date:20110712

On Mon, 2011-07-11 at 21:17 -0700, Joseph Powers wrote:
 Matus,
 
 
 You removed component_getImplementationEnvironment in your patch
 552bb0d5fc2e2d4b0934a9f7f017785962669fbb, While I don't really miss
 the function, some one seems to. From the looks of things, it appears
 like we have a list somewhere that insists that the function should be
 exported. Does anyone know where this list is? (I'm think it would be
 better to fix the list then to just add the code back in.)

It must be from solenv/src/unloadablecomponent.map given
source/animcore/makefile.mk

I suggest you try something like...
coping solenv/src/unloadablecomponent.map to e.g.
solenv/src/unloadablecomponent-merged.map
removing component_getImplementationEnvironment from that copy and
change solenv/src/unloadablecomponent.map in source/animcode/makefile.mk
to solenv/src/unloadablecomponent-merge.map

C.


___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

[Libreoffice] [PUSHED] Accept either 'amd64' or 'x86_64' in the platform id on FreeBSD

On Tue, 2011-07-12 at 00:23 -0700, Mike Eberdt wrote:
 The following patch allows one to configure LO outside of the FreeBSD
 Ports environment as well. Contributed under LGPL3+/MPL.

Pushed now, thanks for this.

C.

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] [PATCH] Implement new rule for xref to numbered paragraphs

On Mon, 2011-07-11 at 23:02 +1000, Troy Rollo wrote:
 ---
  sw/inc/numrule.hxx|   10 ++-
  sw/source/core/doc/number.cxx |   56 
 ++---
  2 files changed, 61 insertions(+), 5 deletions(-)

This looks good, however what does it do :-) Set sane leading/trailing
text/space defaults for a new list numbering level based on the previous
one ?

C.

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] [GSoC][performance] report - week #6

Hi Matus,

On Sun, 2011-07-03 at 22:32 +0200, Matúš Kukan wrote:
This week I'm finally done with school till next year.
I was mostly cleaning some unused components and component_getImplementation
methods.

Cool :-) so - I merged all the patches you sent except the testtool/
performance test removal thing - which ... well ;-) its not doing a
whole lot of harm there I think.

In patch-set there is sample from clone/impress. I've removed those
methods also from other repos and if it's ok I can push that.

Yes - that looks fine - feel free to push all of those :-)

The only method I don't know what to do with is:
http://opengrok.libreoffice.org/xref/libs-core/connectivity/source/drivers/ado/Aservices.cxx#88

94 *ppEnvTypeName = CPPU_CURRENT_LANGUAGE_BINDING_NAME :affine;

That is indeed unusual; I guess it is UNO / appartment related magic,
saying it is thread-affine; I notice we have a libaffine_uno_uno
installed, which presumably is used in this case. It comes from
cppu/source/AffineBridge - which I believe drops that component in it's
own thread [ effectively ] and proxies method calls across from other
threads. Quite how useful that is [ vs. simple mutex protection ] is
unclear to me - but most of the threading code is pretty broken
so ... ;-)

Anyhow - we can leave it for now as long as it is prefixed with a
symbol nicely, so: foo_component_getImplemntation or wahtever.

So ... I guess continuing to remove all the redundant
get_Implementation methods is useful, and of course continuing the
prefixing of component_ methods with baa_component's needs finishing -
where did we get to on that ? it seems that (still) by far the majority
of components are not namespaced: for static linking (for iOS eg.) we're
going to need that completed I think - thoughts on that appreciated.

Thanks,

Michael.

--
michael.me...@novell.com , Pseudo Engineer, itinerant idiot

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

[Libreoffice] [REVIEW] 4 minutes 27 seconds - 11 seconds

cmc-thb: I had some .doc documents which were awesomely slow to load
up, apparently because some polygons had to be split up into large
polypolygons due to a custom dashed line in use to draw them or
something of that nature.

Could you double check my change to basegfx2
http://cgit.freedesktop.org/libreoffice/libs-gui/commit/?id=2f580a32ad5dbb46cd3897316a12aea032b9e10e

I'm *sure* its good, but the scale of speed-up from 4 minutes 27 seconds
to 11 seconds is massively more than I expected. Especially as the
insert point is at the end, rather than at the beginning or middle which
would have required a lot of memmove type operations.

C.

(gcc/libstdcx++ 4.6.0, CXXFLAGS=-Wp,-D_FORTIFY_SOURCE=2
-fstack-protector --param=ssp-buffer-size=4,selinux enabled)

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] Unit tests - cppunit crashing with debug build

2011-07-12 Thread Lubos Lunak

On Monday 11 of July 2011, Caolán McNamara wrote:
 I think the whole mixing code with and without -D_GLIBCXX_DEBUG is sadly
 hopeless. this flag changes the sizes and behavior of standard class
 templates such as std::vector, and therefore you can only link code
 compiled with debug mode and code compiled without debug mode if no
 instantiation of a container is passed between the two translation
 units.

 I reckon we probably have to bite the bullet and drop -D_GLIBCXX_DEBUG
 from solenv ?

 I think we could simply force --without-system-libs (or at least for C++ 
libs) when --enable-debug is used and that could do. I don't how much is 
_GLIBCXX_DEBUG is worth, but --without-system-libs doesn't look like that big 
price to pay.

 p.s. There is another hack e.g. include debug/list vs include list
 which might be work investigating.

 How would that work? The non-debug 3rd-party lib will always use non-debug 
version of list and that means we'd have to manually somehow use non-debug 
version of list as well when interfacing with the library. That sounds pretty 
unfeasible in practice to me.

-- 
 Lubos Lunak
 l.lu...@suse.cz
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] [REVIEW] 4 minutes 27 seconds - 11 seconds

Hi Caolan,

On Tue, 2011-07-12 at 09:27 +0100, Caolán McNamara wrote:
 I'm *sure* its good, but the scale of speed-up from 4 minutes 27 seconds
 to 11 seconds is massively more than I expected.

Oooh - nice :-)

  Especially as the insert point is at the end, rather than at the
 beginning or middle which would have required a lot of memmove
 type operations.

Thorsten is on vacation for another week at least; so ... Radek might
be able to give some sanity checking...

ATB,

Michael.

-- 
 michael.me...@novell.com  , Pseudo Engineer, itinerant idiot


___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

[Libreoffice] component maps without getImpl...

Hi there,

Interestingly, I just got a tinderbox failure on OSX:

On Tue, 2011-07-12 at 08:26 +, nthieb...@gmail.com wrote:
 build failed - error is::
..
 Compiling: animations/source/animcore/targetpropertiescreator.cxx
 /usr/bin/xsltproc --nonet --stringparam uri \
 'vnd.sun.star.expand:$OOO_BASE_DIR/program/libanimcore.dylib' -o 
 ../../unxmacxi.pro/misc/animcore.component \
 /Volumes/Raid0/libo/solenv/bin/createcomponent.xslt animcore.component
 Making:animcore.lib

 cp ../../unxmacxi.pro/misc/unloadablecomponent_animcore.map.exported-symbols 
 ../../unxmacxi.pro/misc/unloadablecomponent_animcore.map
 /Volumes/Raid0/libo/solenv/bin/checkdll.sh -L../../unxmacxi.pro/lib 
 -L/Volumes/Raid0/libo/solver/350/unxmacxi.pro/lib  
 ../../unxmacxi.pro/lib/libanimcore.dylib
 Making:libanimcore.dylib
 Undefined symbols:
   _component_getImplementationEnvironment, referenced from:
  -exported_symbol[s_list] command line option
 ld: symbol(s) not found
 collect2: ld returned 1 exit status

Which seems to come from:

SHL1VERSIONMAP=$(SOLARENV)/src/unloadablecomponent.map

I've added 'nogetimpl' suffixes of both of these to solenv, and
switched to using them. This is a builds fine on Linux, breaks on OSX
issue that I guess we need to make sure we get right, as we remove those
methods :-)

HTH,

Michael.

-- 
 michael.me...@novell.com  , Pseudo Engineer, itinerant idiot

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] Errors building Animation

Hi there,

On Tue, 2011-07-12 at 09:04 +0100, Caolán McNamara wrote:
 It must be from solenv/src/unloadablecomponent.map given
 source/animcore/makefile.mk

I just pushed a fix for this.

Sorry for the interruption - I added a 'nogetimpl' suffix - which is
the right map to use when we've stripped out the redundant getImpl
methods everywhere.

Since we don't see the error on Linux, it is harder to detect, so
perhaps some more will slip through, but the fix in animations/ should
work for them too.

Thanks,

Michael.

-- 
 michael.me...@novell.com  , Pseudo Engineer, itinerant idiot


___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] [GSOC] Report #7: Wizards

Hi Xisco,

First - again nice to hear of your good work :-)

On Tue, 2011-07-12 at 01:24 +0200, Xisco Faulí wrote:
 - Migrate basic to the new resource service rewrote by Björn a while
 ago.

Ooh - right :-) and I assume we've merged this all to master ?

 By the way, I was wondering whether it would be better to convert the
 basic code into python as well. There's some code duplicated in python
 and in basic, it could save us some lines of code. Anyway, I'd like to
 hear your opinion first.

Of course, the priority is to convert Java - python, to reduce our
dependency / warning-on-start problems on Windows, but after that
certainly, I'd personally love to see some StarBasic - Python migration
goodness as well.

ATB,

Michael.

-- 
 michael.me...@novell.com  , Pseudo Engineer, itinerant idiot


___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] [PATCH] Implement new rule for xref to numbered paragraphs

2011-07-12 Thread Troy Rollo

On Tuesday 12 July 2011, Caolán McNamara wrote:
 This looks good, however what does it do :-) Set sane leading/trailing
 text/space defaults for a new list numbering level based on the previous
 one ?

Implements the rules discussed in 
http://lists.freedesktop.org/archives/libreoffice/2011-April/010660.html

(I was busy with study commitments between then and early July - catching up 
now on all the other things I was neglecting then).
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] [REVIEWED] [PUSHED 3-4] patch for fdo#39023: shared formulas are not imported correctly from xlsx, first part

Kohei Yoshida píše v Út 12. 07. 2011 v 00:48 -0400:
 Hello Markus,
 
 On Tue, 2011-07-12 at 04:51 +0200, Markus Mohrhard wrote:
 
  this patch solves the problem that shared formulas are not imported
  correctly from xlsx. The cell containing the shared formula displayed
  =#Name because the range name that is internally used is global but
  the formula compiler thought it is local.
 
 Yes.  The patch makes sense.  Pushed to the -3-4 branch with my
 sign-off.  I believe we haven't branched for 3.4.2 yet (will tomorrow?),
 so this will probably make it for 3.4.2.

Yes, it will be in 3.4.2. I am going to make sure that all patches are
in, do test build, tag, and branch.

Best Regards,
Petr


___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] Remove CREDITS.odt?

Sebastian Spaeth píše v Po 11. 07. 2011 v 14:22 +0200:
 On Mon, 11 Jul 2011 12:24:55 +0100, Michael Meeks michael.me...@novell.com 
 wrote:
   It was planned to create it automatically via a script. For now, I
   simply run it from time to time on my laptop, creating an .html file
   which had been manually converted to .odt.
  
  Ah - so, we can prolly do that during the build with a little effort;
  is your script in git ? perhaps we can add a:
  
  'make prerelease'
  
  type script that will do some checks, and create the updated credits
  file (?). Any chance of a patch for that :-)
 
 My script is currently in bitbucket 
 https://bitbucket.org/spaetz/gitanaly/overview
 and uses mostly Cedric's config files for email aliases and affiliation.
 (git://people.freedesktop.org/~cbosdo/gitdm-lo-config)
 
 I have some local adaptations on who to count as individual and not
 which I promised people to keep private, but that is not needed for the
 credit page generation anyway.
 
 SHM_GET had intended to move things over to the git contrib repo at some
 point and set up the creation on one of the LO servers, but I am not
 sure how far he has gotten there.

It would even make sense to have it bootstrap/bin or bootstrap/solenv.

I like the Michael's idea with 'make prerelease'. I already update
version using the script libreoffice/build/bin/lo-set-version before
doing the release tag. I do not mind to call another script but it
should be easy and easily accessible :-)

Best Regards,
Petr

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

[Libreoffice] [REVIEW] fix for ppt export fdo#32709 (and possibly more issues)

2011-07-12 Thread Radek Doulík

Hi all,

I am looking for 1 review to include the patch in 3-4 branch. It fixes
ppt export and possibly more areas where the code use SvGlobalName class
and  operator.

It is regression in #i101100# strict-alias clean changes.

The original code (before strict alias clean) was using beginning of
pImp-szData in comparison

-else if( *(USHORT *)pImp-szData[ 4 ] == *(USHORT *)rObj.pImp-szData[ 4 
] )
-return *(UINT32 *)pImp-szData  *(UINT32 *)rObj.pImp-szData;

The fix is here

http://cgit.freedesktop.org/libreoffice/libs-gui/commit/?id=345dc7961bc142f167a1b8e5f43f4439e8234f06

commit 345dc7961bc142f167a1b8e5f43f4439e8234f06
Author: Radek Doulik r...@novell.com
Date:   Mon Jul 11 20:36:47 2011 +0200

fix regression in SvGlobalName::operator 

 - it was comparing wrong parts and ommiting part of ID's, resulting
   in wrong results and thus ::std::map didn't work well with default
   less compare function of it's keys
 - fixes fdo#32709

Cheers
Radek


___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] [REVIEWED] + Cherrypick request for 3.4.2

2011-07-12 Thread Muthu Subramanian K

hmm...there are 2 things in pie chart - one to show the values and
second to show the percentages. (the user can choose to display both).
While the data store in impress doesn't seem to hold much of the
formatting information. I assumed it is the responsibility of the
importer (in this case pptx), to interpret the file correctly and tell
impress what to display.
As far as I checked, I couldn't find formatting information being stored
for the series data in pptx (- high probability that I might have missed
this). But, if we have such a provision, we can solve these issues.

(In this case: pptx info says display values, while the series data is
formatted as percentages).

I hope I understood you right ;)

On 07/12/2011 12:39 AM, Michael Meeks wrote:
 
 On Mon, 2011-07-11 at 17:21 +0530, Muthu Subramanian K wrote:
 makes me wonder if we don't want that in our own pie chart
 implementation rather than in the filter though ... :-) But I picked it
 for -3-4.

 Sorry, but I didn't understand this. Did you mean the filter code (oox),
 please?
 
   Sure - can we make our pie chart2 impl. behave in this way, rather than
 having a hack in the filter code to do it ;-)
 
   ATB,
 
   Michael.
 

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] Doubled files: testsax.cxx?

2011-07-12 Thread Thomas Arnhold


Hi Caolán,

I've got it compiled and pushed this to master. As next step I'll check 
if it does something useful.


Thomas
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

[Libreoffice] Running smoketest module standalone.

2011-07-12 Thread Yifan Jiang

Hello all,

I was working on extracting smoketestoo_native module out of the build system,
to investigate the possibility running it with Libreoffice binary build. So
that we could allow QA to run this quick test for a basic level of smoke
testing with a released or pre-released build.

Here is my findings with some questions, I would appreciate if any suggestions
or thinkings are provided.

[Running command]

The smoke test canbe run through a single line command for a specific
binary of libreoffice, provided with the argument -env:arg-soffice :


LD_LIBRARY_PATH=~/smoketest/lib:/opt/libreoffice3.4/ure/lib${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
 \
./cppunittester \

-env:UNO_SERVICES=file:///opt/libreoffice3.4/ure/share/misc/services.rdb \
-env:UNO_TYPES=file:///home/bench/smoketest/types.rdb \
-env:arg-soffice=path:/opt/libreoffice3.4/program/soffice \
-env:arg-user=./user \
-env:arg-env=LD_LIBRARY_PATH${LD_LIBRARY_PATH+=$LD_LIBRARY_PATH} \
-env:arg-testarg.smoketest.doc=~/smoketest/doc/smoketestdoc.sxw \
--protector ~/smoketest/lib/unoexceptionprotector.so 
unoexceptionprotector ~/smoketest/lib/libsmoketest.so

where the ~/smoketest directory contains stuff outside the binary
packages. Namely, beside the libraries provided by ure pakcage, we may need
extra libraries and binary:

Binary:
cppunittester

Library:
libsmoketest.so
libtest.so
unoexceptionprotector.so

I also noticed the services.rdb delivered with ure binary package is Usable
for smoketest:

/opt/libreoffice3.4/ure/share/misc/services.rdb

However the types.rdb delivered with ure binary package will cause a Crash
when smoketesting:

* /opt/libreoffice3.4/ure/share/misc/types.rdb

Then I tried to copy the types.rdb from the master branch:

./unxlngi6.pro/bin/types.rdb

which works fine. Meanwhile, I found another type.rdb in the master build:

./unxlngi6.pro/bin/ure/types.rdb (similar with that in the one delived
in ure package)

whose size is quite smaller than that in the bin/, so I am not sure what's the
difference between them? Would you help me to understand it? Also please
correct me if anything I did is not in the right track. Thank you in advance!

Best wishes,
Yifan

-- 
  Yifan Jiang
  Libreoffice
  Contact: yifan - irc.freenode.net/libreoffice
  =  
  http://www.libreoffice.org/
  http://www.documentfoundation.org/
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

[Libreoffice] [PUSHED] fix for ppt export fdo#32709 (and possibly more issues)

On Tue, 2011-07-12 at 12:09 +0200, Radek Doulík wrote:
 http://cgit.freedesktop.org/libreoffice/libs-gui/commit/?id=345dc7961bc142f167a1b8e5f43f4439e8234f06

mea culpa, oops. Yeah, +1, pushed to 3-4

C.


___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] [REVIEWED][PUSHED:3-4] fix for ppt export fdo#32709 (and possibly more issues)

Radek Doulík píše v Út 12. 07. 2011 v 12:09 +0200:
 Hi all,
 
 I am looking for 1 review to include the patch in 3-4 branch. It fixes
 ppt export and possibly more areas where the code use SvGlobalName class
 and  operator.
 
 It is regression in #i101100# strict-alias clean changes.
 
 The original code (before strict alias clean) was using beginning of
 pImp-szData in comparison
 
 -else if( *(USHORT *)pImp-szData[ 4 ] == *(USHORT *)rObj.pImp-szData[ 
 4 ] )
 -return *(UINT32 *)pImp-szData  *(UINT32 *)rObj.pImp-szData;
 
 The fix is here
 
 http://cgit.freedesktop.org/libreoffice/libs-gui/commit/?id=345dc7961bc142f167a1b8e5f43f4439e8234f06

Makes perfect sense. I see that Caolan already signed if off and pushed
to the 3-4 branch.

Best Regards,
Petr

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] [REVIEW] 4 minutes 27 seconds - 11 seconds

2011-07-12 Thread Radek Doulik

Hi Caolán,

the patch looks good to me.

I am also surprised by the speedup (given that it always inserts at the
end of the vector array in your case). Hard to check without the actual
document. Might be problem in std::vector not detecting that the rest of
reserved space doesn't need to be moved or something?

You might try to load the document into valgrinded libreoffice just to
be sure that it doesn't mess memory.

Cheers
Radek

On Tue, 2011-07-12 at 09:27 +0100, Caolán McNamara wrote:
 cmc-thb: I had some .doc documents which were awesomely slow to load
 up, apparently because some polygons had to be split up into large
 polypolygons due to a custom dashed line in use to draw them or
 something of that nature.
 
 Could you double check my change to basegfx2
 http://cgit.freedesktop.org/libreoffice/libs-gui/commit/?id=2f580a32ad5dbb46cd3897316a12aea032b9e10e
 
 I'm *sure* its good, but the scale of speed-up from 4 minutes 27 seconds
 to 11 seconds is massively more than I expected. Especially as the
 insert point is at the end, rather than at the beginning or middle which
 would have required a lot of memmove type operations.
 
 C.
 
 (gcc/libstdcx++ 4.6.0, CXXFLAGS=-Wp,-D_FORTIFY_SOURCE=2
 -fstack-protector --param=ssp-buffer-size=4,selinux enabled)
 
 ___
 LibreOffice mailing list
 LibreOffice@lists.freedesktop.org
 http://lists.freedesktop.org/mailman/listinfo/libreoffice


___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] Running smoketest module standalone.

On Tue, 2011-07-12 at 18:51 +0800, Yifan Jiang wrote:
 Hello all,
 
 I was working on extracting smoketestoo_native module out of the build system,
 to investigate the possibility running it with Libreoffice binary build. So
 that we could allow QA to run this quick test for a basic level of smoke
 testing with a released or pre-released build.

You could try...
a) just grabbing the smoketest.sxw itself out of a build and manually
open it and click run after turning off macro protection. You don't
need any infrastructure to do that
b) extract out the logic that sets up a temp user dir with macro
protection already turned off in the settings and hack around passing
that to your installed set so that you can leave out clicking the turn
off macro protection step and dig out the command line invocation to
open and automatically run the macro in the smoketest.sxw without manual
intervention.

a is super fast to do.

C.

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] [REVIEW] 4 minutes 27 seconds - 11 seconds

On Tue, 2011-07-12 at 12:51 +0200, Radek Doulik wrote:
 Hi Caolán,
 
 the patch looks good to me.
 
 I am also surprised by the speedup (given that it always inserts at the
 end of the vector array in your case). Hard to check without the actual
 document. Might be problem in std::vector not detecting that the rest of
 reserved space doesn't need to be moved or something?

Yeah, that's sort of what it looks like, though for me the whole
callstack ends up setting up camp in some -DFORTIFY-selinux chain of
goo.

 You might try to load the document into valgrinded libreoffice just to
 be sure that it doesn't mess memory.

Yeah, my test-case is good under valgrind FWIW.

C.

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

[Libreoffice] [REVIEW] regression in cairo canvas setColor

2011-07-12 Thread Radek Doulík

Hi all,

I am looking for 1 review to include the patch in 3-4 branch. It fixes
problem with transparency in cairo canvas.

Fixes fdo#33591 and fdo#35681

The fix is here

http://cgit.freedesktop.org/libreoffice/libs-gui/commit/?id=ac4419bd8a262c6248f06d2124ada924e9b92e99

commit ac4419bd8a262c6248f06d2124ada924e9b92e99
Author: Radek Doulik r...@novell.com
Date:   Tue Jul 12 13:35:43 2011 +0200

fix set color regression in cairo canvas

 - the cairo_set_source_rgba should be called with non-alpha-pre-multiplied
   color values
 - fixes fdo#33591 and fdo#35681

Cheers
Radek



___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] Bug 38831 - [EasyHack] Get rid of SV_DECL_VARARR, SV_DECL_VARARR_PLAIN, SV_DECL_VARARR_SORT ....

2011-07-12 Thread Nigel Hawkins

Hi Maciej,

On Tue, 2011-07-12 at 00:31 +0200, Maciej Rumianowski wrote:
 I picked up this bug and tried to get rid of SvULongs.

I'm the person who's been (slowly) picking away at this, so welcome to
Libreoffice.

 I found for example in ScImportExport::Sylk2Doc and thought it would be
 good to first write a unit test (since there are unit tests of calc).
 But I wasn't successful using ScImportExport::ImportStream and
 ScImportExport::ExportStream, which use Sylk2Doc.

Proper unit tests (as in testing a small piece of functionality) can be
difficult to set up in LO as a lot of stuff is quite tightly coupled.
You can very soon find yourself having to load up the majority of calc
(or writer etc.) to test a very small piece of code. This is not trivial
and I suspect this is the situation you're in.

Check the mailing list archives for the past couple of days - Caolan
McNamara made a post that gave some details about adding unit tests.
That included some examples that might help.

Is it a right way I'm going to start develop LibreOffice or you what me
somewhere else in code?

Anything listed on the EasyHacks page (or marked EasyHack in bugzilla)
is a good place to start. They're all relatively simple but once you've
done a couple you should have a good feel for how the work-flow goes on
this project and how the various components fit together. That will help
a lot when you try something more complex.

Nigel.



___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] [REVIEWED][PUSHED:3-4] regression in cairo canvas setColor

Radek Doulík píše v Út 12. 07. 2011 v 13:40 +0200:
 Hi all,
 
 I am looking for 1 review to include the patch in 3-4 branch. It fixes
 problem with transparency in cairo canvas.
 
 Fixes fdo#33591 and fdo#35681
 
 The fix is here
 
 http://cgit.freedesktop.org/libreoffice/libs-gui/commit/?id=ac4419bd8a262c6248f06d2124ada924e9b92e99

Great catch. Pushed into 3-4 branch. It will be in 3.4.2-rc1.

Best Regards,
Petr

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] [GSOC][patch] Multiline inputbar

2011-07-12 Thread Kohei Yoshida

Hi Anurag,

On Mon, 2011-07-11 at 08:28 +0530, Anurag Jain wrote:

 1: Setting the proper output area when the toggling happens. I've
 tried working on it but was not able to get vertical offset proper
 when in multiline mode. So need to look more into its mathematics.
 
 2: As Noel mentioned in, his last mail that I'm setting the height to
 TBX_WINDOW_HEIGHT instead it can be more than that. I tried fetching
 the current height in ScMultiTextWnd::Resize() and setting the correct
 height but I did some wrong implementation and LO was crashing because
 of that. So I'd need to look into that too.
 
 3: The third thing is, when the resize happens, the ScInputWindow i.e
 toolbox  doesn't resize. The overflow of the child window is hidden
 inside ScInputWindow.

We already talked about the next step in IRC, so this is just a
formality.

We are now focusing on resolving item 3 since without it it's hard to
resolve the other issues.  ScInputWindow is not resizing on button push
because we are not resizing it.  We need to explicitly resize
ScInputWindow on button press since there is no automatic resizing of
any sort with the VCL controls.

When ScInputWindow gets resized, it will in turn resize its immediate
child control ScInputBarGroup, which will resize ScMultiTextWnd, in this
order.  Let's get that resolved first, before looking into the other
issues.

Meanwhile your patch has been committed  pushed to the feature branch.

Kohei

-- 
Kohei Yoshida, LibreOffice hacker, Calc
kyosh...@novell.com

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

[Libreoffice] Don't default to 'Selection' in the print dialog in Writer

2011-07-12 Thread Jan Holesovsky

Hi,

I've committed the following to master:

http://cgit.freedesktop.org/libreoffice/writer/commit/?id=591dfe338ee2e553f822e8bbd3bc602a8ff0d875

This fixes a particularly confusing behavior that you experience when
you want to print something, and have a selection in the document.  In
this case, the print dialog was defaulting to printing the selection
only - which is problematic as printing the selection changes the layout
completely.

The most striking example is when you want to 'deselect' everything in
your document by clicking somewhere.  When you are 'lucky', and click on
a picture, following by trying to print, it will print just the picture
(because it was selected by the click, unlike to clicking to a text that
just positions the cursor, and deselects everything).

Please let me know if this breaks some particularly useful use case; I
hope it does not.  Somebody who wants to print the selection knows what
she wants to achieve, so she is ready to select the 'Selection' option;
this is not true the other way around when somebody selects something
accidentally, and wants to print.

Thank you,
Kendy

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

[Libreoffice] [ANNOUNCE] libreoffice-3.4.2.1 tag created (3.4.2-rc1)

Hi,

there have been created the libreoffice-3.4.2.1 tag for 3.4.2-rc1 release.
The corresponding official builds will be available by the end of the week.

See the attached list of changes against 3.4.1-final.

I am going to create the libreoffice-3-4-2 branch for the final stabilization.

You might switch your current 3-4 source tree to it using:

./g fetch --tags
./g checkout -b tag-libreoffice-3.4.2.1 libreoffice-3.4.2.1

See also the schedule at 
http://wiki.documentfoundation.org/ReleasePlan#3.4_release
and release criteria at http://wiki.documentfoundation.org/Release_Criteria


Best Regards,
Petr
+ base
+ don't crash when selecting target to db document, . (fdo#36545) [Jan Holesovsky]
+ bootstrap
+ sorting additional language packs and dictionaries (fdo#34479) [Andras Timar]
+ calc
+ always set column offset to 0 (fdo#38456) [Katarina Machalkova]
+ better fix for (fdo#38455) [Markus Mohrhard]
+ fix for (fdo#37880) [Markus Mohrhard]
+ fix for don't call vector::back on an empty vector (fdo#38593) [Markus Mohrhard]
+ fix for enable Object/Edit menu item for ole/embedded objects (fdo#38544) [Noel Power]
+ fix for formulas with range names were not imported correctly (fdo#38204) [Markus Mohrhard]
+ fix for name box always selects cell in the first sheet (fdo#36963) [Markus Mohrhard]
+ fix for update local range names (fdo#38380) [Markus Mohrhard]
+ revert Fixed undisplayed calc page and header / footer borders (fdo#38590, fdo#36688) [Petr Mladek]
+ components
+ fix label of Save Keyboard Configuration dialog (fdo#31353) [Andras Timar]
+ filters
+ apparent negative string length (fdo#34023) [Caolán McNamara]
+ do not set solid line border and fill as default for chartarea (bnc#694392) [Radek Doulik]
+ fix for (bnc#693386) [Noel Power]
+ fix for shared formulas are not imported correctly (fdo#39023) [Markus Mohrhard]
+ fix label position for ooxml charts (bnc#694340) [Thorsten Behrens]
+ using formatcode while displaying labels. (bnc#694356) [Muthu Subramanian]
+ help
+ fix broken image paths in help (fdo#37509) [Andras Timar]
+ revert Update the operation of delete and backspace key  (fdo#31652) [Andras Timar]
+ update the operation of delete and backspace key (fdo#31652) [Korrawit Pruegsanusak]
+ impress
+ selecting text box after editing gives rotation handles (fdo#37482) [Luke Symes]
+ libs-core
+ add useronly read permission, while creating tmpfile. (fdo#36733) [Michael Meeks]
+ fix --quickstart shouldn't show start center (fdo#37110) [Robert Nagy]
+ fix for (fdo#37880) [Markus Mohrhard]
+ fixed a crasher. Also, improved batch conversion (bnc#690798, fdo#37412) [Muthu Subramanian]
+ related: ole props buglet (i#35066) [Caolán McNamara]
+ set default min password length in ctor of password dialog (fdo#37771) [Andras Timar]
+ unable to save protected file without pw (fdo#38561) [Markus Mohrhard]
+ use UTF-8 encoding for gstreamer uris (fdo#39091) [Caolán McNamara]
+ libs-gui
+ context menu in spellcheck in RTL UI closes too early (fdo#33509) [Caolán McNamara]
+ crash in number format dialog (fdo#38963) [Markus Mohrhard]
+ duplicate entries in Directory definition are no longer treated as errors (bnc#682484) [Katarina Machalkova]
+ fix regression in SvGlobalName::operator  (fdo#32709) [Radek Doulik]
+ fix set color regression in cairo canvas (fdo#33591, fdo#35681) [Radek Doulik]
+ make sure we send resize event to fullscreen windows (i#86302) [Radek Doulik]
+ related: busted ole2 wrapper (i#46388) [Caolán McNamara]
+ related: make simpleMapNativeRoleFrom::accessibleContext safer (fdo#38704) [Caolán McNamara]
+ writer
+ additional fix for bug (fdo#33781) [Roland Baudin]
+ avoid infinite loop when writting out docs with bookmarks (bnc#700370) [Luboš Luňák]
+ basic handling of lines in RTF import (bnc#695479) [Cédric Bosdonnat]
+ copy fieldmark properties when copying/pasting, . (fdo#38374) [Jan Holesovsky]
+ fix the computation of grid character pitch, . (fdo#37516) [Jan Holesovsky]
+ handle wrong text len gracefully (i#95550) [Caolán McNamara]
+ hang on load of .doc (fdo#34023) [Caolán McNamara]
+ imported .doc crashes layout (fdo#38623) [Caolán McNamara]
+ more conservative fix for aliasing problems (i#99665, fdo#33781) [Petr Mladek]
+ protect against missing insertion position (i#105349) [Caolán McNamara]
+ related: make abo levels safer (i#86988) [Caolán McNamara]
+ related: use cached data values instead of pointer to sprm cache (i#115768) [Caolán McNamara]
+ related: use standard sprm iterator (fdo#37057) [Caolán McNamara]
+ common
+ version 3.4.2.1, tag libreoffice-3.4.2.1 (3.4.2-rc1) [Petr Mladek]
+ base
+ don't crash when selecting target to db document, . (fdo#36545) [Jan Holesovsky]
+ bootstrap
+ --enable-online-update in the Linux universal builds. [Jan Holesovsky]
+ better

[Libreoffice] [ANNOUNCE] Branch libreoffice-3-4-2 created

Hi all,

there have been created the libreoffice-3-4-2 branch. It will be used for fine
tuning of the 3.4.2 bugfix release. It is based on the the tag
libreoffice-3.4.2.1 for 3.4.2-rc1 release.

The following rules apply:

+ preferably just translation or blocker fixes
+ only cherry-picking from libreoffice-3-4 branch
+ 2 additional reviews needed; 2nd reviewer pushes
+ no regular merges back to anything

The 'libreoffice-3-4' branch is still active, will be used for the next
bugfix release (3.4.3). Please read more at

   http://wiki.documentfoundation.org/ReleasePlan
   http://wiki.documentfoundation.org/Development/Branches


Now, if you want to switch your clone to the libreoffice-3-4-2 branch, please
do:

./g pull -r
./g checkout -b libreoffice-3-4-2 origin/libreoffice-3-4-2

Hopefully it will work for you :-)  Most probably, you will also want to
do (if you haven't done it yet):

git config --global push.default tracking

When you do git push with this, git will push only the branch you are
on; e.g. libreoffice-3-4-2 when you have switched to it.  This will
save you some git shouting at you.


Happy hacking,
Petr

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

[Libreoffice] gimli.documentfoundation.org down

2011-07-12 Thread Jan Holesovsky

Hi,

Unfortunately, gimli is down due to a failing disk :-(  This means the
following services are currently unavailable:

dev-builds.libreoffice.org
help.libreoffice.org
opengrok.libreoffice.org
tinderbox.libreoffice.org

The admins assured me that the data should survive, let's hope it is
true.

Regards,
Kendy

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] gimli.documentfoundation.org down

2011-07-12 Thread Jan Holesovsky

Hi,

On 2011-07-12 at 16:50 +0200, Jan Holesovsky wrote:

 Unfortunately, gimli is down due to a failing disk :-(  This means the
 following services are currently unavailable:
 
 dev-builds.libreoffice.org
 help.libreoffice.org
 opengrok.libreoffice.org
 tinderbox.libreoffice.org
 
 The admins assured me that the data should survive, let's hope it is
 true.

The machine is up again, and it seems that the services are running;
should you have any problems, please let me know.

Regards,
Kendy

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

[Libreoffice] development summary: year 2011, week 27

Hi,

this time a brief summary of what happened during the 27th week in 2011
on LibreOffice repositories and the living branches:

+ master:LO-3.5 development
+ libreoffice-3-3:   fixes for LO-3.3.4 bug fix release
+ libreoffice-3-4:   fixes for LO-3.4.2 bug fix release

There are two logs for each branch:

+ bugfixes-branch-week-year-week.txt lists all commits that 
reference a proper
  bug id from a variety of trackers, i.e. #i... referring to the 
OpenOffice
  issuezilla, fdo# to freedesktop, rhbz# to RedHat bugzilla

+ commit-log-branch-week-year-week.txt lists all relevant commits 
on the actual
  source repositories

Many thanks to all contributors - you make all the difference!


Best Regards,
Petr
nothing this week
+ calc
+ fix for (fdo#37880) [Markus Mohrhard]
+ fix for update local range names (fdo#38380) [Markus Mohrhard]
+ filters
+ apparent negative string length (fdo#34023) [Caolán McNamara]
+ do not set solid line border and fill as default for chartarea (bnc#694392) [Radek Doulik]
+ fix for (bnc#693386) [Noel Power]
+ fix label position for ooxml charts (bnc#694340) [Thorsten Behrens]
+ using formatcode while displaying labels. (bnc#694356) [Muthu Subramanian]
+ impress
+ selecting text box after editing gives rotation handles (fdo#37482) [Luke Symes]
+ libs-core
+ add useronly read permission, while creating tmpfile. (fdo#36733) [Michael Meeks]
+ fix for (fdo#37880) [Markus Mohrhard]
+ related: ole props buglet (i#35066) [Caolán McNamara]
+ use UTF-8 encoding for gstreamer uris (fdo#39091) [Caolán McNamara]
+ libs-gui
+ crash in number format dialog (fdo#38963) [Markus Mohrhard]
+ duplicate entries in Directory definition are no longer treated as errors (bnc#682484) [Katarina Machalkova]
+ related: busted ole2 wrapper (i#46388) [Caolán McNamara]
+ writer
+ additional fix for bug (fdo#33781) [Roland Baudin]
+ basic handling of lines in RTF import (bnc#695479) [Cédric Bosdonnat]
+ handle wrong text len gracefully (i#95550) [Caolán McNamara]
+ hang on load of .doc (fdo#34023) [Caolán McNamara]
+ more conservative fix for aliasing problems (i#99665, fdo#33781) [Petr Mladek]
+ protect against missing insertion position (i#105349) [Caolán McNamara]
+ related: make abo levels safer (i#86988) [Caolán McNamara]
+ related: use cached data values instead of pointer to sprm cache (i#115768) [Caolán McNamara]
+ related: use standard sprm iterator (fdo#37057) [Caolán McNamara]
+ calc
+ fix for Option to display calc gridlines on colored cells (fdo#30800) [André Schnabel]
+ filters
+ fix label position for ooxml charts (bnc#694340) [Thorsten Behrens]
+ using formatcode while displaying labels. (bnc#694356) [Muthu Subramanian]
+ help
+ help for help for changed grid line option (fdo#30800) [André Schnabel]
+ libs-core
+ add useronly read permission, while creating tmpfile. (fdo#36733) [Muthu Subramanian]
+ call the handler once all changes are done on FrameSelector (fdo#32257) [Cédric Bosdonnat]
+ fix for Option to display grid lines on colored cells (fdo#30800) [André Schnabel]
+ fixed crasher by using boost::shared_ptr (fdo#37403) [Cédric Bosdonnat]
+ long/int mix causes 32bit LONG_MAX to be 64bit -1 (fdo#34450) [Caolán McNamara]
+ related: ole props buglet (i#35066) [Caolán McNamara]
+ use UTF-8 encoding for gstreamer uris (fdo#39091) [Caolán McNamara]
+ libs-gui
+ crash in number format dialog (fdo#38963) [Markus Mohrhard]
+ duplicate entries in Directory definition are no longer treated as errors (bnc#682484) [Katarina Machalkova]
+ writer
+ copy fieldmark properties when copying/pasting, . (fdo#38374) [Jan Holesovsky]
+ crash on 0 width FrmFmt table cell layout (fdo#33012) [Caolán McNamara]
+ more conservative fix for aliasing problems (i#99665, fdo#33781) [Petr Mladek]
+ related: crash in SwTxtSizeInfo::GetMultiCreator (rhbz#718976) [Caolán McNamara]
+ related: make abo levels safer (i#86988) [Caolán McNamara]
+ related: use cached data values instead of pointer to sprm cache (i#115768) [Caolán McNamara]
+ writer: bug correction : selected text was always replaced (fdo#36181) [Dolives Benoit]
+ wrong word counts for USHRT_MAX len paras (fdo#34779) [Caolán McNamara]
+ libs-gui
+ handle busted emf lengths [Caolán McNamara]
+ bootstrap
+ add distro-configs/SUSE-*.conf [Petr Mladek]
+ better support for distro packaging [Petr Mladek]
+ bin/unpack-sources: new helper script for packaging LO [Petr Mladek]
+ calc
+ fix for (fdo#37880) [Markus Mohrhard]
+ fix for update local range names (fdo#38380) [Markus Mohrhard]
+ make sure to only export up to 1026 manual page breaks to XLS. [Kohei Yoshida]
+ components
+ fix --enable-dbgutil build [Caolán McNamara]
+ fix wrong capitalization in case of non-standard hyphenation

Re: [Libreoffice] Extension-Repository: Need help to solve a Python error

2011-07-12 Thread Andreas Mantke

Hi Caolan, *,

Am Montag, 11. Juli 2011, 18:27:09 schrieb Caolán McNamara:
 On Tue, 2011-07-05 at 21:14 +0200, Andreas Mantke wrote:
(...)
 given that a contenttype is required, some total guesswork suggests...
 
 -blob = BlobWrapper()
 -if isinstance(value, basestring):
 -# make StringIO from string, because StringIO may be adapted to
 -#Blobabble
 -value = StringIO(value)
 -if value is not None:
 -blobbable = IBlobbable(value)
 -blobbable.feed(blob.getBlob())
 -blob.setContentType(blobbable.mimetype())
 
 might need to become
 
 +if isinstance(value, basestring):
 +# make StringIO from string, because StringIO may be adapted to
 +#Blobabble
 +value = StringIO(value)
 +if value is not None:
 +blobbable = IBlobbable(value)
 +#move blob ctor down to where we know the mimetype
 +blob = BlobWrapper(blobbable.mimetype())
 +blobbable.feed(blob.getBlob())
 +blob.setContentType(blobbable.mimetype()) #maybe drop this then
 
 or something of that nature, you get my drift anyway. That the
 super-class BlobWrapper may take an additional argument in your plone
 over the version that the template that you based the above on works
 with.

thanks, your patch proposal solved the issue for me. I commited the patch to 
the 
Plone community repository and it is now inside our extension-template-website. 
We 
can store binaries (extensions / templates) inside the file system now.

Regards,
Andreas
-- 
## Developer LibreOffice
## Freie Office-Suite für Linux, Mac, Windows
## http://LibreOffice.org
## Support the Document Foundation (http://documentfoundation.org)
## Meine Seite: http://www.amantke.de 
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: [Libreoffice] Remove CREDITS.odt?