SvxCaseMapItem)rHt).GetCaseMap() )
{
case SVX_CASEMAP_NOT_MAPPED:
diff --git a/sw/source/filter/html/htmlatr.cxx b/sw/source/filter/html/htmlatr.cxx
index 2d10e6f..f7a0ec1 100644
--- a/sw/source/filter/html/htmlatr.cxx
+++ b/sw/source/filter/html/htmlatr.cxx
@@ -1505,8 +1505,8 @@ HTMLOnOffState HTMLEndPosLst::GetHTMLItemState( const SfxPoolItem rItem )
break;
case RES_CHRATR_CASEMAP:
-if( IsHTMLMode(HTMLMODE_SMALL_CAPS) )
-eState = HTML_STYLE_VALUE;
+eState = HTML_STYLE_VALUE;
+break;
case RES_CHRATR_KERNING:
if( IsHTMLMode(HTMLMODE_FULL_STYLES) )
--
1.7.5.4
From e0d8e9114c6778bb534a61ee0f50dcf11a393ad0 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Harri=20Pitk=C3=A4nen?= hatap...@iki.fi
Date: Sat, 18 Jun 2011 21:12:01 +0300
Subject: [PATCH 2/2] Support CSS text-transform property in HTML import
---
sw/source/filter/html/svxcss1.cxx | 35 +++
1 files changed, 35 insertions(+), 0 deletions(-)
diff --git a/sw/source/filter/html/svxcss1.cxx b/sw/source/filter/html/svxcss1.cxx
index 117ef4e..2fcf673 100644
--- a/sw/source/filter/html/svxcss1.cxx
+++ b/sw/source/filter/html/svxcss1.cxx
@@ -128,6 +128,14 @@ static CSS1PropertyEnum const aFontVariantTable[] =
{ 0, 0 }
};
+static CSS1PropertyEnum const aTextTransformTable[] =
+{
+{ sCSS1_PV_uppercase, SVX_CASEMAP_VERSALIEN },
+{ sCSS1_PV_lowercase, SVX_CASEMAP_GEMEINE },
+{ sCSS1_PV_capitalize, SVX_CASEMAP_TITEL },
+{ 0, 0 }
+};
+
static CSS1PropertyEnum const aDirectionTable[] =
{
{ sCSS1_PV_ltr, FRMDIR_HORI_LEFT_TOP },
@@ -1314,6 +1322,32 @@ static void ParseCSS1_font_variant( const CSS1Expression *pExpr,
}
}
+static void ParseCSS1_text_transform( const CSS1Expression *pExpr,
+SfxItemSet rItemSet,
+SvxCSS1PropertyInfo /*rPropInfo*/,
+const SvxCSS1Parser /*rParser*/ )
+{
+OSL_ENSURE( pExpr, no expression );
+
+// none | capitalize | uppercase | lowercase
+
+switch( pExpr-GetType() )
+{
+case CSS1_IDENT:
+{
+sal_uInt16 nCaseMap;
+if( SvxCSS1Parser::GetEnum( aTextTransformTable, pExpr-GetString(),
+nCaseMap ) )
+{
+rItemSet.Put( SvxCaseMapItem( (SvxCaseMap)nCaseMap,
+aItemIds.nCaseMap ) );
+}
+}
+default:
+;
+}
+}
+
/* */
static void ParseCSS1_color( const CSS1Expression *pExpr,
@@ -3140,6 +3174,7 @@ static CSS1PropEntry aCSS1PropFnTab[] =
CSS1_PROP_ENTRY(text_align),
CSS1_PROP_ENTRY(text_decoration),
CSS1_PROP_ENTRY(text_indent),
+CSS1_PROP_ENTRY(text_transform),
CSS1_PROP_ENTRY(margin_left),
CSS1_PROP_ENTRY(margin_right),
CSS1_PROP_ENTRY(margin_top),
--
1.7.5.4
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] [PATCH] fix fdo#37110 --quickstart shouldn't show start center
---
desktop/source/app/app.cxx |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/desktop/source/app/app.cxx b/desktop/source/app/app.cxx
index d8483a2..8c86886 100644
--- a/desktop/source/app/app.cxx
+++ b/desktop/source/app/app.cxx
@@ -1784,7 +1784,7 @@ int Desktop::Main()
if ( !pExecGlobals-bRestartRequested )
{
-if ((!rCmdLineArgs.WantsToLoadDocument()
!rCmdLineArgs.IsInvisible() !rCmdLineArgs.IsHeadless() )
+if ((!rCmdLineArgs.WantsToLoadDocument()
!rCmdLineArgs.IsInvisible() !rCmdLineArgs.IsHeadless()
!rCmdLineArgs.IsQuickstart())
(SvtModuleOptions().IsModuleInstalled(SvtModuleOptions::E_SSTARTMODULE))
(!bExistsRecoveryData
)
(!bExistsSessionData
)
--
1.7.5
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] Interesting difference in thread-specific data function semantics for Windows vs. Unix
I was investigating why the osl_Thread unit test
(sal/qa/osl/process/osl_Thread.cxx) fails most of the times on Windows. If I
understand correctly, the problem is a quite fundamental difference in the
semantics of thread-specific data keys created with a destructor callback
passed to osl_createrThreadKey() (or its wrapper, the constructor of the
ThreadData class):
On Windows, if the value of a key is changed with osl_setThreadKeyData() (or
its wrapper, ThreadData::setData()), the destructor callbac is called for the
old value. (See the code in sal/osl/w32/thread.c. On Unix this does not seem to
happen. There osl_setThreadKeyData() is just a thin wrapper for
pthread_setspecific(), and the documentation for that does not say anything
about the destructor being called. (Experimentation confirms this, at least on
Linux.)
Now, this means that on Linux, if you set a thread-specific key's value to some
pointer to dynamically allocated memory, and then set it to another value, the
first value will indeed leak unless you otherwise free it yourself. This is why
Caolán added some explicit delete calls in April. But at least some of those
delete calls now then cause double deletes of the same data on Windows, causing
to slight (non-fatal in this case) heap corruption, and in most runs of the
test then (apparently a bit timing dependent) a failed assertion.
So, the interesting question now then is whether this difference in
osl_setThreadKeyData() semantics is intentional and known, and handled
specifically in those places in the code where thread-specific data is used? Or
whether the code assumes the semantics on Windows and just then leaks on Unix?
Or what...
Here is a patch that adds some debugging output to osl_Thread.cxx:
diff --git a/sal/qa/osl/process/osl_Thread.cxx
b/sal/qa/osl/process/osl_Thread.cxx
index aa6ad67..982a82e
--- a/sal/qa/osl/process/osl_Thread.cxx
+++ b/sal/qa/osl/process/osl_Thread.cxx
@@ -1896,7 +1896,8 @@ namespace osl_Thread
// destroy function when the binding thread terminate
void SAL_CALL destroyCallback(void * data)
{
-delete[] (char *) data;
+fprintf (stderr, del: %p (data in destroyCallback)\n, data); fflush
(stderr);
+//delete[] (char *) data;
}
static ThreadData myThreadData(destroyCallback);
@@ -1919,6 +1920,7 @@ private:
void SAL_CALL run()
{
char * pc = new char[2];
+fprintf (stderr, new: %p (pc in myKeyThread::run)\n, pc); fflush
(stderr);
// strcpy(pc, m_nData);
memcpy(pc, m_nData, 1);
pc[1] = '\0';
@@ -2032,6 +2034,7 @@ namespace osl_ThreadData
{
// at first, set the data a value
char* pc = new char[2];
+fprintf (stderr, new: %p (pc in setData_002)\n, pc); fflush
(stderr);
char m_nData = 'm';
// LLA: this is a copy functions only and really only for \0 terminated strings
// m_nData is not a string, it's a character
@@ -2068,6 +2071,7 @@ namespace osl_ThreadData
{
// at first, set the data a value
char* pc = new char[2];
+fprintf (stderr, new: %p (pc in setData_003)\n, pc); fflush
(stderr);
char m_nData = 'm';
memcpy(pc, m_nData, 1);
pc[1] = '\0';
@@ -2080,6 +2084,7 @@ namespace osl_ThreadData
// aThread1 and aThread2 should have not terminated yet
// setData the second time
char* pc2 = new char[2];
+fprintf (stderr, new: %p (pc2 in setData_003)\n, pc2);
fflush (stderr);
m_nData = 'o';
memcpy(pc2, m_nData, 1);
pc2[1] = '\0';
@@ -2100,8 +2105,10 @@ namespace osl_ThreadData
cData1 == 'a' cData2 == 'b' aChar == 'o'
);
-delete [] pc2;
-delete [] pc;
+fprintf (stderr, del: %p (pc2 in setData_003\n, pc2); fflush
(stderr);
+// delete [] pc2;
+fprintf (stderr, del: %p (pc in setData_003)\n, pc); fflush
(stderr);
+// delete [] pc;
}
CPPUNIT_TEST_SUITE(setData);
@@ -2127,6 +2134,7 @@ namespace osl_ThreadData
void getData_001()
{
char* pc = new char[2];
+fprintf (stderr, new: %p (pc in getData_001)\n, pc); fflush
(stderr);
char m_nData[] = i;
strcpy(pc, m_nData);
myThreadData.setData(pc);
@@ -2150,7 +2158,8 @@ namespace osl_ThreadData
cData1 == 'c' cData2 == 'd' aChar == 'i'
);
-delete [] pc;
+fprintf (stderr, del: %p (pc in getData_001)\n, pc); fflush
(stderr);
+// delete [] pc;
}
// setData then change the value in the address data pointer points,
@@ -2158,6 +2167,7 @@
[Libreoffice] [GSoC][performance] report - week #4
Hi all, I've started adding prefixes for components, so we can then merge libraries. Just in already gnumake-ified modules. First was toolkit. There is just one component and it seems to be working. I have tested that component. Then some components in configmgr, vcl, comphelper, sfx2, framework, svtools, svl. It's not straightforward always. For example in sfx2 there were also component_ methods in workben directory. I've ignored them. What's in workben directories? Now I'm playing with makefiles and next week I want to merge some libraries. All the best, Matus ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] QA-hints ? - question for devs !
Hi *, Cor Nouws wrote (19-06-11 11:51) And thanks to you and others for pointing to the current summaries as a good starting point for that, anyway as far as I am concerned. I understand the limitations now too. Since there is much on the route in our QA-process/work, we can see if there shows up a natural, not too complicated, possibility to enhance it. Is there a script available, or a combination of git-commands, or ... that is used to extract certain information for the summaries? If so, I could use/adapt that to get information on certain weeks, branches, ... ? Cheers, Cor -- - Cor - http://nl.libreoffice.org ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] Making our Own Banned.h
Hello list. As you all know, there are a bunch of old C APIs that make security vulnerabilities trivial to implement. And doing a git grep tells me that we use those a plenty. Now, not all of it may create vulnerabilities, but it is good practice to migrate away from those as much as possible. Microsoft has compiled a useful list: http://msdn.microsoft.com/en-us/library/bb288454.aspx And they have made a header (I'm attaching here) that works on their compiler. Now, I think we should make it multi-platform, so that the whole code base can benefit from it. The transition must be gradual, for sure, but I think we'd benefit a lot from it in the long run. What are the compilers that we must handle? - Gcc TODO - Microsoft's DONE - Sun's cc family ??? - Intel's ??? Regards, -- Marc-André Laverdière Software Security Scientist Innovation Labs, Tata Consultancy Services Hyderabad, India /*** * banned.h - list of Microsoft Security Development Lifecycle (SDL) banned APIs * * Purpose: * This include file contains a list of banned APIs which should not be used in new code and * removed from legacy code over time. * * History * 01-Jan-2006 - mikehow - Initial Version * 22-Apr-2008 - mikehow - Updated to SDL 4.1, commented out recommendations and added memcpy * 26-Jan-2009 - mikehow - Updated to SDL 5.0, made the list sane, added SDL compliance levels * 10-Feb-2009 - mikehow - Updated based on feedback from MS Office * 12-May-2009 - jpardue - Added wmemcpy * 08-Jul-2009 - mikehow - Fixed header #ifndef/#endif logic, made the SDL recommended compliance level name more obvious * 05-Nov-2009 - mikehow - Added vsnprintf (ANSI version of _vsnprintf) * 01-Jan-2010 - mikehow - Added better strsafe integration, now the following works: * #include strsafe.h * #include banned.h * 04-Jun-2010 - mikehow - Small #if bug fix * * ***/ #ifndef _INC_BANNED # define _INC_BANNED # if defined(_MSC_VER) # pragma once // SDL 5.0 and later Requirements # if defined(_STRSAFE_H_INCLUDED_) !defined(STRSAFE_NO_DEPRECATE) // Only deprecate what's not already deprecated by StrSafe # pragma deprecated (_mbscpy, _mbccpy) # pragma deprecated (strcatA, strcatW, _mbscat, StrCatBuff, StrCatBuffA, StrCatBuffW, StrCatChainW, _tccat, _mbccat) # pragma deprecated (strncpy, wcsncpy, _tcsncpy, _mbsncpy, _mbsnbcpy, StrCpyN, StrCpyNA, StrCpyNW, StrNCpy, strcpynA, StrNCpyA, StrNCpyW, lstrcpyn, lstrcpynA, lstrcpynW) # pragma deprecated (strncat, wcsncat, _tcsncat, _mbsncat, _mbsnbcat, lstrncat, lstrcatnA, lstrcatnW, lstrcatn) # pragma deprecated (IsBadWritePtr, IsBadHugeWritePtr, IsBadReadPtr, IsBadHugeReadPtr, IsBadCodePtr, IsBadStringPtr) # pragma deprecated (memcpy, RtlCopyMemory, CopyMemory, wmemcpy) # else // StrSafe not loaded, so deprecate everything! # pragma deprecated (strcpy, strcpyA, strcpyW, wcscpy, _tcscpy, _mbscpy, StrCpy, StrCpyA, StrCpyW, lstrcpy, lstrcpyA, lstrcpyW, _tccpy, _mbccpy, _ftcscpy) # pragma deprecated (strcat, strcatA, strcatW, wcscat, _tcscat, _mbscat, StrCat, StrCatA, StrCatW, lstrcat, lstrcatA, lstrcatW, StrCatBuff, StrCatBuffA, StrCatBuffW, StrCatChainW, _tccat, _mbccat, _ftcscat) # pragma deprecated (sprintfW, sprintfA, wsprintf, wsprintfW, wsprintfA, sprintf, swprintf, _stprintf) # pragma deprecated (wvsprintf, wvsprintfA, wvsprintfW, vsprintf, _vstprintf, vswprintf) # pragma deprecated (strncpy, wcsncpy, _tcsncpy, _mbsncpy, _mbsnbcpy, StrCpyN, StrCpyNA, StrCpyNW, StrNCpy, strcpynA, StrNCpyA, StrNCpyW, lstrcpyn, lstrcpynA, lstrcpynW) # pragma deprecated (strncat, wcsncat, _tcsncat, _mbsncat, _mbsnbcat, StrCatN, StrCatNA, StrCatNW, StrNCat, StrNCatA, StrNCatW, lstrncat, lstrcatnA, lstrcatnW, lstrcatn) # pragma deprecated (gets, _getts, _gettws) # pragma deprecated (IsBadWritePtr, IsBadHugeWritePtr, IsBadReadPtr, IsBadHugeReadPtr, IsBadCodePtr, IsBadStringPtr) # pragma deprecated (memcpy, RtlCopyMemory, CopyMemory, wmemcpy) # endif //defined(_STRSAFE_H_INCLUDED_) !defined(STRSAFE_NO_DEPRECATE) // SDL 5.0 and later Recommendations # if defined(_SDL_BANNED_RECOMMENDED) # if defined(_STRSAFE_H_INCLUDED_) !defined(STRSAFE_NO_DEPRECATE) // Only deprecate what's not already deprecated by StrSafe # pragma deprecated (wnsprintf, wnsprintfA, wnsprintfW) #
[Libreoffice] [PATCH] Replace List with std::vector
I'm sending the patch for review; not because it needs a lot but just to have someone verify that the memory was leaking and needs to be fixed. Current fixes: 1. XPropertyList::Clear() looks like it has a memory leek; so I fixed it. Future Fixes: 1. List* pBmpList; Should be replaced with a std::vector Bitmap* 2. I thinking that once pBmpList is a vector I should cleanup the constructors that pass initial size and resize values. 3. I'm noticing that several methods have extra parameters that aren't used. Should I include a patch to clean them up also? Thanks, Joe P. 0001-Replace-List-with-std-vector-XPropertyEntry.patch Description: Binary data ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] QA-hints ? - question for devs !
On Sun, Jun 19, 2011 at 3:45 PM, Cor Nouws oo...@nouenoff.nl wrote: Hi *, Is there a script available, or a combination of git-commands, or ... that is used to extract certain information for the summaries? If so, I could use/adapt that to get information on certain weeks, branches, ... ? If you have a git repository then you can get info on that repo that is as fresh as your last pull. The lo-commit-stat script [Petr Mladek] in root/bin is capable of slicing out whatever you want. The top-dir and --log-suffix arguments are required. From your local repo root $ cd bin $ lo-commit-stat --help # gives you the help Use the trailing git-args --since (or --after) and --before (or --until) to pick out a range of times: a last argument of --after=2011-05-31 gives bugs/bugnumbers/commits after that date to the present. Use the --bugs arg to get just the commits with associated issue numbers - the default is all commits. So, in any repo_root/bin directory, the command $ lo-commit-stat ../ --bugs --log-suffix='test' --after=2011-05-31 gives the list of the commits made so far this month that have an issue # in their summary as a log file in repo_root/bin named: bugfixes-current_branch-test.log Again, the list i only as recent as your latest git pull . Hope that helps, LeMoyne Cheers, Cor ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
