date:20121022

On 10/20/12 10:06 AM, jan iversen wrote:
 You are quite right, I might not be the typical volunteer, and it is very
 important to find a hook where you can start, I had the luck that juergen
 and andrea gave me a starting point.
 
 Your list is quite ok, just lets call it something neutral, like help to
 get started.

I think part of the problem is that we have sometimes too much
information for the same things and it is not well structured. For
example the building guides, we have several pages describing how to
build AOO. Some of these pages are old and out-dated or incomplete. And
we have the new guide which is not the first hit when you search for it
(at least I find always the old ones first). I think cleanup of the
available info would be also of great help.

And we should make clear that asking questions is ok and wanted, it
gives not really stupid questions. We can try to use any question to
improve things, we can ask ourselves why was this question asked and why
wasn't the potentially available information not found...

Juergen


 
 jan
 
 On 20 October 2012 00:24, Rob Weir robw...@apache.org wrote:
 
 On Fri, Oct 19, 2012 at 6:16 PM, jan iversen jancasacon...@gmail.com
 wrote:
 I think it is important to remember, that a volunteer is not signing up
 for
 anything.

 A volunteer, in my view, is a person who wants to help with his/hers
 skillset...so if we start saying you have to pass level x before
 continuing
 we have already lost (At least I can relate that to myself)


 That might be true for you.  But I can tell you from experience that
 we've had volunteer after volunteer who have posted a note to this
 list, said they wanted to help, stuck around for a few days, and then
 were never heard of again.  They never found a hook that they could
 attach themselves to.  They never figured out how to get started.  The
 couldn't find where to get started.  The lack of accomplishment and
 progress leads to frustration, and then they are gone.

 Maybe we can find some way of expressing this without offering too
 much offense ?

 -Rob

 I have been in this business since 1975, and I have never made it through
 any of all these master classes and other exams. I am just one of the
 guys who get things done, like in the early days before tcp/ip.

 What I am trying to say is, let´s help people work with usthat´s what
 it´s all about, if we can help people to easier help us, then we have a
 win-win situation.

 And in respect of introducing myself, which I forgot please read this
 resume:
 http://wiki.openoffice.org/wiki/User:JanIversen

 jan.

 Jan.



 On 19 October 2012 23:08, Rob Weir rabas...@gmail.com wrote:

 On Oct 19, 2012, at 4:45 PM, Kay Schenk kay.sch...@gmail.com wrote:



 On 10/19/2012 01:07 PM, Rob Weir wrote:
 On Fri, Oct 19, 2012 at 12:17 PM, Rob Weir robw...@apache.org
 wrote:
 I am thinking about what new project volunteers need to get started.
 Obviously there are area-specific things.  For example, developers
 need to know how to download and build.  Translation volunteers need
 to understand Pootle, etc.  But there are also some basic things
 that
 all volunteers should probably do.

 Although we have all of this information (or at least most of it) on
 the website or wikis or mailing list archives, it is scattered all
 over the place.  I think it would be good if we could collect this
 information (or at least links to this information) into one place
 and
 put a linear order behind it, a step of specific steps we want new
 volunteers to take.

 Now, I can hear the objections already -- you can't tell volunteers
 what to do.  That is why they are volunteers.  You can't regiment
 them, etc.  This is true.  But at the scale we need to operate at --
 I'm aiming to attract dozens of new volunteers on the project by the
 end of the year -- we need some structure.  So what can we do to
 make
 their first 2 weeks in the project easier for them, and easier for
 us?

 One idea:  Think of the new volunteer startup tasks in terms of
 stages or levels, a defined set of reading and other activities
 that leads them to acquire basic skills in our community.

 For example:

 To make it more concrete, this is what Level 1 might look like:

 http://incubator.apache.org/openofficeorg/orientation/level-1.html

 -Rob

 This is very good! I esp like the last part about providing a way for
 volunteers to sign up if you will. This will be a nice touch.

 I'm also wondering if there's some way to tie this in to our current
 Help Wanted page:

 https://cwiki.apache.org/confluence/display/OOOUSERS/Help+Wanted

 Yes, It is worth looking at the new volunteer view of things, from end
 to
 end.

 My current thinking is this: as we scale the number of volunteers
 we'll soon want a better way to track items like these. Maybe putting
 them into BZ would work?  Introduce a new field to record difficulty
 in BZ and filters to list unassigned easy issues?


 Maybe someone has some ideas?


 Level 1 tasks:

 1) Read

Another graphic needs updating, formerly Re: Another logo needs updating: Get it here!

2012-10-22 Thread Kevin Grignon

On Monday, October 22, 2012, Hagar Delest wrote:

 Le 21/10/2012 21:18, RGB ES a écrit :

 CC to Hagar, because he proposed the download logo used on the forums:

 http://forum.openoffice.org/**en/forum/viewtopic.php?p=**238128#p238128http://forum.openoffice.org/en/forum/viewtopic.php?p=238128#p238128

 See below for more details

 [...]


 Hy, Hagar. There is a discussion on the MLs about a new download button
 (without the incubator part) and the download logo you built for the forums
 seems a really good starting point. Do you still have the source file?


 Hi Ricardo,

 The one on the forum is the source file. I just used the different logos
 available in the different areas of the site and assembled them to make the
 pic.
 So there is no SVG version, just that one in .png

 Hagar



Hey all,

We should be clear in the distinction between a logo and a graphic image
that includes branding elements or the actual logo.

We need not design new or different logos for different contexts, rather we
should apply our branding and logo guidelines to such graphics.

Re: Apache OO General Questions

2012-10-22 Thread Michael Meeks


On Thu, 2012-10-18 at 16:05 -0500, Alexandro Colorado wrote:
 Recently i reported a 7yr old bug on memory management, i am sure LibO
 also inherit that bug and they haven't fix it mainly because of the
 low level technical knowledge to devote to the performance project.

I seem to remember doing some work on performance somewhen, perhaps in
glibc, gcc, binutils  misc. plumbing etc. in the past - seemingly it
has even had an effect, linking is nowadays only a small proportion of
startup. It can be nice to have a low as well as a high level view.

IIRC the OO.o performance project where we mentored and tried to advise
others on improving things was lead by myself and Mathias Heutsch - but
my memory management is often flaky ;-)

We have certainly inherited lots of bugs though; perhaps
you have a bug number ?

All the best,

Michael.

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

Re: [DISCUSS] defining roles for management, coordination, work items...

2012-10-22 Thread Kevin Grignon

On Friday, October 19, 2012, Kay Schenk wrote:

On 10/17/2012 10:35 PM, Kevin Grignon wrote:

KG01 - see comments online

On Oct 18, 2012, at 5:24 AM, Kay Schenk kay.sch...@gmail.com wrote:

On 10/16/2012 03:40 PM, Rob Weir wrote:

On Tue, Oct 16, 2012 at 6:06 PM, Kay Schenk kay.sch...@gmail.com
wrote:

On 10/16/2012 02:47 PM, Rob Weir wrote:

On Tue, Oct 16, 2012 at 5:29 PM, Kay Schenk kay.sch...@gmail.com
wrote:

[top posting -- old discussion/business]

I just created a little wiki schematic page based on this discussion
at:

https://cwiki.apache.org/**confluence/display/OOOUSERS/**
Project+Management+Roleshttps://cwiki.apache.org/confluence/display/OOOUSERS/Project+Management+Roles

which will make it easy for us to add roles, people to roles, etc.
[The second column, intended for actual names, is blank so far.]

I'm not really sure what problem we're solving here.

Better defining activities that we do/need to do.
Hopefully soliciting/inviting individuals to take ownership of some of
these
activities.

But we can have a full list of roles on the wiki, but not define a
single task... But maybe this is the first step.

yes...and an oversight in my initial enthusiasm. Obviously, it is
difficult to determine if any list is sufficient unless the roles are
defined. :/

One way of making this scale and be more self-maintaining is something
like:

1) Work with Infra to create a new issue tracking DB for the project.
Maybe use JIRA rather than BZ. This new DB would be for tracking
tasks, not for tracking bugs. (Why a separate DB? So we don't drive
QA team crazy.)

KG01 - Rob, I like the idea of a new system that supports scope items in
addition to work items. We could benefit from more PPM tooling.

Kevin--

Maybe you could raise this as an additional issue or a separate discussion
item, and explain to the community what you mean by this. I don't know how
many folks involved here know what PPM tooling is.

KG02 - PPM = project and portfolio management. The PPM process include
objects for scope and work. Some change management systems that manage work
items realize the PPM conceptual model.

2) In the database we put in all the tasks that we know needs to get
done, from running a RAT scan before a release, to verifying the new
Norwegian translation, to UX exploration tasks, etc. There are
probably 100+ tasks that we could think of, more than would be fun to
track on the wiki.

KG01 - In fact, I'm already in the midst of populating a ux opportunity
backlog as BZ does not serve my business needs.

3) Project members can comment on issues and take ownership of them.

4) If a project member is focused deeply on a specific issue, then
they can set themselves as the default owner for that issue
category.

In this way, roles are defined implicitly by what tasks you claim and
what categories you set yourself as the default owner.

KG01 - agreed, less role oriented and more work item oriented.

work items pretty much determine role I would think except in some
specialized cases at Apache.

KG02 - Not necessarily, many work items are boundary objects that apply, or
could be performed by a variety of role. My previous comment was to
encourage us to focus on work and tasks, and focus less on role.

The nice thing about this approach is it helps with the communication
challenges of a large project. None of us can read everything on
every mailing list. But we can all read the JIRA notifications that
come directly to us as an issue owner.

Well this part is certainly true. It's difficult to overlook what you
signed up for, when you get notifications.

Another nice thing about this approach is it lets us set up a backlog
of things that are nice to do someday but where we have no
volunteer. For example, spell checking the website, or updating the
Indonesian translation. We don't even have a person in the role of an
Indonesian Translator today. But if the tasks are defined in JIRA
then we can the unassigned tasks as a way to recruit new volunteers.
It makes it easier to see where we need help.

KG01 - backlogs also allow is to create candidate release and iteration
plans

I think the JIRA/BZ approach definitely has merit in the long run.
For now, I think it would be advantageous to just flesh out the page a bit
more to supply definitions and see what we're missing in terms of typical
actions/roles that we take on, or should be taken on.

And, although people could indeed enroll for tasks right on the page,
I'm not explicitly suggesting that as some categories, for example,
developer, would have MANY entries. However, I do think it is valuable for
site visitors to at least identify mailing list moderators, and maybe BZ,
wiki, etc. admins.

For example, if person X is assigned role Y, I assume we don't want to
encourage people to contact person X directly for questions or
assistance. We should do

Re: Dmake binaries ?

2012-10-22 Thread Andre Fischer


On 21.10.2012 02:38, Pedro Giffuni wrote:

Hi;

Apparently the Dmake binaries that I made available on the Dmake site don't 
work.

I think it would save some time, especially for Windows users to have a binary
installable package. Perhaps some would like to provide one?

best regards,

Pedro.


I don't see the problem.   When you build AOO it takes roughly 95 
minutes on my system.   Building dmake takes about one minute (just 
timed bootstrap with 1m7.555s) and requires tools that you need anyway.  
Moreover, you can copy dmake.exe to /usr/bin and don't have to build it 
again.


I admit that the configure process could be made easier regarding dmake 
but then again the configure process could be made easier regarding 
anything else, too.


Andre

Re: Directory main/swext/mediawiki

2012-10-22 Thread Herbert Duerr


On 21.10.2012 18:16, jan iversen wrote:

Now I am the one that is confused (again).

Is the directory mediaWiki active or not ?


The extension in that directory is built only if you used the
  --enable-wiki-publisher
configure option.

Herbert

Re: Directory main/swext/mediawiki

Thanks, I thought --all meant just compiling it all :-)

I tried to do man build to get a list of all switches, which of course
did not work.

Is there a place (build config file or so) to look for all the switches.

JanI

On 22 October 2012 10:36, Herbert Duerr h...@apache.org wrote:

 On 21.10.2012 18:16, jan iversen wrote:

 Now I am the one that is confused (again).

 Is the directory mediaWiki active or not ?


 The extension in that directory is built only if you used the
   --enable-wiki-publisher
 configure option.

 Herbert

Re: OpenOffice Developer Room (devroom) at FOSDEM

On 10/21/12 11:06 PM, Andrea Pescetti wrote:
 On 16/10/2012 Andrea Pescetti wrote:
 I've asked the organizers to inform us about their decisions. I expect
 to receive news in the next days; I'll keep you informed when I do.
 
 I received the answer just a few minutes ago: organizers decided to give
 OpenOffice its own devroom, which will be available all day Saturday (2
 Feb 2013). Now we will have to organize a call for talks and so on, but
 this is not urgent. As agreed with the FOSDEM organizers, I'm reposting
 their answer in public, see below.

ok, a little bit surprising but a clear signal.

We should try to find out if other Apache projects will be there as well
and if we can share a booth with them. Having a booth is a good
opportunity to talk with people and to answer questions.

Juergen



 
 Regards,
   Andrea.
 
---
 
 Dear Andrea,
 
 Thank you for considering a merged devroom with LibreOffice. After some
 internal discussion, we have come to the conclusion that perhaps the
 best way to foster collaboration and cross-polination between AOO and LO
 is not to force a merged devroom, but rather to give each projects its
 own room on a different day. We hope both communities will use this
 opportunity in the best way possible.
 
 It is hence our pleasure to inform you that your request for a devroom
 at FOSDEM 2013 has been accepted.
 
 
 We offer you the following:
 - a room on Saturday
 - with video projector and best effort wifi
 - in building AW (76 seats)
 
 We must add that as in previous years, visitor interest will outnumber
 room capacity for most devrooms. We can not offer you larger rooms --
 instead we will provide you with signalization to indicate that the room
 is full.
 
 If you prefer to decline this offer, please inform us as soon as possible.
 
 
 By accepting this offer, you accept to be responsible for this devroom.
 This means to spread a call for papers in your community, compose a
 schedule, and provide us with that information. It also includes on-site
 responsibilities such as making sure the capacity limit is respected and
 that the devroom is clean by the end of the day.
 
 
 Next: call for talks
 
 We invite you to send an open call for talks to relevant mailinglists.
 Please include a copy to fos...@lists.fosdem.org
 We will collect all the calls on our website.
 
 Regarding content and format, you are free to schedule your devroom as
 you wish.
 Do allow us to repeat the obvious: FOSDEM is a FOSS community event, by
 and for the community, and the content and topics _must_ reflect that
 (arguably, that is pretty obvious).
 
 
 We offer you a name-devr...@fosdem.org mailman mailinglist, in case you
 want to have a mailinglist 'on neutral grounds'. This can be especially
 interesting for joint devrooms.
 Contact us to take on this offer.
 
 
 Deadlines
 -
 The hard deadline for sending us the full schedule is *2013-01-10*.
 
 For your own call for talks, we advise mid-december as deadline, with
 notifications to the speakers around christmas. (The sooner, the more
 time they have to book travel/accomodation).
 
 
 
 Lastly, we wish to clarify that there are no reimbursements of travel or
 hotel costs for devroom organizers/speakers.
 
 
 And, of course, if you have any questions, please do contact us at
 devro...@fosdem.org
 
 
 Looking forward to your participation,
 On behalf of the FOSDEM devroom team,
 Tias

Re: Directory main/swext/mediawiki

Sorry, You wrote configure option...please forget my mail.

janI

On 22 October 2012 10:49, jan iversen jancasacon...@gmail.com wrote:

 Thanks, I thought --all meant just compiling it all :-)

 I tried to do man build to get a list of all switches, which of course
 did not work.

 Is there a place (build config file or so) to look for all the switches.

 JanI


 On 22 October 2012 10:36, Herbert Duerr h...@apache.org wrote:

 On 21.10.2012 18:16, jan iversen wrote:

 Now I am the one that is confused (again).

 Is the directory mediaWiki active or not ?


 The extension in that directory is built only if you used the
   --enable-wiki-publisher
 configure option.

 Herbert

Re: Wiki robots.txt

2012-10-22 Thread Herbert Duerr


On 21.10.2012 15:13, imacat wrote:

 I found the following rule in the robots.txt of our wiki:

User-Agent: *
Disallow: /

 Does any know if there is any special reason why it is set so?  Does
any have any reason to keep it?  I'm thinking of removing this rule.


+1, blocking all search robots makes no sense.
Google etc. are also much more successful in finding relevant results to 
non-trivial searches. The wiki-builtin search had many problems [1], 
many of which are fixed in the meantime though.


[1] http://www.mediawiki.org/wiki/Search_issues

Herbert

Re: Wiki robots.txt

2012-10-22 Thread Rory O'Farrell

On Mon, 22 Oct 2012 11:29:18 +0200
Herbert Duerr h...@apache.org wrote:

 On 21.10.2012 15:13, imacat wrote:
   I found the following rule in the robots.txt of our wiki:
 
  User-Agent: *
  Disallow: /
 
   Does any know if there is any special reason why it is set so?  Does
  any have any reason to keep it?  I'm thinking of removing this rule.
 
 +1, blocking all search robots makes no sense.
 Google etc. are also much more successful in finding relevant results to 
 non-trivial searches. The wiki-builtin search had many problems [1], 
 many of which are fixed in the meantime though.
 
 [1] http://www.mediawiki.org/wiki/Search_issues
 
 Herbert
 

+1

-- 
Rory O'Farrell ofarr...@iol.ie

Re: Directory main/swext/mediawiki

On 10/21/12 6:16 PM, jan iversen wrote:
 Now I am the one that is confused (again).
 
 Is the directory mediaWiki active or not ?
 
 SPI is an interface if I understand it right, and that is hopefully not
 hidden in the mediawiki directory ?

Herbert answered already, sorry I didn't wanted to confuse you. It was
more related to Alexandros reply which I of course didn't understand ;-)

Juergen

 
 jan
 
 On 21 October 2012 17:40, Juergen Schmidt jogischm...@gmail.com wrote:
 

 Am Samstag, 20. Oktober 2012 um 23:52 schrieb Alexandro Colorado:
 On Sat, Oct 20, 2012 at 2:50 PM, jan iversen jancasacon...@gmail.com
 wrote:

 I am a bit confused.

 We have a directory named main/swext/mediawiki.

 Is that the sun wiki publisher 1.1 or are there 2 different mediawiki
 export extensions ?

 I ask because I sun wiki publisher 1.1 installed, but if I change
 the XLS
 and rebuilt AOO but it does not seem to have an effect.



 Yes I think it was started on core, and then sent to a separate
 extensions.
 Same thing happened with smarttags IIRC.



 what do you mean here exactly, it doesn't make sense. Smarttags  API (or
 better the SPI) is available in the core and it is now possible to develop
 smarttags either in the core or as extension.

 Juergen



 Either I make a wrong assumption or life is not so simple as I would
 it to
 be :-)

 thanks in advance.
 jan.





 --
 Alexandro Colorado
 PPMC Apache OpenOffice
 http://es.openoffice.org

build --all and clean.

I have made a build --all, after which I use configure to set other options.

Now is there any way I can clean all objects etc. from the previous build.

I thought
build --all --clean
would do it, but build does not recognize --clean

I cannot find it in wiki.

thanks in advance for help.
JanI.

Re: Automatically Generated extension.update.xml

On 10/21/12 11:13 PM, Andrea Pescetti wrote:
 imacat wrote:
 I wonder if it
 is possible to generate the extension update information files
 extension.update.xml automatically on
 http://extensions.openoffice.org, instead of maintaining the update
 information files by the authors manually?
 
 This requires a slightly complex answer since there are differences
 between theory and practice.
 
 In theory, when you upload an extension to extensions.openoffice.org,
 you needn't provide an update XML file: the repository knows about the
 mechanism and automatically generates it. Moreover, OpenOffice knows
 where to ask for the update so, if you use the official repository to
 host your extension, updates will magically work with no need of actions
 from the developer (not even specify an update URL), aside possibly from
 making sure that the version number always increases.
 
 In practice, the process is broken. The point where it breaks is the
 fact that extensions.openoffice.org does not provide the real update
 information but it sends a hardcoded answer instead, saying that no
 updates are available. This was done back in the Oracle times for some
 reasons - probably traffic, but I have no clue here, maybe it was just
 broken.
 
 Fact is, what you ask for is already available in theory and I have seen
 it working for a period before it was disabled in 2010. But it isn't
 available in practice at the moment, and it hasn't been for years.
 
 On the application side (the OpenOffice program) everything should still
 be working, even though I've recently seen an issue where Ariel
 explained we might have some regressions on the OpenOffice side too:
 https://issues.apache.org/ooo/show_bug.cgi?id=121201

mmh, I was not ware that this function was disabled in 2010 and I don't
know the reason for it. But Roberto has explained it quite well and the
idea behind it was exactly that for extensions hosted in our repo the
update information is generated automatically and the extensions
shouldn't add further update information in their description.xml.

We should figure out if the mechanism still works and should enable it.

Ariel reported indeed a problem with his updates but we have to check
this in general because I at least was not able to reproduce all
reported problems. But this is a separate issue and have to fixed
independently.

Juergen

Re: File: readme.xrm

On 10/21/12 2:16 PM, jan iversen wrote:
 There is exactly one file with extension .xrm
 
 main/read_license_oo/docs/readme/readme.xrm
 
 is there a reason (apart from history) for it being in .xrm or could it be
 converted to e.g. .xhp ?
 
 If so we could get rid of one more conversion tool (read: does not need to
 be converted to new code).
 

I am not sure if xhp would be a good option here. But we can probably
switch to something else. Maybe a common readme file that gets extended
with platform specific portions from other files. When I remember it
correctly the xrm files contains the content for the readme file and
depending on the platform different content is extracted from this file.

Juergen

Re: File: readme.xrm

As far as I can see on the usage are your assumption correct, and there
must be other ways to make different readme text platform dependent.

Would it not be ok, to have one readme for all platforms, and in the text
mention the specics ?

janI

On 22 October 2012 13:34, Jürgen Schmidt jogischm...@gmail.com wrote:

 On 10/21/12 2:16 PM, jan iversen wrote:
  There is exactly one file with extension .xrm
 
  main/read_license_oo/docs/readme/readme.xrm
 
  is there a reason (apart from history) for it being in .xrm or could it
 be
  converted to e.g. .xhp ?
 
  If so we could get rid of one more conversion tool (read: does not need
 to
  be converted to new code).
 

 I am not sure if xhp would be a good option here. But we can probably
 switch to something else. Maybe a common readme file that gets extended
 with platform specific portions from other files. When I remember it
 correctly the xrm files contains the content for the readme file and
 depending on the platform different content is extracted from this file.

 Juergen

[RELEASE]: propose a new developer snapshot based on revision 1400796

Hi,

we still have problems with our public build bots, a problem that we
should address immediately. Nevertheless would I like to propose that we
start a developer snapshot build based on revision 1400796.

I am currently build MacOS and Windows and I hope that the revision will
build on both platform without error. And that it will build on Linux as
well.

I added 5 further languages (gd da pl ko nb) to make them available for
our new volunteers to review and check the result of their translations
or to simply review what's already there.

In case of an error we have to adapt the revision accordingly.

@Ariel, are you able to build the Linux version? In case you are not
able for whatever reason please let us know and we have to find an
alternative. I really hope that we can use build bots in the near future
for providing the snapshot builds.

Regards
Juergen

[ApacheCon] Where you booking the room?

2012-10-22 Thread Raphael Bircher

Hi all

Question: Where you booking your Room. In wich hotel you stay during the
ApacheCon?

Greetings Raphael
-- 
My private Homepage: http://www.raphaelbircher.ch/

Re: [RELEASE]: propose a new developer snapshot based on revision 1400796

On Mon, Oct 22, 2012 at 01:57:31PM +0200, Jürgen Schmidt wrote:
 Hi,
 
 we still have problems with our public build bots, a problem that we
 should address immediately. Nevertheless would I like to propose that we
 start a developer snapshot build based on revision 1400796.
 
 I am currently build MacOS and Windows and I hope that the revision will
 build on both platform without error. And that it will build on Linux as
 well.
 
 I added 5 further languages (gd da pl ko nb) to make them available for
 our new volunteers to review and check the result of their translations
 or to simply review what's already there.
 
 In case of an error we have to adapt the revision accordingly.
 
 @Ariel, are you able to build the Linux version? 

Yes, I can.
Did you plan real Developer Snapshots with no system integration?
did you plan only en-US full install sets + language packs?

 In case you are not able for whatever reason please let us know and we
 have to find an alternative. I really hope that we can use build bots
 in the near future for providing the snapshot builds.

With the current Linux build bots is impossible to build
releases/Dev.Snapshots, they have very high system requirements. I guess
I should open a JIRA ticket...


Regards
-- 
Ariel Constenla-Haile
La Plata, Argentina


pgpath90uhCxU.pgp
Description: PGP signature

Re: [RELEASE]: propose a new developer snapshot based on revision 1400796

On 10/22/12 1:57 PM, Jürgen Schmidt wrote:
 Hi,
 
 we still have problems with our public build bots, a problem that we
 should address immediately. Nevertheless would I like to propose that we
 start a developer snapshot build based on revision 1400796.
 
 I am currently build MacOS and Windows and I hope that the revision will
 build on both platform without error. And that it will build on Linux as
 well.
 
 I added 5 further languages (gd da pl ko nb) to make them available for
 our new volunteers to review and check the result of their translations
 or to simply review what's already there.
 
 In case of an error we have to adapt the revision accordingly.

ok, I detected the first problem. With the new updated tomcat version
the servlet-api.jar is not delivered correctly and the build breaks
under MacOS and Windows.

I start a new build with revision 1400866 and will let you know when it
builds successful and if we will use it for the dev build.

Juergen

 
 @Ariel, are you able to build the Linux version? In case you are not
 able for whatever reason please let us know and we have to find an
 alternative. I really hope that we can use build bots in the near future
 for providing the snapshot builds.
 
 Regards
 Juergen

Re: Apache OO General Questions

2012-10-22 Thread Michael Meeks

Hi Alexandro,

On Mon, 2012-10-22 at 03:17 -0500, Alexandro Colorado wrote:
 On Sat, Oct 20, 2012 at 2:39 AM, Michael Meeks michael.me...@suse.comwrote:
  On Thu, 2012-10-18 at 16:05 -0500, Alexandro Colorado wrote:
   Recently i reported a 7yr old bug on memory management, i am sure LibO
   also inherit that bug and they haven't fix it mainly because of the
   low level technical knowledge to devote to the performance project.
...
  We have certainly inherited lots of bugs though; perhaps
  you have a bug number ?
 
 Headless memory leak #105191

I was under the impression from your mail, that this was fixed in
Apache OpenOffice but that doesn't seem to be reflected in the bug
state.

I have a personal bug / TODO to fix what (may well) be the same
underlying issue titled Image caching / management is utterly
shambolic - which has a number of duplicates.

ATB,

Michael.

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

Re: [RELEASE]: propose a new developer snapshot based on revision 1400796

On 10/22/12 3:06 PM, Ariel Constenla-Haile wrote:
 On Mon, Oct 22, 2012 at 01:57:31PM +0200, Jürgen Schmidt wrote:
 Hi,

 we still have problems with our public build bots, a problem that we
 should address immediately. Nevertheless would I like to propose that we
 start a developer snapshot build based on revision 1400796.

 I am currently build MacOS and Windows and I hope that the revision will
 build on both platform without error. And that it will build on Linux as
 well.

 I added 5 further languages (gd da pl ko nb) to make them available for
 our new volunteers to review and check the result of their translations
 or to simply review what's already there.

 In case of an error we have to adapt the revision accordingly.

 @Ariel, are you able to build the Linux version? 
 
 Yes, I can.
 Did you plan real Developer Snapshots with no system integration?
 did you plan only en-US full install sets + language packs?

sorry for being not detailed enough. I plan dev snapshot without system
integration for now.

I can live the en-US full install sets  + language packs. But I am open
here and I would provide what helps our volunteers doing testing and
translation best.

What do you think?

Juergen


PS: before you start please wait until I have at least finished one
platform with my latest fix ;-)

 
 In case you are not able for whatever reason please let us know and we
 have to find an alternative. I really hope that we can use build bots
 in the near future for providing the snapshot builds.
 
 With the current Linux build bots is impossible to build
 releases/Dev.Snapshots, they have very high system requirements. I guess
 I should open a JIRA ticket...
 
 
 Regards

Re: [ApacheCon] Where you booking the room?

2012-10-22 Thread Pedro Giffuni

I will be in Hotel Bär. Not sure if they are full already but
there is information about nearby hotels here:

http://www.apachecon.eu/accommodation/


cheers,

Pedro.




 From: Raphael Bircher r.birc...@gmx.ch
To: ooo-dev@incubator.apache.org 
Sent: Monday, October 22, 2012 7:47 AM
Subject: [ApacheCon] Where you booking the room?
 
Hi all

Question: Where you booking your Room. In wich hotel you stay during the
ApacheCon?

Greetings Raphael
-- 
My private Homepage: http://www.raphaelbircher.ch/

Re: Why a mailing list is superior to any web-based forum

2012-10-22 Thread Shane Curcuru

In a community-driven project, whichever tool is easier for the people 
contributing the work to use is the superior tool.  But this is an [OT] 
thread since the AOO *development* community is already decided that 
ooo-dev@ is it's primary home (although it will shortly move to 
@openoffice.apache.org!).


Luckily, AOO has both a set of mailing lists and a rich set of forums, 
so we get the best of both worlds!


For those interested in following or linking to past mailing threads on 
any ASF mailing list, I also recommend MarkMail like Joe does in this 
thread, which has an awesome search interface:


  http://apache.markmail.org/

- Shane

Re: [RELEASE]: propose a new developer snapshot based on revision 1400796

On Mon, Oct 22, 2012 at 03:22:33PM +0200, Jürgen Schmidt wrote:
  @Ariel, are you able to build the Linux version? 
  
  Yes, I can.
  Did you plan real Developer Snapshots with no system integration?
  did you plan only en-US full install sets + language packs?
 
 sorry for being not detailed enough. I plan dev snapshot without system
 integration for now.
 
 I can live the en-US full install sets  + language packs. But I am open
 here and I would provide what helps our volunteers doing testing and
 translation best.
 
 What do you think?
 
 Juergen
 
 
 PS: before you start please wait until I have at least finished one
 platform with my latest fix ;-)

is this your with-lang switch?

--with-lang=ar cs da de en-GB en-US es fi fr gd gl hu it ja km ko nb nl pl 
pt-BR ru sk sl zh-CN zh-TW


Regards
-- 
Ariel Constenla-Haile
La Plata, Argentina


pgpi5g7Ihx1Rm.pgp
Description: PGP signature

Re: Dmake binaries ?

2012-10-22 Thread Pedro Giffuni

Hi Andre;

 From: Andre Fischer awf@gmail.com
...
Subject: Re: Dmake binaries ?

On 21.10.2012 02:38, Pedro Giffuni wrote:
 Hi;

 Apparently the Dmake binaries that I made available on the Dmake site don't 
 work.

 I think it would save some time, especially for Windows users to have a 
 binary
 installable package. Perhaps some would like to provide one?

 best regards,

 Pedro.

 I don't see the problem.   When you build AOO it takes roughly 95 minutes on 
 my system. 
 Building dmake takes about one minute (just timed bootstrap with 1m7.555s) 
 and requires
 tools that you need anyway.  Moreover, you can copy dmake.exe to /usr/bin and 
 don't have
 to build it again.

My system is much slower but it's not a problem here either.

I was under the impression that you had to copy some configuration files
though (startup/*).

I committed some changes in the dmake project to address recent cygwin.
I am still unsure if it builds with MSVC 2010 but I guess if I provide a binary
release we will never find out ;)..

cheers,

Pedro.

Re: [RELEASE]: propose a new developer snapshot based on revision 1400796

On Mon, Oct 22, 2012 at 9:22 AM, Jürgen Schmidt jogischm...@gmail.com wrote:
 On 10/22/12 3:06 PM, Ariel Constenla-Haile wrote:
 On Mon, Oct 22, 2012 at 01:57:31PM +0200, Jürgen Schmidt wrote:
 Hi,

 we still have problems with our public build bots, a problem that we
 should address immediately. Nevertheless would I like to propose that we
 start a developer snapshot build based on revision 1400796.

 I am currently build MacOS and Windows and I hope that the revision will
 build on both platform without error. And that it will build on Linux as
 well.

 I added 5 further languages (gd da pl ko nb) to make them available for
 our new volunteers to review and check the result of their translations
 or to simply review what's already there.

 In case of an error we have to adapt the revision accordingly.

 @Ariel, are you able to build the Linux version?

 Yes, I can.
 Did you plan real Developer Snapshots with no system integration?
 did you plan only en-US full install sets + language packs?

 sorry for being not detailed enough. I plan dev snapshot without system
 integration for now.

 I can live the en-US full install sets  + language packs. But I am open
 here and I would provide what helps our volunteers doing testing and
 translation best.

 What do you think?


What is the issue with snapshots that have system integration?   Do we
never have bugs with system integration?  Shouldn't we encourage
testing with something that is as close as possible to what we will
release?

Ditto with fill translations:  We are currently discussing some major
changes to the translation workflow.  Are there possible bugs that
could be introduced and not detected if we did only language packs?

-Rob

 Juergen


 PS: before you start please wait until I have at least finished one
 platform with my latest fix ;-)


 In case you are not able for whatever reason please let us know and we
 have to find an alternative. I really hope that we can use build bots
 in the near future for providing the snapshot builds.

 With the current Linux build bots is impossible to build
 releases/Dev.Snapshots, they have very high system requirements. I guess
 I should open a JIRA ticket...


 Regards

Re: [RELEASE]: propose a new developer snapshot based on revision 1400796

On Mon, Oct 22, 2012 at 10:38:47AM -0400, Rob Weir wrote:
 What is the issue with snapshots that have system integration?   Do we
 never have bugs with system integration?  Shouldn't we encourage
 testing with something that is as close as possible to what we will
 release?

The issue is that they replace the current stable installation, unless
you install them in parallel, in which case it make no sense to build
with system integration at all.

Developer Snapshots where named in OpenOffice.org times only the build
that could be installed beside a current stable installation, because
they use a different name: OOo-Dev. In Linux, for example, it installs
on 

/opt/ooo-dev
/opt/ooo-dev3

In OpenOffice.org times, builds with system integration where done only
for the Release Candidates (not even Alphas or Betas, IIRC).


 Ditto with fill translations:  We are currently discussing some major
 changes to the translation workflow.  Are there possible bugs that
 could be introduced and not detected if we did only language packs?

AFAIK, no. Well, the Windows localized installers for the full install
sets could not be tested with a only en-US full install set.


Regards
-- 
Ariel Constenla-Haile
La Plata, Argentina


pgp3w0j2lHVVe.pgp
Description: PGP signature

Re: [RELEASE]: propose a new developer snapshot based on revision 1400796

On 10/22/12 4:24 PM, Ariel Constenla-Haile wrote:
 On Mon, Oct 22, 2012 at 03:22:33PM +0200, Jürgen Schmidt wrote:
 @Ariel, are you able to build the Linux version? 

 Yes, I can.
 Did you plan real Developer Snapshots with no system integration?
 did you plan only en-US full install sets + language packs?

 sorry for being not detailed enough. I plan dev snapshot without system
 integration for now.

 I can live the en-US full install sets  + language packs. But I am open
 here and I would provide what helps our volunteers doing testing and
 translation best.

 What do you think?

 Juergen


 PS: before you start please wait until I have at least finished one
 platform with my latest fix ;-)

1400866 seems to be good, on MacOS I package already the install sets

 
 is this your with-lang switch?
 
 --with-lang=ar cs da de en-GB en-US es fi fr gd gl hu it ja km ko nb nl pl 
 pt-BR ru sk sl zh-CN zh-TW

yes

Juergen

Re: [RELEASE]: propose a new developer snapshot based on revision 1400796

2012-10-22 Thread Pedro Giffuni

Heck!


- Original Message -
 From: Jürgen Schmidt

 
  In case of an error we have to adapt the revision accordingly.
 
 ok, I detected the first problem. With the new updated tomcat version
 the servlet-api.jar is not delivered correctly and the build breaks
 under MacOS and Windows.
 

Sorry, don't know how I got everything building fine in my system without
that change :(.

FWIW, it's over now ... Tomcat will have no further 5.5 releases.

Pedro.

Re: Need new logo for openoffice.apache.org

On 10/20/12 5:34 PM, Alexandro Colorado wrote:
 On Sat, Oct 20, 2012 at 9:28 AM, Rob Weir robw...@apache.org wrote:
 
 See upper left here:  http://openoffice.apache.org

 The Incubating is integrated into the graphic.

 The underlying file is here:  a PNG with transparent background.

 http://incubator.apache.org/openofficeorg/images/300x100_dj_trans.png

 What do we want to do here?

 1) Edit that graphic to remove Incubating?

 2) Use a different graphic?

 Note that the http://www.openoffice.org/ site uses a different form of
 the branding.  Are we intentionally using two different logos here?
 Do we want to continue this?

 
 
 I am confused about having the branding all disjointed, OO is used
 implemented halfway, we use Oracle brand refresh elements just on certain
 parts. The brand refreshed used this:
 - Symbols
 - Logo
 - Icons
   - application
   - modules
   - filetype
 
 There was some big drawback on the monotone look of the filetype, but for
 some unknown reason there was no implementation on the modules either.
 Since, we have been keeping Application logos from 3.1, modules logos from
 2.4 and filetypes of 2.0.
 
 Specifically on the logo, we have 2 versions, one with the gulls and the
 other with the orb. I am not sure if there is a representation.
 
 There is also the issue with the fonts, where the fonts are not open, and
 we currently don't have a specific vector format logo, between using
 Nimbus, Liberation or Bitstreams we still dont have a good functional
 specification.
 
 My proposal is:
 - Get rid of incubating
 - Normalize the font-face of the logo
 - Update the modules logo to the latest refresh
 - Review new filetypes contribution
 
 Rather than changing and implementing a new design, work on the proposals
 to have a consistent look.
 

you seem to have a good overview. Maybe you can create an overview page
in the wiki of what we have now (and where it comes from) and a further
column where we can list what we want use in the next version. This can
help us to get an overview in one single place.

And a brief description what the element is and where it is used!

And of course we only have to change things that don't fit together in
the first place. Everything else can evolve over time.

I have to confess that I am a little bit lost to get a complete overview
of all the branding relevant things.

Juergen

Bringing social capabilities into the core SDK

(Responses please to the ooo-api list)

I wanted to confirm some thinking I've been having recently.  Please
let me know if any of this sounds plausible or even useful.

The problem I'm trying to solve is how to make OpenOffice better
connected with the online tools that we all use today, from DropBox to
Drupal, Twitter to Facebook, Sharepoint to SAP.   We do have a few
extensions in these areas.  That is good.  But I'm wondering if there
are some things we can do in the API to make these kinds of extensions
much easier.

The key observation is that there is a set of protocols and formats
that have emerged as the basic foundation of connecting with these
kinds of apps.  HTTP/XML at the base, but also on top things like RSS,
Atom, OData, OAuth, OpenID, OpenSocial, CMIS, etc.

What if we made available in our SDK, access to a cross-platform
library that provided these capabilities without any extra coding
required?  Where needed, provide integration with native credential
stores, etc.  So the SDK does all the heavy lifting.  The extension
authors can then focus on things at a higher level.

In other words, make socially-connected OpenOffice apps be in the
reach of every app developers, rather than requiring a protocol
expert.

I'd be interested in hearing from extension authors on whether such a
capability would be useful.  And how useful?

Regards,

-Rob

Re: Directory main/swext/mediawiki

Now it generates ok, thanks to that little tip.

I have taken the wiki-publisher.oxt file and put it Documents (that where
my extension manager looks).

When I open the oxt file with extension manager, I get a strange error:

(com.sun.star.deployment.DeploymentException) {{Message=Could not obtain
path to license. Possible error in description.xml,
Context=(com.sun.star.uno.Xinterface) @0}, cause=(any) {void}}

description.xml has tag license_test  that refers to
xlink:href:licene/LICENSE

The error text seems to come from: desktop/deployment.

Can that be because I have just done a build --all and should have done
with --with-lang ?

thanks in advance.
rgds
Jan I

On 22 October 2012 10:36, Herbert Duerr h...@apache.org wrote:

 On 21.10.2012 18:16, jan iversen wrote:

 Now I am the one that is confused (again).

 Is the directory mediaWiki active or not ?


 The extension in that directory is built only if you used the
   --enable-wiki-publisher
 configure option.

 Herbert

Re: File: readme.xrm

2012-10-22 Thread Keith N. McKenna


jan iversen wrote:

As far as I can see on the usage are your assumption correct, and there
must be other ways to make different readme text platform dependent.

Would it not be ok, to have one readme for all platforms, and in the text
mention the specics ?

janI

On 22 October 2012 13:34, Jürgen Schmidt jogischm...@gmail.com wrote:


On 10/21/12 2:16 PM, jan iversen wrote:

There is exactly one file with extension .xrm

main/read_license_oo/docs/readme/readme.xrm

is there a reason (apart from history) for it being in .xrm or could it

be

converted to e.g. .xhp ?

If so we could get rid of one more conversion tool (read: does not need

to

be converted to new code).



I am not sure if xhp would be a good option here. But we can probably
switch to something else. Maybe a common readme file that gets extended
with platform specific portions from other files. When I remember it
correctly the xrm files contains the content for the readme file and
depending on the platform different content is extracted from this file.

Juergen





Jan;

That may indeed be one way to do it. My concern is that users will get 
frustrated trying to wade through the info for the other platforms and 
just not bother with it at all. Of course based on the way they read the 
release notes they probably don't read it anyway. Be that as it may do 
we want to give them another reason not to read it and possible miss 
pertinent information.


Regards
Keith

Re: File: readme.xrm

+1, That is a very good point !!

But may we can still write it as plain text, put some tags in, and use
sed to split when generating installation sets ?

janI

On 22 October 2012 18:00, Keith N. McKenna keith.mcke...@comcast.netwrote:

 jan iversen wrote:

 As far as I can see on the usage are your assumption correct, and there
 must be other ways to make different readme text platform dependent.

 Would it not be ok, to have one readme for all platforms, and in the text
 mention the specics ?

 janI

 On 22 October 2012 13:34, Jürgen Schmidt jogischm...@gmail.com wrote:

  On 10/21/12 2:16 PM, jan iversen wrote:

 There is exactly one file with extension .xrm

 main/read_license_oo/docs/**readme/readme.xrm

 is there a reason (apart from history) for it being in .xrm or could it

 be

 converted to e.g. .xhp ?

 If so we could get rid of one more conversion tool (read: does not need

 to

 be converted to new code).


 I am not sure if xhp would be a good option here. But we can probably
 switch to something else. Maybe a common readme file that gets extended
 with platform specific portions from other files. When I remember it
 correctly the xrm files contains the content for the readme file and
 depending on the platform different content is extracted from this file.

 Juergen



  Jan;

 That may indeed be one way to do it. My concern is that users will get
 frustrated trying to wade through the info for the other platforms and just
 not bother with it at all. Of course based on the way they read the release
 notes they probably don't read it anyway. Be that as it may do we want to
 give them another reason not to read it and possible miss pertinent
 information.

 Regards
 Keith

Re: [RELEASE]: propose a new developer snapshot based on revision 1400796

On 10/22/12 4:50 PM, Jürgen Schmidt wrote:
 On 10/22/12 4:24 PM, Ariel Constenla-Haile wrote:
 On Mon, Oct 22, 2012 at 03:22:33PM +0200, Jürgen Schmidt wrote:
 @Ariel, are you able to build the Linux version? 

 Yes, I can.
 Did you plan real Developer Snapshots with no system integration?
 did you plan only en-US full install sets + language packs?

 sorry for being not detailed enough. I plan dev snapshot without system
 integration for now.

 I can live the en-US full install sets  + language packs. But I am open
 here and I would provide what helps our volunteers doing testing and
 translation best.

we provide snapshot only every 2 weeks or every 4 weeks, maybe we should
provide full install sets. But we can easy change that if it doesn't work.

Would it help if we provide the Linux 64 bit version? I think we have
proper build client in place to build. That would reduce the amount of
data for the upload for you.

Juergen



 What do you think?

 Juergen


 PS: before you start please wait until I have at least finished one
 platform with my latest fix ;-)
 
 1400866 seems to be good, on MacOS I package already the install sets
 

 is this your with-lang switch?

 --with-lang=ar cs da de en-GB en-US es fi fr gd gl hu it ja km ko nb nl pl 
 pt-BR ru sk sl zh-CN zh-TW
 
 yes
 
 Juergen

Re: [RELEASE]: propose a new developer snapshot based on revision 1400796

On Mon, Oct 22, 2012 at 1:09 PM, Jürgen Schmidt jogischm...@gmail.com wrote:
 we provide snapshot only every 2 weeks or every 4 weeks, maybe we should
 provide full install sets. But we can easy change that if it doesn't work.

 Would it help if we provide the Linux 64 bit version? I think we have
 proper build client in place to build. That would reduce the amount of
 data for the upload for you.

I have no problems as long as we use the same switches and system base.

are you planning full install sets + language packs for every language?


Regards

Re: Automatically Generated extension.update.xml

2012-10-22 Thread imacat

So, who is managing extensions.openoffice.org?  Is it possible that we
can turn on the online update and try to solve this problem now?

On 2012/10/22 19:27, Jürgen Schmidt said:
 On 10/21/12 11:13 PM, Andrea Pescetti wrote:
 imacat wrote:
 I wonder if it
 is possible to generate the extension update information files
 extension.update.xml automatically on
 http://extensions.openoffice.org, instead of maintaining the update
 information files by the authors manually?

 This requires a slightly complex answer since there are differences
 between theory and practice.

 In theory, when you upload an extension to extensions.openoffice.org,
 you needn't provide an update XML file: the repository knows about the
 mechanism and automatically generates it. Moreover, OpenOffice knows
 where to ask for the update so, if you use the official repository to
 host your extension, updates will magically work with no need of actions
 from the developer (not even specify an update URL), aside possibly from
 making sure that the version number always increases.

 In practice, the process is broken. The point where it breaks is the
 fact that extensions.openoffice.org does not provide the real update
 information but it sends a hardcoded answer instead, saying that no
 updates are available. This was done back in the Oracle times for some
 reasons - probably traffic, but I have no clue here, maybe it was just
 broken.

 Fact is, what you ask for is already available in theory and I have seen
 it working for a period before it was disabled in 2010. But it isn't
 available in practice at the moment, and it hasn't been for years.

 On the application side (the OpenOffice program) everything should still
 be working, even though I've recently seen an issue where Ariel
 explained we might have some regressions on the OpenOffice side too:
 https://issues.apache.org/ooo/show_bug.cgi?id=121201
 
 mmh, I was not ware that this function was disabled in 2010 and I don't
 know the reason for it. But Roberto has explained it quite well and the
 idea behind it was exactly that for extensions hosted in our repo the
 update information is generated automatically and the extensions
 shouldn't add further update information in their description.xml.
 
 We should figure out if the mechanism still works and should enable it.
 
 Ariel reported indeed a problem with his updates but we have to check
 this in general because I at least was not able to reproduce all
 reported problems. But this is a separate issue and have to fixed
 independently.
 
 Juergen


-- 
Best regards,
imacat ^_*' ima...@mail.imacat.idv.tw
PGP Key http://www.imacat.idv.tw/me/pgpkey.asc

Woman's Voice News: http://www.wov.idv.tw/
Tavern IMACAT's http://www.imacat.idv.tw/
Woman in FOSS in Taiwan http://wofoss.blogspot.com/
Apache OpenOffice http://www.openoffice.org/
EducOO/OOo4Kids Taiwan http://www.educoo.tw/
Greenfoot Taiwan http://greenfoot.westart.tw/



signature.asc
Description: OpenPGP digital signature

Re: OpenOffice Developer Room (devroom) at FOSDEM

2012-10-22 Thread Louis Suárez-Potts


On 12-10-22, at 04:49 , Jürgen Schmidt jogischm...@gmail.com wrote:

 On 10/21/12 11:06 PM, Andrea Pescetti wrote:
 On 16/10/2012 Andrea Pescetti wrote:
 I've asked the organizers to inform us about their decisions. I expect
 to receive news in the next days; I'll keep you informed when I do.
 
 I received the answer just a few minutes ago: organizers decided to give
 OpenOffice its own devroom, which will be available all day Saturday (2
 Feb 2013). Now we will have to organize a call for talks and so on, but
 this is not urgent. As agreed with the FOSDEM organizers, I'm reposting
 their answer in public, see below.
 
 ok, a little bit surprising but a clear signal.
 
 We should try to find out if other Apache projects will be there as well
 and if we can share a booth with them. Having a booth is a good
 opportunity to talk with people and to answer questions.

Let's focus attention on this set of tasks and carry discussions on marketing 
list?

And be fairly rigorous about issuing the CfP; and also, external to Apache, 
soliciting funds, if that is okay with the AOO team, for subsidizing would-be 
participants' travel/lodging.

Louis

Re: svn commit: r835763 - /infrastructure/buildbot/aegis/buildmaster/master1/projects/openofficeorg.conf

2012-10-22 Thread Andrew Rist


Herbert,
This was a good try - but 
http://ci.apache.org/builders/aoo-win7/builds/356/steps/compile_1/logs/err.html

will look at it a bit later...
A.

On 10/22/2012 9:18 AM, h...@apache.org wrote:

Author: hdu
Date: Mon Oct 22 16:18:02 2012
New Revision: 835763

Log:
merge aoo-win7 weekly and nightly builders by using the scheduler property 
rsync_opts

Modified:
 
infrastructure/buildbot/aegis/buildmaster/master1/projects/openofficeorg.conf

Modified: 
infrastructure/buildbot/aegis/buildmaster/master1/projects/openofficeorg.conf
==
--- 
infrastructure/buildbot/aegis/buildmaster/master1/projects/openofficeorg.conf 
(original)
+++ 
infrastructure/buildbot/aegis/buildmaster/master1/projects/openofficeorg.conf 
Mon Oct 22 16:18:02 2012
@@ -27,14 +27,16 @@ c['schedulers'].append(Nightly(name='ope
minute=20))

  c['schedulers'].append(Nightly(name='aoo-win7',
-  builderNames=[],
+  builderNames=['aoo-win7'],
branch='incubator/ooo/trunk',
+  properties={'rsync_opts':''},
hour=4,
minute=30))

  c['schedulers'].append(Nightly(name='aoo-win7w',
-  builderNames=[],
+  builderNames=['aoo-win7'],
branch='incubator/ooo/trunk',
+  properties={'rsync_opts':'--delete'},
dayOfWeek=0,
hour=6,
minute=30))
@@ -239,7 +241,10 @@ c['builders'].append(b_ooo_1r)



-# building ooo source on Win7 - nightly incremental build
+# building ooo source on Win7, either
+# - a weekly full build
+# - a nightly incremental build
+# depending on whether the property rsync_clean is set to --delete or 
  # First lets create a variable to full path of cygwin bash.
  winbash = r'C:\cygwin\bin\bash.exe'

@@ -250,7 +255,7 @@ f_ooo_w1.addStep(Compile(command=[winbas
  haltOnFailure=True, workdir=build,
  ))
  # sync files from svn - deleting all other files for clean build - i.e. 
clobber
-f_ooo_w1.addStep(Compile(command=[winbash , '-cliex', 'cd 
/cygdrive/e/slave14/aoo-win7/ ; rsync -av --delete build_svn/ build/'],
+f_ooo_w1.addStep(Compile(command=[winbash , '-cliex', WithProperties('cd 
/cygdrive/e/slave14/aoo-win7/ ; rsync -av %(rsync_opts) build_svn/ build/')],
  haltOnFailure=True, workdir=build,
  ))
  f_ooo_w1.addStep(SetProperty(command=svnversion e:\\slave14\\aoo-win7\\build_svn, 
property=got_revision))
@@ -340,125 +345,16 @@ f_ooo_w1.addStep(MasterShellCommand(comm
  alwaysRun=True,
  ))

-b_ooo_w1 = {'name': aoo-win7w,
+b_ooo_w1 = {'name': aoo-win7,
'slavename': bb-win7,
'builddir': aoo-win7,
'factory': f_ooo_w1,
'category': aoo-win7,
}

-# c['builders'].append(b_ooo_w1)
+c['builders'].append(b_ooo_w1)


-
-# building ooo source on Win7 weekly clean build
-# First lets create a variable to full path of cygwin bash.
-
-f_ooo_w2 = factory.BuildFactory()
-
-f_ooo_w2.addStep(Compile(command=[winbash , '-cliex', 'cd 
/cygdrive/e/slave14/aoo-win7/ ; svn up build_svn'],
-haltOnFailure=True, workdir=build,
-))
-# sync files from svn - without deleting build files
-f_ooo_w2.addStep(Compile(command=[winbash , '-cliex', 'cd 
/cygdrive/e/slave14/aoo-win7/ ; rsync -av build_svn/ build/'],
-haltOnFailure=True, workdir=build,
-))
-f_ooo_w2.addStep(SetProperty(command=svnversion e:\\slave14\\aoo-win7\\build_svn, 
property=got_revision))
-
-f_ooo_w2.addStep(ShellCommand(command=[winbash , '-cliex', 'cd 
/cygdrive/e/slave14/aoo-win7/build/main ; cp -arfv ../../main_xtra/* .'],
-haltOnFailure=False, workdir=build/main,
-))
-



-f_ooo_w2.addStep(ShellCommand(command=[winbash , '-cliex', 'cd 
/cygdrive/e/slave14/aoo-win7/build/main ; /cygdrive/c/Program Files/MicrosoftVisual 
Studio 9.0/VC/bin/vcvars32.bat ; autoconf'],
-haltOnFailure=False, workdir=build/main,
-))
-
-f_ooo_w2.addStep(ShellCommand(command=[winbash , '-cliex', WithProperties('cd /cygdrive/e/slave14/aoo-win7/build/main ; /cygdrive/c/Progra~1/Microsoft Visual Studio 
9.0/VC/bin/vcvars32.bat ; ./configure --enable-verbose --with-dmake-url=http://dmake.apache-extras.org.codespot.com/files/dmake-4.12.tar.bz2; 
--with-cl-home=/cygdrive/c/Progra~1/MicrosoftVisual Studio 9.0/VC --with-jdk-home=/cygdrive/c/PROGRA~1/Java/jdk1.6.0_30 --with-psdk-home=/cygdrive/c/PROGRA~1/Microsoft 
SDKs/Windows/v6.1 --with-directx-home=/cygdrive/c/PROGRA~1/Microsoft DirectX SDK (June 2010) --with-mspdb-path=/cygdrive/c/Progra~1/Microsoft Visual Studio 
9.0/Common7/IDE --with-frame-home=/cygdrive/c/Progra~1/Microsoft SDKs/Windows/v6.1 --with-midl-path=/cygdrive/c/Progra~1/Microsoft SDKs/Windows/v6.1/Bin 
--with-asm-home=/cygdrive/c/Progra~1/Microsoft Visual Studio 9.0 --with-csc-path=/cygdrive/c/Progra~1/Microsoft Visual Studio 9.0/SDK/v3.5

Is there a way to force website rebuild?

It looks like parts of the CMS were down from last night until
recently.  I had some changes pending at the time on ooo-site and
site.  Now when I go in I see that the staged content does not reflect
the source files.

For example:

http://ooo-site.staging.apache.org/contact_us.html

But if you look at the source file you see the addresses have been
changed to, e.g.,  http://openoffice.apache.org/mailing-lists.html

I tried forcing an update from SVN to staging.  But that did not help.

I'd rather not retouch all of the files in SVN just to force a
rebuild.  Is there a way to force a rebuild?

-Rob

Re: Is there a way to force website rebuild?

2012-10-22 Thread Joe Schaefer

Make a trivial change to a file in templates/ or lib/

Sent from my iPhone

On Oct 22, 2012, at 2:19 PM, Rob Weir robw...@apache.org wrote:

 It looks like parts of the CMS were down from last night until
 recently.  I had some changes pending at the time on ooo-site and
 site.  Now when I go in I see that the staged content does not reflect
 the source files.
 
 For example:
 
 http://ooo-site.staging.apache.org/contact_us.html
 
 But if you look at the source file you see the addresses have been
 changed to, e.g.,  http://openoffice.apache.org/mailing-lists.html
 
 I tried forcing an update from SVN to staging.  But that did not help.
 
 I'd rather not retouch all of the files in SVN just to force a
 rebuild.  Is there a way to force a rebuild?
 
 -Rob

Re: Is there a way to force website rebuild?