Re: OpenOffice development with Eclipse

2013-12-19 Thread Jürgen Schmidt
On 12/19/13 7:54 AM, Abhishek Kumar wrote:
 Hello Andre and Jürgen,
 
 I have recently started working with OpenOffice, and wanted to get into
 deeper details of the massive codebase.
 
 Although I am used to working with  vim+gnuToochain, but given the codebase
 size, an IDE for indexing and viewing the code seems to be a good way
 forward for someone new to this codebase. Eclipse as specified here seems
 to be quite good for debugging and viewing code.
 
 Is there some standard IDE that is used for debugging/editing core OO code
 in SC/SD/WriterFilter etc?

on Windows the MS Dev Studio Express Edition is used normally, the Ms
debugger is the best I have seen so far. On Mac I use gdb in Emacs or
Xcode. Especially with the new Mac env I will use Xcode. On Linux I
haven't seen a very good tool wrapping the gdb. I use gdb in Emacs or
via  command line but will try in the future Eclipse.

Juergen


 
 Thanks,
 Abhishek
 
 
 On Tue, Dec 17, 2013 at 3:04 PM, Jürgen Schmidt jogischm...@gmail.comwrote:
 
 On 12/17/13 9:58 AM, Andre Fischer wrote:
 Hi,

 About a week or so ago I tried again to import the OpenOffice source
 code into Eclipse.  The last time I tried that, a few years ago, this
 did not work at all.  Our source code was just to big and too complex.
 This time however, it worked out of the box.  If you do it right :-)
 The developers of the CDT (the C/C++ Development Tooling) have made
 great progress in the last years.

 As this allows newcomers a much easier way to become OpenOffice
 developers and also can make the work of experienced developers more
 productive I would like to share how to set up Eclipse for OpenOffice.
 This process is still a bit rough around the edges. Any help is
 appreciated.

 I have started a Wiki page [1] which explains the manual setup. It
 should work on all platforms but at the moment only Windows is
 described.  Maybe you can help to fill in missing values for other
 platforms.  I am working on a few Eclipse plugins to automate this
 process:

 - A wizard for setting up an Eclipse workspace.  The prototype has just
 to be pointed to the (compiled) source code.  It will then collect all
 the necessary include paths and set the necessary compiler defines so
 that the indexer can resolve even the platform dependent definitions.
 This is important because even something as simple as sal_Int16 will not
 be recognized without the indexer being told about platform specific
 preprocessor defines.

 - The idea of CDT about how C/C++ files are built is different from how
 OpenOffice does it.  Therefore I am developing a short term solution
 that provides a couple of buttons/menu entries/whatever will be
 appropriate to build the current module, directory or file and then
 deploy the new libraries into an installed office.


 Here are some highlight of what Eclipse and CDT allow us to do:

 - Easily switch between header and implementation file.

 - Follow header inclusion (just Ctrl-click on an #include statement and
 the referenced header file is opened in the editor.

 - Get information about variables and classes by simply hovering over
 their names.  A popup box shows you the variable definition or the class
 implementation.  That is something you have to see to believe it.

 - Find out from where a function or method is called

 - Explore the type hierarchy of a class.

 - Macro expansion.  Hover over a macro use and see the expansion of it,
 several iterations if necessary. Ever wanted to know how IMPL_LINK is
 defined?  Just move your mouse pointer over it.

 - Code assist while typing.  Define a variable, eg.
   ::rtl::OUString sText
   Then type sText., wait a little (default is 500ms) and a box shows
 you all the available methods of OUString.

 - An overview at the right side shows you all the functions, methods,
 includes, macro defines etc, that are made in the currently edited
 file.  Click on one and jump to the corresponding source code.


 All of this may sound familiar to Java developers.  But for C++,
 especially for OpenOffice, I think this is quite remarkable.

 And to underline what Andre has described above, this is really cool and
 we are sure it will help us a lot.

 I watched a video from Bjoern Michaelsen from LO where he presented
 something similar for KDevelop and pointed out what's possible with
 their new complete gbuild system  I didn't understand the relation
 between the build system and the integration and talked with Andre about
 and others about it.

 Andre who is quite familiar with Eclipse (I am more a NetBeans guy)
 spend a first hour on experimenting with Eclipse. We didn't knew that
 and after an hour he presented us the first results which were impressive.

 I did a similar test on my Mac with the same result and was able to
 compile and debug (old gcc env) in Eclipse. Well I had some problems
 with the new upcoming env (clang, lldb) because there is no lldb tool
 chain for Eclipse yet. But anyway the whole support and the 

Re: createRegistryServiceFactory: cannot get uno environments!

2013-12-19 Thread Herbert Duerr

Hello Απόστολος,

On 18.12.2013 18:39, Απόστολος Συρόπουλος wrote:


Anyway, if we're lucky the problem is just a missing library or the failed 
loading
of a library. In that case the patch below might help you to analyze the 
problem:



Indeed the problem is a missing library:

:  
LD_LIBRARY_PATH=/extra/sources/OpenOffice/aoo4/main/solver/410/unxsogi.pro/lib${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
  ../../../unxsogi.pro/bin/saxparser en_AU en_AU.xml 
../../../unxsogi.pro/misc/localedata_en_AU.cxx \
 
file:///extra/sources/OpenOffice/aoo4/main/i18npool/source/localedata/data/../../../unxsogi.pro/misc/saxparser.rdb
 /extra/sources/OpenOffice/aoo4/main/solver/410/unxsogi.pro/bin/types.rdb \
 
-env:OOO_INBUILD_SHAREDLIB_DIR=file:///extra/sources/OpenOffice/aoo4/main/solver/410/unxsogi.pro/lib
loadModule( 
/extra/sources/OpenOffice/aoo4/main/solver/410/unxsogi.pro/lib/libg++_uno.so, 
257) failed with ld.so.1: saxparser: fatal: 
/extra/sources/OpenOffice/aoo4/main/solver/410/unxsogi.pro/lib/libg++_uno.so: open 
failed: No such file or directory
loadModule( 
/extra/sources/OpenOffice/aoo4/main/solver/410/unxsogi.pro/lib/libg++_uno.so, 
257) failed with ld.so.1: saxparser: fatal: 
/extra/sources/OpenOffice/aoo4/main/solver/410/unxsogi.pro/lib/libg++_uno.so: open 
failed: No such file or directory
Exception on createRegistryServiceFactory: cannot get uno environments!
dmake:  Error code 1, while making 
'../../../unxsogi.pro/misc/localedata_en_AU.cxx'

Google returns nothing about libg++_uno...


A library like that should be automatically built as one of the bridges 
in main/bridge/source/cppu_uno/. In your case the gcc3_solaris_intel 
bridge might be a good candidate but as the name says it probably builds 
a libgcc3_uno.so instead of the libg++_uno.so your other binaries 
expect. Check the makefile.mk in the bridge. Or find out why your 
binaries look for a libg++_uno.so instead of a libgcc3_uno.so.


Herbert

-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



Re: OpenOffice development with Eclipse

2013-12-19 Thread Andre Fischer

Am 18.12.2013 20:33, schrieb Andrea Pescetti:

Jürgen Schmidt wrote:

We will blog about
it in the new year and I ma sure we can then present a little bit more.


This may be an interesting topic for FOSDEM too and I think you (Andre 
included, obviously!) already have quite clear ideas, so you could 
submit it as explained in

https://blogs.apache.org/OOo/entry/join_apache_openoffice_at_fosdem
by the Dec 22 deadline... unless you have already done it!


I have written two proposals, and the Eclipse support is one of them.

I am also working towards a blog post.  But first I would like to finish 
a couple of Eclipse addons that make the OpenOffice integration a little 
easier.  Christmas may slow me down a bit :-)


In the mean time there is the Wiki page, now with some screenshots and a 
section about the Mylyn (Bugzilla) integration.


-Andre



Regards,
  Andrea.

-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org




-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



Re: OpenOffice development with Eclipse

2013-12-19 Thread Oliver-Rainer Wittmann

Hi,

On 17.12.2013 09:58, Andre Fischer wrote:

Hi,

About a week or so ago I tried again to import the OpenOffice source
code into Eclipse.  The last time I tried that, a few years ago, this
did not work at all.  Our source code was just to big and too complex.
This time however, it worked out of the box.  If you do it right :-) The
developers of the CDT (the C/C++ Development Tooling) have made great
progress in the last years.

As this allows newcomers a much easier way to become OpenOffice
developers and also can make the work of experienced developers more
productive I would like to share how to set up Eclipse for OpenOffice.
This process is still a bit rough around the edges. Any help is
appreciated.

I have started a Wiki page [1] which explains the manual setup. It
should work on all platforms but at the moment only Windows is
described.  Maybe you can help to fill in missing values for other
platforms.  I am working on a few Eclipse plugins to automate this process:

- A wizard for setting up an Eclipse workspace.  The prototype has just
to be pointed to the (compiled) source code.  It will then collect all
the necessary include paths and set the necessary compiler defines so
that the indexer can resolve even the platform dependent definitions.
This is important because even something as simple as sal_Int16 will not
be recognized without the indexer being told about platform specific
preprocessor defines.

- The idea of CDT about how C/C++ files are built is different from how
OpenOffice does it.  Therefore I am developing a short term solution
that provides a couple of buttons/menu entries/whatever will be
appropriate to build the current module, directory or file and then
deploy the new libraries into an installed office.


Here are some highlight of what Eclipse and CDT allow us to do:

- Easily switch between header and implementation file.

- Follow header inclusion (just Ctrl-click on an #include statement and
the referenced header file is opened in the editor.

- Get information about variables and classes by simply hovering over
their names.  A popup box shows you the variable definition or the class
implementation.  That is something you have to see to believe it.

- Find out from where a function or method is called

- Explore the type hierarchy of a class.

- Macro expansion.  Hover over a macro use and see the expansion of it,
several iterations if necessary. Ever wanted to know how IMPL_LINK is
defined?  Just move your mouse pointer over it.

- Code assist while typing.  Define a variable, eg.
   ::rtl::OUString sText
   Then type sText., wait a little (default is 500ms) and a box shows
you all the available methods of OUString.

- An overview at the right side shows you all the functions, methods,
includes, macro defines etc, that are made in the currently edited
file.  Click on one and jump to the corresponding source code.


All of this may sound familiar to Java developers.  But for C++,
especially for OpenOffice, I think this is quite remarkable.



Yes, it is - I will definitely try it out.

Thanks for the sharing your findings.


Best regards, Oliver.

-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



Re: AOO is now also available as GIT (READ-ONLY).

2013-12-19 Thread jan i
On 19 December 2013 11:22, Herbert Duerr h...@apache.org wrote:

 On 18.12.2013 18:38, jan i wrote:

 Infra-5590 has now been solved, thanks to an effort from jfarrell:

 jfarrell 16:11:32 janIV: christmas gift for you. took a little over a
 week to wrap, had to use duct tape along the way. git ls-remote git://
 git.apache.org/openoffice.git


 Please have a look at our new READ-ONLY git mirror.


 That's great news, thanks! Kudos to Jake. Having a reference
 git-repository will allow us to do some nifty things:

 I'll prepare some git-packs with pre-AOO history from its older Mercurial,
 Subversion and maybe even from its CVS-based open-source ancestors. Using
 git's graft facilities this will make it much easier to research the
 history of parts of the codebase where noone is left to tell the story or
 its details. Having the individual commit comments and their issue numbers
 is invaluable then.

 Is there a gitweb or similar frontend to our git-mirror? Then we could
 e.g. provide a link from the nightly builds to their respective version
 history.

 As expected in my JIRA-request the two-level deep svn-branches (i.e.
 alg/aw080 and alg/sysdepgs) have been munged into one git-branch...

We have anyhow to consider it the GIT branch is a practial solution, I just
did a branch switch and encountered 2 problem:
1. The switch alone to 20 minutes
2. I had to rebuilt everything because (in my case) unxlngx6.pro is used
for trunk and branch.

We can with a relative simple patch platform= to configure avoid 2, but 1
is going to be PITN.



 Also the symphony branch is missing because it was checked into svn at a
 non-canonic location so the git-mirror couldn't pick it up.

If those branches are changed to normal, the mirror should pick it up.




  Once everybody is happy, we can consider moving fully to GIT, I will
 however recommend to it imidiatly after next release.


 I'll have to adjust the svn2bz bot, we'll have to adjust the buildbots and
 there are some other scripts that have to be updated before we can think
 about a full switch.

 But I look forward to git's proper author attributions, interactive
 rebasing, intact individual commits after merges, nice cherry-picks, etc.

I think we should have a discussion on the list. I am convinced that git
offer some advantages but also some disadvantages (like branch switching)
and I feel everybody should have a clear view of pro/con before we decide.

rgds
jan I.



 Herbert


 -
 To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
 For additional commands, e-mail: dev-h...@openoffice.apache.org




Re: AOO is now also available as GIT (READ-ONLY).

2013-12-19 Thread Herbert Duerr

Hello Jan,

On 19.12.2013 15:37, jan i wrote:

We have anyhow to consider it the GIT branch is a practial solution, I just
did a branch switch and encountered 2 problem:
1. The switch alone to 20 minutes


A svn switch between two branches takes about the same time, doesn't it? 
Window's file access is unfortunately quite slow. On Linux or on Mac the 
filesystems are no bottleneck so checkouts don't take long.



2. I had to rebuilt everything because (in my case) unxlngx6.pro is used
for trunk and branch.


Well, if trunk and branch are incompatible (i.e. some project wide 
headers changed, build options changed or there are binary 
incompatibilites between some libs) then you need clean rebuilds or 
different output trees regardless of the version control system. Does 
subversion have some magic that makes such incompatibilities go away? If 
there is such a magic then that would certainly be nobel price worthy ;-)


But maybe you are using a different setup for your svn repos? Do you 
have one checkout per branch you are interested in? You can do the same 
with git of course and with git-clone's --reference option you can even 
share the meat of the different repositories, which is something very 
nice that I missed in Mercurial or now in Subversion.



We can with a relative simple patch platform= to configure avoid 2, but 1
is going to be PITN.




Also the symphony branch is missing because it was checked into svn at a
non-canonic location so the git-mirror couldn't pick it up.


If those branches are changed to normal, the mirror should pick it up.


Yes.


But I look forward to git's proper author attributions, interactive
rebasing, intact individual commits after merges, nice cherry-picks, etc.


I think we should have a discussion on the list. I am convinced that git
offer some advantages but also some disadvantages (like branch switching)
and I feel everybody should have a clear view of pro/con before we decide.


+1

Herbert

-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



switch trunk from Mac 32bit to 64bit

2013-12-19 Thread Herbert Duerr
The new Mac port looks quite good. I uploaded a current version to my 
page [1]. Jürgen already mentioned it will only work for OSX 10.7 and 
up. It is based on todays trunk, which already contains a lot of fixes 
and enhancements compared to our latest release. For details you can 
have a look at our progress tracking page [2].


[1] http://people.apache.org/~hdu/
[2] http://people.apache.org/~hdu/izlist9.htm

In the early days of next year I plan to update our trunk so the new 
port becomes active. To build it yourself you'll need XCode4 then. 
XCode4 comes with the 10.7 SDK.


Herbert


-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



Re: AOO is now also available as GIT (READ-ONLY).

2013-12-19 Thread jan i
On 19 December 2013 16:28, Herbert Duerr h...@apache.org wrote:

 Hello Jan,


 On 19.12.2013 15:37, jan i wrote:

 We have anyhow to consider it the GIT branch is a practial solution, I
 just
 did a branch switch and encountered 2 problem:
 1. The switch alone to 20 minutes


 A svn switch between two branches takes about the same time, doesn't it?
 Window's file access is unfortunately quite slow. On Linux or on Mac the
 filesystems are no bottleneck so checkouts don't take long.


The difference (of course I might use git wrongly), is the in svn, the
branch is stored in its directory. With git I same disk space (trunk and
branches are in 1 directory) but it needs to write a lot of files.




  2. I had to rebuilt everything because (in my case) unxlngx6.pro is used
 for trunk and branch.


 Well, if trunk and branch are incompatible (i.e. some project wide headers
 changed, build options changed or there are binary incompatibilites between
 some libs) then you need clean rebuilds or different output trees
 regardless of the version control system. Does subversion have some magic
 that makes such incompatibilities go away? If there is such a magic then
 that would certainly be nobel price worthy ;-)

He, I think it boils down to what I wrote in 1). Git standard keeps
trunk/branches in one directory, and for small projects its very fast to
switch.

I have managed with a trix to have a git trunk in one directory and a git
branch in another directory. However when I read the git manual that is not
how its supposed to work.



 But maybe you are using a different setup for your svn repos? Do you have
 one checkout per branch you are interested in? You can do the same with git
 of course and with git-clone's --reference option you can even share the
 meat of the different repositories, which is something very nice that I
 missed in Mercurial or now in Subversion.


I have the setup:
 /x1/aoo/trunk /x1/aoo/branches
so one big aoo directory with everything (I have asked svn to ignore the
branches am not interested in).

But please bear with me, I am no git expert, just read the manual and try
to follow it.

rgds
jan I.




  We can with a relative simple patch platform= to configure avoid 2, but
 1
 is going to be PITN.



 Also the symphony branch is missing because it was checked into svn at a
 non-canonic location so the git-mirror couldn't pick it up.


 If those branches are changed to normal, the mirror should pick it up.


 Yes.


  But I look forward to git's proper author attributions, interactive
 rebasing, intact individual commits after merges, nice cherry-picks, etc.

  I think we should have a discussion on the list. I am convinced that git
 offer some advantages but also some disadvantages (like branch switching)
 and I feel everybody should have a clear view of pro/con before we decide.


 +1


 Herbert

 -
 To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
 For additional commands, e-mail: dev-h...@openoffice.apache.org




How to distinguish whether a 32 or 64 bit AOO is installed on systems that may host both bitnesses ? (Re: switch trunk from Mac 32bit to 64bit

2013-12-19 Thread Rony G. Flatscher (Apache)
Congratulations for allowing AOO be compiled in 64 bit for the MacOSX!

Is or will be there a possiblity to find out which version of AOO is installed, 
the 32 or the 64 bit
version?

There is at least one (mine ;) ) third party extension that may try to add an 
extension to AOO. As
the app can only be installed in either 32 or 64 bit, the extension support for 
AOO should only be
installed, if bitnesses between both match. So on operating systems where AOO 
can be installed
either as 32 or as 64 bit it would be necessary to figure out (in a platform 
independent way) what
bitness the installed AOO carries.

As unoinfo seems to not carry the bitness information, is there another 
(portable, platform
independent) way to find out what bitness the installed AOO has?

TIA,

---rony

P.S.: Expecting that after all PCs migrated from 32 to 64 bit, that then the 
migration to 128 will
start, such that this remains an interesting piece of information for the times 
to come.



On 19.12.2013 17:29, Herbert Duerr wrote:
 The new Mac port looks quite good. I uploaded a current version to my page 
 [1]. Jürgen already
 mentioned it will only work for OSX 10.7 and up. It is based on todays trunk, 
 which already
 contains a lot of fixes and enhancements compared to our latest release. For 
 details you can have
 a look at our progress tracking page [2].

 [1] http://people.apache.org/~hdu/
 [2] http://people.apache.org/~hdu/izlist9.htm

 In the early days of next year I plan to update our trunk so the new port 
 becomes active. To build
 it yourself you'll need XCode4 then. XCode4 comes with the 10.7 SDK.


-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



Re: switch trunk from Mac 32bit to 64bit

2013-12-19 Thread jan i
On 19 December 2013 17:29, Herbert Duerr hdu_...@alice.de wrote:

 The new Mac port looks quite good. I uploaded a current version to my page
 [1]. Jürgen already mentioned it will only work for OSX 10.7 and up. It is
 based on todays trunk, which already contains a lot of fixes and
 enhancements compared to our latest release. For details you can have a
 look at our progress tracking page [2].

 [1] http://people.apache.org/~hdu/
 [2] http://people.apache.org/~hdu/izlist9.htm

 In the early days of next year I plan to update our trunk so the new port
 becomes active. To build it yourself you'll need XCode4 then. XCode4 comes
 with the 10.7 SDK.

+1 the wiki build instructions should also be updated.

Do we also need to update information on the download page (e.g. that we
only support OSX 10.7 and up) ?

rgds
jan I.


Herbert


 -
 To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
 For additional commands, e-mail: dev-h...@openoffice.apache.org




Re: AOO is now also available as GIT (READ-ONLY).

2013-12-19 Thread Herbert Duerr

On 19.12.2013 17:38, jan i wrote:

On 19 December 2013 16:28, Herbert Duerr h...@apache.org wrote:

On 19.12.2013 15:37, jan i wrote:


We have anyhow to consider it the GIT branch is a practial solution, I
just
did a branch switch and encountered 2 problem:
1. The switch alone to 20 minutes



A svn switch between two branches takes about the same time, doesn't it?
Window's file access is unfortunately quite slow. On Linux or on Mac the
filesystems are no bottleneck so checkouts don't take long.



The difference (of course I might use git wrongly), is the in svn, the
branch is stored in its directory. With git I same disk space (trunk and
branches are in 1 directory) but it needs to write a lot of files.


Ah, so you were not checking out trunk or branches individually but you 
conceptually had all of http://svn.apache.org/repos/asf/openoffice/ 
which contains the website, all branches and trunk. With some trickery 
you didn't actually check out the directories you were not interested 
in. So when you built the branches in their individual directories they 
couldn't interfere.


If you like this setup with the individual directories per branch then I 
suggest to first do a reference clone of our git mirror:

git clone --bare git://git.apache.org/openoffice.git aoo.git
and then checkout each branch you are interested in, e.g.
	git clone --reference aoo.git --branch capstone2013 
git://git.apache.org/openoffice.git mycapstone2013


You can then work on these individual git branches as you were doing 
with the svn branches.



He, I think it boils down to what I wrote in 1). Git standard keeps
trunk/branches in one directory, and for small projects its very fast to
switch.


Try out the setup I suggested above. There is no need to switch (and 
rebuild from scratch) if you like it that way.



I have managed with a trix to have a git trunk in one directory and a git
branch in another directory. However when I read the git manual that is not
how its supposed to work.


If your preferred workflow uses different directories then THAT's how 
its supposed to work ;-)



But maybe you are using a different setup for your svn repos? Do you have
one checkout per branch you are interested in? You can do the same with git
of course and with git-clone's --reference option you can even share the
meat of the different repositories, which is something very nice that I
missed in Mercurial or now in Subversion.



I have the setup:
  /x1/aoo/trunk /x1/aoo/branches
so one big aoo directory with everything (I have asked svn to ignore the
branches am not interested in).


As mentioned above you can have its equivalent in git. Without the 
ignore-trickery you had to use ;-)



But please bear with me, I am no git expert, just read the manual and try
to follow it.


We'll help each other on the way to a perfect workflow.

Herbert

-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



Re: Reopening the Vietnamese Forum

2013-12-19 Thread Hagar Delest

Le 18/12/2013 13:05, Drew Jensen a écrit :

FWIW I would close and remove the site. The real problem is lack of
interest by users. The VI user activity never really was adequate to
support a forum IMO.


Basically, I agree.
But since someone thought that there would be worth a try, I'd like to wait a 
bit and see he manages to revive it.
Even if it doesn't seem to be the case for the moment, the forum has reopened 
rather recently so let's wait and see.
As long as spam is under control, there is no big issue.

Hagar

-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



Re: OpenOffice passwords: What are we telling users on the forums?

2013-12-19 Thread Hagar Delest

Top posting.
For the record, a query about password topics: 
https://forum.openoffice.org/en/forum/search.php?keywords=passwordterms=allauthor=sc=1sf=allsr=topicssk=tsd=dst=0ch=300t=0submit=Search

Indeed, we are several users advising to switch from AOO password to 3rd party 
password (pwd archive for example).
We have seen some cases of data loss. See:
- https://forum.openoffice.org/en/forum/viewtopic.php?f=7t=65258
- https://forum.openoffice.org/en/forum/viewtopic.php?f=9t=55698
- https://forum.openoffice.org/en/forum/viewtopic.php?f=7t=54816 (not much 
information, I agree)
- https://forum.openoffice.org/en/forum/viewtopic.php?f=9t=10401 (with a fix)

Other issues:
- https://forum.openoffice.org/en/forum/viewtopic.php?f=6t=49537
- https://forum.openoffice.org/en/forum/viewtopic.php?f=9t=64901
- https://forum.openoffice.org/en/forum/viewtopic.php?f=9t=63279
- https://forum.openoffice.org/en/forum/viewtopic.php?f=9t=54865

I remember some other topics but couldn't find them right now. I confirm that 
we are talking about the file encryption and that in most cases, there is no 
question on the password itself

Hagar


Le 17/12/2013 21:44, Rob Weir a écrit :


I noticed this note, which I thought was odd:

http://listarchives.libreoffice.org/global/users/msg35699.html

I'm hoping this is just a confusion, but we do need to be careful to
avoid confusion in this area, since it can cause users to panic.

The facts, as I understand them:

There are two features in OpenOffice (and in LibreOffice and Microsoft
Office) that users refer to when they talk about passwords:

1) Password protected encrypted documents

2) Password protected sections, cells, read-only files, etc.

An encrypted document is as good as your password.  We use good, high
quality encryption in ODF documents by default.  And we use MS Office
compatible encryption, which is also good, with Office files.

But in practice most users have far weaker passwords than they should.
  The context of a password protected file is much more vulnerable than
a website password.  A typical website will allow you to attempt a log
in 3 or 5 times before locking you out for an hour or more.  But
someone who has your encrypted document can attempt to guess the
password without any such restriction.  They can run sophisticated
programs, standalone password crackers, with GPU hardware acceleration
to attempt billions of passwords.  So a casual password of 6
alphanumeric characters will be quickly broken. So given the context
users should be using longer, more complex passwords.  Of course, that
makes it more likely that they will forget the password and show up on
the forums when they forget.  However you look at it, document-based
passwords are a 1985 solution to a problem that is better solved today
in other ways.

As for the protected sections, we should all know that these are
honor system protection mechanisms, essentially child safety locks,
and offer no real cryptographic protection.  This is true in MS Office
is well.  The feature is there to help the user define sections that
they don't want accidentally deleted, but the password protection can
be trivially defeated in 30 seconds with a text editor and a copy of
unzip.  This is not a flaw in OpenOffice.  This is not a bug.  This is
how the feature was designed and has been used in Microsoft Office and
even 1-2-3 before then.

Hopefully we're telling users something that is consistent with what I
outlined above.  Of course, it is quite possible that many users will
not understand this and all they hear is My password can be broken so
OpenOffice is bad.

Regards,

-Rob

-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



Re: openoffice build issues

2013-12-19 Thread FR web forum
in python module i am getting multiple errors. Can you please help me out.

I'm not a developper but Python 2.7.5 is very old.
Maybe it can be upgrade to 3.3 like LibreOffice?
http://python3porting.com/

-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



RE: EXTERNAL: Re: Building comphelper

2013-12-19 Thread Steele, Raymond
Herbert,

Raymond and I are in the process of rebuilding OpenOffice with the debug flags, 
but have run into some errors that didn't occur the first time through in the 
build. The current error has caused me quite a bit of problems.

We are getting a link error, unrecognized symbol in 
/svx/source/engine3d/e3dundo.cxx. However the symbol that it is looking for is 
not present inside e3dundo.cxx. The symbol is 
ParagraphDataParagraphData::operator=(const ParagraphData) [sic] which we 
have located in /editeng/inc/editeng/paragraphdata.hxx and the implementation 
appears to in  /editeng/source/outliner/paralist.cxx.

There are some very odd things going on here.

1)  paragraphdata.hxx is not included in paralist.cxx even though this is 
where the prototype is declared.
2)  e3dundo.cxx doesn't seem to use the ParagraphData class at all.
3)  e3dundo.cxx doesn't seem to include the libraries where these classes 
are used.

I want to note that we first started the debug build using multiple processors 
(-P8) and perhaps that has caused problems in the ordering in which the 
libraries were compiled and linked. It is conceivable that this problem may be 
resolved with a complete system clean and rebuild in linear mode, but before we 
committed to another day of building the product, we wanted to check with you 
to see if this problem has occurred before, is it recognized or if there is a 
solution that you could recommend?

We'd appreciate any help or insight you could provide. 

David Meffe

P.S. I will be on holiday for the next two weeks and I believe Raymond will be 
as well so we may be out of communication.

-Original Message-
From: Herbert Duerr [mailto:h...@apache.org] 
Sent: Wednesday, December 18, 2013 12:11 AM
To: dev@openoffice.apache.org
Cc: Steele, Raymond; Meffe, David K
Subject: Re: EXTERNAL: Re: Building comphelper

Hello David,
Hello Raymond,

On 17.12.2013 23:31, Steele, Raymond wrote:
 At long last Raymond and I have successfully gotten through the entire 
 build script without a fault. The executable seemed to create just 
 fine and we were able to launch soffice. We got the splash screen, 
 then the executable started and the options window appeared and 
 everything looked to be in order,

That's a great success!

 but when we selected one of the document types to open, the executable 
 crashed with the memory error:

 ./soffice[121]: wait: 6980: Memory fault

I'm afraid we need a stack trace of this crash to understand what's going on. 
When the stack trace shows which libraries are involved then these libraries 
should be recompiled with debug infos (i.e. created with the debug=1 option 
for make or for build). Then replace the libs in your installation with the 
debug libs and get the stack trace of this crash.

 I don't think Raymond or I was expecting everything to work right on first 
 launch and we were happy to see the splash screen, the executable start and 
 the option menu appear. However, have you seen this type of memory error 
 before or has it been seen by others? Do you have an idea what might cause 
 this?

If you single stepped through the huge amounts of code that is needed to get to 
there you'd see there is more than plenty of opportunity for something to go 
wrong ;-)

 My fear is that we have made too many modifications to the code in order to 
 get the system to build correctly. We have removed, in small select areas, 
 const as a qualifier to certain parameters to member functions because the 
 system would not compile with these parameters declared as a const. We can 
 provide a list of these locations that we have edited, I believe it is about 
 3 or 4 files that might have been edited in this way.

The changes in the few source files you mentioned in your other mail are 
probably uncritical. And e.g. Calc isn't changed at all, so if it doesn't work 
either then I suspect the problem in a deeper layer.

 If you have any ideas, it would greatly help us out. Thanks.

The stack trace of the crash when debug-enabled libraries were used would be 
very helpful

Herbert

-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



Re: switch trunk from Mac 32bit to 64bit

2013-12-19 Thread Jürgen Schmidt
On 12/19/13 5:58 PM, jan i wrote:
 On 19 December 2013 17:29, Herbert Duerr hdu_...@alice.de wrote:
 
 The new Mac port looks quite good. I uploaded a current version to my page
 [1]. Jürgen already mentioned it will only work for OSX 10.7 and up. It is
 based on todays trunk, which already contains a lot of fixes and
 enhancements compared to our latest release. For details you can have a
 look at our progress tracking page [2].

 [1] http://people.apache.org/~hdu/
 [2] http://people.apache.org/~hdu/izlist9.htm

 In the early days of next year I plan to update our trunk so the new port
 becomes active. To build it yourself you'll need XCode4 then. XCode4 comes
 with the 10.7 SDK.

 +1 the wiki build instructions should also be updated.
 
 Do we also need to update information on the download page (e.g. that we
 only support OSX 10.7 and up) ?

yes, we should make this clear on the download page and in the release
notes. But the new version itself will not run on versions lower 10.7
and have a version check, see
http://people.apache.org/~jsc/screenshots/MacOS_version_check_AOO4.1.png

And if somebody tries to start it, the error box with the related
version info pops up, see again the screen shot.

Juergen

 
 rgds
 jan I.
 
 
 Herbert


 -
 To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
 For additional commands, e-mail: dev-h...@openoffice.apache.org


 


-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org



Re: How to distinguish whether a 32 or 64 bit AOO is installed on systems that may host both bitnesses ? (Re: switch trunk from Mac 32bit to 64bit

2013-12-19 Thread Jürgen Schmidt
On 12/19/13 5:44 PM, Rony G. Flatscher (Apache) wrote:
 Congratulations for allowing AOO be compiled in 64 bit for the MacOSX!
 
 Is or will be there a possiblity to find out which version of AOO is 
 installed, the 32 or the 64 bit
 version?
 
 There is at least one (mine ;) ) third party extension that may try to add an 
 extension to AOO. As
 the app can only be installed in either 32 or 64 bit, the extension support 
 for AOO should only be
 installed, if bitnesses between both match. So on operating systems where AOO 
 can be installed
 either as 32 or as 64 bit it would be necessary to figure out (in a platform 
 independent way) what
 bitness the installed AOO carries.
 
 As unoinfo seems to not carry the bitness information, is there another 
 (portable, platform
 independent) way to find out what bitness the installed AOO has?

no I think today it is not so easy to detect but it sounds like a useful
enhancement.

Juergen

 
 TIA,
 
 ---rony
 
 P.S.: Expecting that after all PCs migrated from 32 to 64 bit, that then the 
 migration to 128 will
 start, such that this remains an interesting piece of information for the 
 times to come.
 
 
 
 On 19.12.2013 17:29, Herbert Duerr wrote:
 The new Mac port looks quite good. I uploaded a current version to my page 
 [1]. Jürgen already
 mentioned it will only work for OSX 10.7 and up. It is based on todays 
 trunk, which already
 contains a lot of fixes and enhancements compared to our latest release. For 
 details you can have
 a look at our progress tracking page [2].

 [1] http://people.apache.org/~hdu/
 [2] http://people.apache.org/~hdu/izlist9.htm

 In the early days of next year I plan to update our trunk so the new port 
 becomes active. To build
 it yourself you'll need XCode4 then. XCode4 comes with the 10.7 SDK.

 
 -
 To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
 For additional commands, e-mail: dev-h...@openoffice.apache.org
 


-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org