Re: Suspicious target for files of commit r1556127
On 07.01.2014 13:26, Jürgen Schmidt wrote: Hi Regina, I think that was by mistake and Yushen Fan has already corrected this. She started to collect test documents for OOXML and make them available under openoffice/test/test-data/ooxml/... The naming of the openoffice/test directory is a bit unfortunate, because there is also the openoffice/trunk/test directory, which contains the automatic testing framework... That will allow our developers/testers easy access to the test files with a simple checkout. We should collect all kind of test document that are used in defined test-cases or for general QA purposes in this new area in svn. Juergen On 1/7/14 12:33 PM, Regina Henschel wrote: Hi, I see, that with commit r1556127 a lot of files has been saved directly to openoffice and not into a subfolder. That looks suspicious to me. Kind regards Regina - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Build braker in rejuvenate01 cppu
Hi Raphael, On 11.11.2013 10:47, Raphael Bircher wrote: I run in a build breaker in the rejuvenate01 branche dmake: Error: -- `uno_purpenvhelpers5abi.map' not found, and can't be made I just committed r1540968 to add the missing file by renaming its older cousin. Has someone a idea what happend here, and what's wrong? The rejuvenate01 branch had its UNO bridge renamed. UNO bridges are named after the calling convention supported by them. C++ calling conventions used to be very specific to the compiler/platform/etc. but things have settled and became standardized. With the rejuvenate branch we are using the standard C++ tool chain on the OSX platform (XCode4/5) and build for the standard C++ calling conventions [1]. So the name I originally chose (cxx) emphasized the fact that the new UNO bridge was for the standard C++ ABI on the target platform. [1] https://developer.apple.com/library/mac/documentation/developertools/conceptual/LowLevelABI/000-Introduction/introduction.html Other developers disagreed with naming it cxx (for C++) as that name is quite generic. So I mentioned the fully qualified ABI name of the standard C++ ABI on this platform which is System V Application Binary Interface [for the] AMD64 Architecture and Jürgen abbreviated it as s5abi. Thanks for looking into the rejuvenate01 branch. With the renaming and the recent major update from trunk it has become a construction site though. After the RmMoz work I can now focus on getting the branch round again and make it a viable port. So stay tuned, in a few days it will build+run out of the box again and it will have gained an important feature it missed so far: support for password protected documents! Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: XML files are binary?
On 12.11.2013 15:34, Andre Fischer wrote: On 12.11.2013 15:19, Rob Weir wrote: On Tue, Nov 12, 2013 at 8:59 AM, Andre Fischer awf@gmail.com wrote: I just checked in some files for creating Windows patches (no, the work is not yet finished) and saw that the .xml files are flagged as binary. Checking their properties revealed that their mime type is correctly set to application/xml. Does anybody know why they are still flagged as binary? Do you have svn:eol-style set? No, svn:mime-type is the only property that was (automatically) set. Maybe its value should be text/xml instead of application/xml. Thats also what http://www.apache.org/dev/svn-eol-style.txt recommends: Please do not use svn:mime-type=application/xml for xml files as that causes SVN to treat them as binary - use text/xml or omit the setting entirely. Now if svn's automatic property setting mechanism agreed with this too all would be fine. Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
AOO Security Features without Mozilla
About everyone who ever built OpenOffice in the last couple of years wondered why an almost complete (and obsolete/unmaintained/ancient) version of Mozilla Seamonkey was needed when building OpenOffice with its security features enabled such as support for password protected documents. The branch Remove_MOZ shows that it is possible to get rid of that dependency and I suggest we do that as soon as possible. The branch was inappropriately named because it is only about the removing the mozilla dependency of security related stuff. But the old Seamonkey binaries still have another purpose: for now they are needed for providing its own address books that used to be in the Mork format. It also provides access to some address books [1] such as LDAP, Outlook and Outlook Express. [1] http://www.openoffice.org/dba/specifications/address_book_architecture.html Other address sources such as JDBC, ODBC, CSV-Text, MySql and dBase already work without Mozilla. On Mac the native Address Book is already supported directly. Since issue 91209 the mozilla address books were disabled on Mac altogether anyway, so on Mac we could rid AOO of its heavy Seamonkey dependency really soon without removing any features by using NSS instead of bundling a large set of Seamonkey libraries. On the other platforms a very high percentage of our user base wouldn't notice any missing features if the Mozilla address book support was removed there too. Developing mozilla-less replacements should be possible and this would remove a lot of complexity. As a first idea the replacements could be implemented as extensions using something like [2] for LDAP, [3] for Mork and [4] for WAB if there was an UNO API to facility that support. Comparing the complexity of the scripts below vs the complexities and maintenance headaches the ancient Seamonkey and its XPCOM-UNO bridge is like comparing the weight of mice to elephants... [2] http://www.python-ldap.org/ [3] https://bug241438.bugzilla.mozilla.org/attachment.cgi?id=175024action=view [4] http://stackoverflow.com/questions/11538550/retrieving-outlook-contacts-via-python But splitting off the security dependency is much more important. I plan to integrate the changes needed for that soon. They will be enabled either with --enable-nss-module or with the more general option --enable-category-b For the mozilla address books I plan to add the option --enable-mozab-module to replace the then way too broadly named option --enable-mozilla Until the replacements outlined above have been developed this new option will allow bundling of the old Seamonkey binaries for users that depend on its address book support. Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: AOO Security Features without Mozilla
On 22.10.2013 13:46, janI wrote: On 22 October 2013 13:30, Herbert Dürr h...@apache.org wrote: [...] Since issue 91209 the mozilla address books were disabled on Mac altogether anyway, so on Mac we could rid AOO of its heavy Seamonkey dependency really soon without removing any features by using NSS instead of bundling a large set of Seamonkey libraries. On the other platforms a very high percentage of our user base wouldn't notice any missing features if the Mozilla address book support was removed there too. I have no problem with that, since it makes our product lighter and simpler. But for this I think we need user opinions. Yes, whether we build our next release with the --enable-mozab-module option or without it is open for discussion. Now if there were volunteers that implemented extensions for mapping mork/ldap/wab address books to AOO's SDBC API then the whole mozab module would be superfluous anyway and the discussion would have only one reasonable result. [...] For the mozilla address books I plan to add the option --enable-mozab-module to replace the then way too broadly named option --enable-mozilla Just to be sure, you will not add the option, but rename the other option, so we only have --enable-mozab-module ? Up to now the enable-mozilla option was for enabling security and the mozilla address book. With the rework only the mozab aspect would remain and so the option should be renamed to be more concise. Until the replacements outlined above have been developed this new option will allow bundling of the old Seamonkey binaries for users that depend on its address book support. please keep the number of new options and changes in configure as low as possible, that helps me :-) I intend to keep it as simple and concise as possible. Its a good initative, which I highly support, and once you have integrated it into trunk I will update my branches. Thanks! Actually the rejuvenate branch seems to be a bigger candidate for conflicting changes, but we will take that when its ready. Yes, working on and improving the rejuvenate branch is very important. The security-enabling branch removemoz is a prerequisite though as the old seamonkey is not nearly ready to be compiled on modern platforms. Investing energy in resurrecting the obsolete version or replacing it with a more modern version would be a waste of time, especially when there is a way for us to kill this incredibly heavy dependency for good. Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: [VOTE]: Release OpenOffice 4.0 (RC2)
On 17.07.2013 09:40, Jürgen Schmidt wrote: [...] The RC is based on the release branch AOO400, revision 1503704! Please vote on releasing this package as Apache OpenOffice 4.0. The vote starts now and will be open until: UTC 3:00pm on Friday, 19 July: 2013-07-19 3:00 UTC. But we invite all people to vote (non binding) on this RC. We would like to provide a release that is supported by the majority of our project members. [ ] +1 Release this package as Apache OpenOffice 4.0 [ ] 0 Don't care [ ] -1 Do not release this package because... +1 Release this package as Apache OpenOffice 4.0 Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Build problem on epm in the rejuvenate01 branche
Hi Raphael, On 2013/05/30 8:27 PM, Raphael Bircher wrote: Mac OS X 10.7 with XCode 4.3 My build stops at the epm modul. I know there is a workflow to find more details about the breaker, but I fergot it :-( Can sameone help me? Go into the module that had the problem and run build verbose=t there. Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: [RELEASE]: proposed further schedule towards AOO 4.0
[regarding dropping stlport4] The changes to make the codebase ready for native STL support are done. Builds with stlport4 enabled will continue to work as before. I suggest to use the --without-stlport option for all new builds though: Stlport is a great project, but the versions that OOo depended on had been released more than ten years ago. The library improved greatly since then from a feature, performance and standard compliance perspective. And of course many many bugs have been fixed [1]. In their stlport5 version they continue to improve significantly. Platform STLs have been inspired by stlport, improved greatly too and in the C++11 standardization process divergent views have consolidated. We can rely on the platform STLs. I agree that the timing of the suggested switch is not so good but the switch itself is overdue. A major version change is the right time to do this. [1] relevant examples of fixes that got into stlport releases newer than the ones OOo depended on can be seen at e.g. http://stlport.git.sourceforge.net/git/gitweb.cgi?p=stlport/stlport;a=blob;f=etc/ChangeLog;hb=refs/tags/STLport-STLPORT_4_6 On 2013/05/28 2:38 PM, Rob Weir wrote: In theory every fix can cause bugs. But some fixes are more localized than others. Fixes with localized impact are easier to test. Widespread fixes are harder to test, because more code is potentially broken. The switch was rendered possible by many little changes over the last couple of months which got our code base more in line with C++11 expectations. Snapshots based on these changes have been and are already extensively tested by our great QA community. The switch itself is just another step in evolving towards a high quality release. Additionally testing has it much easier to find issues introduced by the switch should there be any. E.g. we have many testers and almost a thousand automatic tests. They work on different platforms. They cover a lot of different areas. The risk that a regression in that layer could remain undetected is very low. Automated testing ran its 940 autotests (in BVT, FVT, SVT and PVT) on different operating systems for 32bit and 64bit versions. The cross-correlation between pre- and post-switch builds is the same as the auto-correlation for test reruns: the same tests were successful on both sides, the same tests failed for both sides. Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: GIT mirror.
On 2013/05/28 2:50 PM, janI wrote: We have INFRA 5590 pending. When I (and infra) reads it, its only a request for a RO mirror. Yes, it was assumed that a read-only mirror of our codebase would be an essential step regardless on how we want to proceed. A read-only mirror would have the immediate benefit of being a stable reference: As SVN allows editing properties of older commits individual imports of the codebase via git-svn are different and result in unpredictable commit-ids which prevents an efficient DVCS working style. An official git-mirror of our SVN sub-tree would solve all that. The questions in https://issues.apache.org/jira/browse/INFRA-5590 need to be addressed first in either way: do we want the localizations split from the codebase? How to map the one-level/two-level deep branch names? Do we want to have the incubator/ooo history grafted before the post-graduation history? And how about the OOo-history? etc. I understood from another discussion, that we wanted to be able to do git commit, that automatically updated the svn (orignial) repository. I'm afraid only unidirectional mirroring will work reliably. This means we'll either a) have SVN as upstream of a read-only mirror b) have a git repository as new upstream a) as an intermediate step is needed anyway and it would solve a lot of problems. I personally would prefer option b in the long term, but not having it would only be a minor inconvenience. Not having any official git-mirror is a much bigger problem. I my understanding is correct the ticket should be updated to reflect that (and maybe less open possibilities). I suggest to keep the original request and add comments or answers to the JIRA issue. As the author of INFRA-5590 I'll be happy to change the subject line then to clarify what we want. Do we have committers who would prefer to continue with SVN? Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Question about: OpenOffice Calc. =-1^2 like expressions
On 2013/05/27 11:20 AM, Torokhov Sergey wrote: There is a some strange behaviour in OpenOffice Calc (and in Microsoft Office too). Did it was made in OpenOIffice for compability with MSO? This seems to concern all combinations of even degrees and minus sign in the begining of formula 1. Enter in cell the expression: =-1^2 The result will be equal 1 instead of -1 Note that =1-2^2 results in -3 and it's true but -2^2 results 4 instead of -4 2. Well, now enter the expression: =-(1)^2 The result will be equal 1 instead of -1 again (or expression like =-(1+2)^2 the result will be 9 instead of -9) 3. Only if enter the expression: = -(1^2) The result now will be right and equal -1 Such behaviour contradicts mathematical notation of formula. I'm quite sure AOO's way of interpreting such an expression was determined by compatibility considerations. For what its worth Wikipedia [1| also states these different conventions: There exist differing conventions concerning the unary operator − (usually read minus). In written or printed mathematics, the expression −32 is interpreted to mean −(32) = −9,[3] but in some applications and programming languages, notably the application Microsoft Office Excel and the programming language bc, unary operators have a higher priority than binary operators, that is, the unary minus (negation) has higher precedence than exponentiation, so in those languages −32 will be interpreted as (−3)2 = 9.[4] In cases where there is the possibility that the notation might be misinterpreted, parentheses are usually used to clarify the intended meaning, however due to the syntax of most major programming languages, it is usually hard or impossible to be ambiguous. [1] http://en.wikipedia.org/wiki/Order_of_operations#Exceptions_to_the_standard Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Alternative to ooo-extras?
Just a heads up: We are using ooo-extras [1] as reliable mirror to the upstream locations of third party code with non-permissive licenses. Today's google opensource blog [2] says that this method will stop working in 2014. We need to find an alternative. Herbert [1] https://code.google.com/a/apache-extras.org/p/ooo-extras/ [2] http://google-opensource.blogspot.com/2013/05/a-change-to-google-code-download-service.html - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Layout improvements (Re: [Pootle]Different behaviour on different browsers)
On 2013/05/19 12:16 AM, Andrea Pescetti wrote: On 16/05/2013 Ariel Constenla-Haile wrote: On Thu, May 16, 2013 at 03:41:11PM +0200, Jürgen Schmidt wrote: a more generic layout engine for VCL. Thinking of an extended or modified format to describe dialogs and include layout info/constraints that are used by the layout engine etc. Christian Lippka had this almost finished. I recall the picture from this thread http://www.mail-archive.com/dev@de.openoffice.org/msg30669.html it had a side panel defined in xml. It would be interesting to investigate if the work was done on some private repository in Hamburg, or if it is in some CWS on http://hg.services.openoffice.org/; and rescue it, avoiding to reinvent the wheel. He was also doing a nice job with the svtools::ToolbarMenu in the toolbar controllers: http://www.youtube.com/watch?v=_eO0X_eO6Ww Christian is an Apache OpenOffice committer and already contributed his improved color picker to Apache OpenOffice, so if he wants he can merge his improvements any time. The video seems to be from the Renaissance4 CWS, still in HG. I absolutely don't know whether this CWS contains the layout engine improvement too, though. I talked to Christian and he confirmed that the video was from the code in the http://hg.services.openoffice.org/cws/renaissance4/ child workspace. He doesn't think that the XML-defined sidepanel work got far enough to be worth picking. I'm afraid Christian is too busy in his new job to work further on such OpenOffice topics, but everyone knows him to be very constructive and cooperative. I'm BCC'ing him in case he wants to jump in. Hi Christian! :-) Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: svn commit: r1482736 - /openoffice/trunk/main/sal/inc/rtl/string.hxx
On 2013/05/16 1:23 AM, Ariel Constenla-Haile wrote: On Wed, May 15, 2013 at 09:22:51AM -, h...@apache.org wrote: [...] #i122208# avoid signedness warnings in signed-char build environments [...] This breaks on Linux: Sorry about that. Fixed in r1483216. Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: svn commit: r1483216 - /openoffice/trunk/main/sal/inc/rtl/string.hxx
On 2013/05/16 4:08 PM, Ariel Constenla-Haile wrote:
On Thu, May 16, 2013 at 09:03:26AM -0500, jorge ivan poot diaz wrote:
Hello Ariel,
In my local directory:
size_t operator()( const char* p) const {
size_t n = 0;
while( *p)
n += 4*n + *static_castunsigned char*(p++);
return n;
Update your local copy, Herbert fixed this already.
Yup. Sorry again. It worked on my current main system which defaults to
unsigned chars...
Herbert
-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: New branch rejuvenate01
On 2013/05/08 8:25 PM, janI wrote: On 8 May 2013 20:16, Herbert Dürr h...@apache.org wrote: I'd like to point to the new branch named rejuvenate01 for advancing AOO's platform support. The main focus is the upgrade of our codebase. E.g. the MacOSX platform support had nearly stagnated since 2007 and thus even getting the build requirements right on a current Mac was becoming very very difficult and with the rejuvenation it should not only work out of the box, but make leaps ahead. This is very interestingmay I come with one small suggestion...I do understand your choice of branch name, but it would be nice if the name reflected the work being done...that makes it easier to identify and later to find (if needed) once reintegrated. I agree, but since a lot of this stuff had co-dependencies (e.g. 64bit JRE - 64bit UNO - XCode4 - clang - allows C++11 - no-stlport4 and XCode4 - newer SDK) it wasn't easy to come up with a name that covered the different aspects. Platform refresh 2013 and rejuvenate were the only ones that came to my little mind. Maybe I should have asked around more for better ideas. In my defense I mentioned the name pr2013 as a suggestion for the branch in my 2013/04/09 mail on the XCode4 topic and nobody complained or had a better idea then. If you need help with testing on ubuntu 12.04 I will be happy to assist. Thanks! I'm afraid the configure scripts for this target platform are not yet fully in shape for direct building and testing... but as far as I know you that doesn't scare you away in the least :-) Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: [build] source of epm source package no longer available
On 2013/05/02 9:39 PM, Andrea Pescetti wrote: Andrew Rist wrote: On 5/2/2013 12:22 AM, Andre Fischer wrote: There is also the version that we use at the moment: http://www.msweet.org/files/project2/epm-3.7-source.tar.gz changes made and linux 64 build kicked off... I've just updated the Building Guide accordingly: http://wiki.openoffice.org/wiki/Documentation/Building_Guide_AOO#Configuration_and_bootstrapping Thanks! The only thing that has to be changed is configure.in, where there is a suggestion for the download (line 132 where the --with-epm-url option is defined). But copying it to apache extras would probably be better. Any volunteers to make the necessary changes? Done [1]. I also added an alternative location [2] that seems to be more generic. It also has the matching MD5. [1] http://svn.apache.org/viewvc?view=revisionrevision=r1478665 [2] http://epm.sourcearchive.com/downloads/3.7-1/epm_3.7.orig.tar.gz Both proposed changes make sense too, of course. In general, we should consolidate build dependencies (libraries, dictionaries) on as few servers as possible, to avoid problems like this one. +1 We already have [3] as backup to upstream locations for the category-B stuff. Can we put category-X stuff up there too? [3] https://code.google.com/a/apache-extras.org/p/ooo-extras/downloads/list Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: OUStringBuffer
Hi Jorge, like Regina I have a hard time to understand the question. My best guess is that you have an OUStringBuffer object and need an UniString (aka tool's String [1]). If that is so you can get it using the UniString constructor that takes an OUString argument: UniString( buf.makeStringAndClear() ) [1] http://opengrok.adfinis-sygroup.org/source/xref/aoo-trunk/main/tools/inc/tools/solar.h#269 Hope that helps, Herbert On 2013/04/25 8:00 PM, jorge ivan poot diaz wrote: I've already made a concatenation, in this file: http://opengrok.adfinis-sygroup.org/source/xref/aoo-trunk/main/cui/source/tabpages/tpcolor.cxx#546 Something like this: OUStringBuffer buf( 128 ); buf.append( aName ); buf.append( 4 ); OUString string = buf.makeStringAndClear(); OString oString = rtl::OUStringToOString( string , RTL_TEXTENCODING_ASCII_US ); Substituting aName by oString in this fragment of code http://opengrok.adfinis-sygroup.org/source/xref/aoo-trunk/main/cui/source/tabpages/tpcolor.cxx#549, I have the following error, http://ooo.pastebin.ca/2366521 So, I want to convert this code to a STRING: OString oString = rtl::OUStringToOString( string , RTL_TEXTENCODING_ASCII_US ); Because the definition of aName is STRING, as you can see here: http://ooo.pastebin.ca/2366523 The file is located at: http://opengrok.adfinis-sygroup.org/source/xref/aoo-trunk/main/svx/inc/svx/xtable.hxx#57 - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Unable to build on MacOS 10.8
On 2013/04/23 5:24 AM, Dehua Xu wrote: http://wiki.openoffice.org/wiki/AquaBuild#Requirements; it's old for the mac os x 10.8,and downloading xcode 3 is difficult. Yes, that is one important reason why we need to update it. The branch will be announced as soon as it successfully configures and compiles. And eventually even before the configuration checks are ready... could you give some suggestion about configures on xcode 4 ? For now the best way is to manually set the CC, CXX and LD environment variables, disable all the checks in configure that depend on gcc4.0 and update the SDK references from 10.4 to e.g. 10.7. This all should become simpler soon though. Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Xcode 4.6 compiled aoo 3.5.0
On 2013/04/22 9:29 AM, Dehua Xu wrote: Recently ,i want to build aoo 3.5.0 on mac x 10.8 with Xcode4.6. when i run ./configure,it's error about gcc-4.0 compiler. Please see http://wiki.openoffice.org/wiki/AquaBuild#Requirements Currently the build requirements are XCode 3 and a OSX 10.4 SDK. This will change in the near future to XCode4 and the SDKs provided by it, though. Could you please tell which branch i can successfully compile? The branch will be announced as soon as it successfully configures and compiles. And eventually even before the configuration checks are ready... Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Clean build proposal
On 2013/04/16 9:12 AM, Andre Fischer wrote: On 15.04.2013 18:12, Herbert Dürr wrote: On 2013/04/15 4:00 PM, Andre Fischer wrote: I would like to change the build type of the Windows7 nightly build from incremental to clean. Reasons in favor of a clean build are: - Developers don't have to know if a change is incompatible and thus have to (somehow) trigger a clean build. - Developers don't have to learn how to trigger a clean build. - We avoid very subtle bugs that are introduced when an incremental build is made on incompatible changes. Issues 122047 and 122048 may be caused by such a broken build: the office works in general, only commands are mapped to the wrong ids. It is not acceptable to have builds that we can not rely on. I'm having a deja vue moment ;-) [1] http://markmail.org/message/lpprmeahwurvepx5 No, this thread does not apply. It was centered on the idea to change the bootstrap file and thus to affect every build, not just the ones on the buildbot. I don't want to change any files in SVN, only the buildbot setup. The central suggestion was to default to a clean build because of the observation that dependencies currently don't work reliably which results in artifacts that are not reproducible and waste a lot of time was a direct hit. It is clear that the cleaning has to be somewhere between configure and the build steps. Where exactly was almost irrelevant. Bootstrap was only suggested because it is the only step between configure and the actual build. The build --clean suggested by Oliver had the problem that a simple build restart, e.g. after one of the infamous dmake failures, should not accidentally delete all the work already done. I am thankful for your and Andrew's work but you describe a problem with the current process, not with clean builds. I prefer clean builds too and now that I'm hearing Andrew that is happy with the change I'm all the happier with that. I already did it weeks ago for the snapshot builds a couple of weeks ago which resulted in some (minor) opposition then ;-) BTW, was that a -1 or a 0 to my proposal? It was a support vote for Andrew's evaluation of the situation who quite clearly described why he liked the incremental build. Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Clean build proposal
On 2013/04/15 4:00 PM, Andre Fischer wrote: I would like to change the build type of the Windows7 nightly build from incremental to clean. Reasons in favor of a clean build are: - Developers don't have to know if a change is incompatible and thus have to (somehow) trigger a clean build. - Developers don't have to learn how to trigger a clean build. - We avoid very subtle bugs that are introduced when an incremental build is made on incompatible changes. Issues 122047 and 122048 may be caused by such a broken build: the office works in general, only commands are mapped to the wrong ids. It is not acceptable to have builds that we can not rely on. I'm having a deja vue moment ;-) [1] http://markmail.org/message/lpprmeahwurvepx5 If no one objects in the coming 72 hours then I will change the buildbot configuration file to make clean Windows7 nightly builds. I will then try to pinpoint and fix any build errors. Any help is appreciated. As Andrew pointed out in his mail he prefers an incremental build on the w7bot for several reasons such as problems with hung processes and wrong dependencies that currently have to be cleaned out manually by him, as he is the only one with machine level access. So forcing a clean build would add additional work on him as volunteer. I trust Andrew's assessment. Server load was a another serious concern. Regarding the argument that a buildbot should always do a clean build whereas builds outside of that system should default to be incremental: There is some merit to the concept that the buildbot shouldn't always have to use some extra steps that the default build does not do. I have started to collect the available documentation on the buildbots [1] but many parts are missing. For example I have not yet found much on the buildbot configuration file. Anyone who knows more, please update the wiki page. Anyone considering a change to the configuration file [2] is well advised to get familiar with its extensive documentation [3] provided by the buildbot project first. Every slave-bot only has about a dozen steps... [2] https://svn.apache.org/repos/infra/infrastructure/buildbot/aegis/buildmaster/master1/projects/openofficeorg.conf [3] http://docs.buildbot.net/current/manual/configuration.html Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: How to make install package name changed to 4.0
Hi On 2013/04/11 4:50 PM, Ji Yan wrote: I have a question about the product version in installation package name, is the version controlled by aoo.ver in srcrelease.xml under folder https://svn.apache.org/repos/asf/incubator/ooo/trunk/main/solenv/bin/srcrelease.xml Please not the our repository moved from incubator/ooo to openoffice so the file you mentioned is now at https://svn.apache.org/repos/asf/openoffice/trunk/main/solenv/bin/srcrelease.xml Currently, the version is 3.5, should we update it to 4.0? The file at the new location is already updated to 4.0. Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Build breaker and clean build
On 2013/04/10 10:09 AM, Andre Fischer wrote: tonight we had a build breaker in the windows build: a slot id that is used in SW had been removed in SVX. The reference in SW had also been removed, so this change should not be a problem. But the windows build is still not a full build. Therefore the old SW slot header files where used and the build broke. There is an easy fix for situations like this: a clean build. Incremental build are known to have problems thats why I suggested [1] to default to a clean build. That didn't receive consensus though and indeed there are good reasons against it: The incremental build both tests the dependency system and it reduces the load when building significantly. On the already strained buildbot this means a factor of almost five improvement as clean build takes about 4.5h whereas an incremental build takes only 0.5-1.0h. Andrew even had to reschedule the snapshot build away from the weekly clean build because the buildbot load is a real problem. [1] http://markmail.org/message/wmlhc5f5zaiiyu2o [2] http://markmail.org/message/7q64ijlwygdqmwf3 This may become important in the coming weeks when we have to fix some bugs in the sidebar (which is about to be merged back into trunk). The sidebar is implemented in several modules. Without a clean windows build we will run into build breakers very regularly. It is possible to force a clean build manually. Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Build breaker and clean build
On 2013/04/10 10:27 AM, janI wrote: I am no expert in what we can do in buildbot, but isnt it possible to run a script before build, that removes all unxlngx6.pro directories and solver, that should secure a clean build ?? Each nightly unxlngx6.pro buildbot run already does clean build. Only the nightly windows build is incremental whereas the weekly windows trunk and snapshot builds are done from scratch. I suggested to default to a better safe than sorry approach of a clean build but that was rejected at that time (2012/2) [1] [1] http://markmail.org/message/wmlhc5f5zaiiyu2o Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Build breaker and clean build
On 2013/04/10 11:15 AM, Andre Fischer wrote: On 10.04.2013 10:33, Herbert Dürr wrote: On 2013/04/10 10:09 AM, Andre Fischer wrote: tonight we had a build breaker in the windows build: a slot id that is used in SW had been removed in SVX. The reference in SW had also been removed, so this change should not be a problem. But the windows build is still not a full build. Therefore the old SW slot header files where used and the build broke. There is an easy fix for situations like this: a clean build. Incremental build are known to have problems thats why I suggested [1] to default to a clean build. That didn't receive consensus though and indeed there are good reasons against it: The incremental build both tests the dependency system and it reduces the load when building significantly. On the already strained buildbot this means a factor of almost five improvement as clean build takes about 4.5h whereas an incremental build takes only 0.5-1.0h. Hm, we prefer a fast build over a correct build? Some developers apparently do. Please see the mail thread at [1] [1] http://markmail.org/message/wmlhc5f5zaiiyu2o I don't think that that is the right priority. Also, a build that breaks due to errors caused by not starting clean is wasted time, ours and that of the build server. I agree. Andrew even had to reschedule the snapshot build away from the weekly clean build because the buildbot load is a real problem. Would it help to do the windows build only every second day (night)? Doing the regular build for trunk only every second night would only help if the nightly builds for branches such as l10n or ia2 would also be reduced. Do they still need nightly builds? Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: XCode 4.x, any progress?
Hi Raphael,
Do you have any progress in make Apache OpenOffice run on XCode 4? Do
you have allready any code in this area?
This has some different aspects:
1. compiling with XCode4's clang
2. using a standard compliant STL
3. using SDKs directly supported by XCode4
4. replacing deprecated APIs with their modern counterparts
5. supporting newer architectures such as AMD64
For the clang compilation (1.) I already committed many changes for this
that don't interfere with the other builds. Clang was quite good in
finding real problems. Many regressions happened because the new
environment resulted in problems in the bridges, in UNO, etc, that were
mostly identified and solved. They need a to be cleaned up more. There
are many more things to do such as extending the configure script and
the platform specific solenv parts.
As clang doesn't like stlport4 and using a standard compliant STL is a
good idea anyway, especially since we now have the chance to switch to
it for the major version change, I'm working on this (2) too. For now
I'd like to keep these changes minimally invasive to reduce the risk of
regressions. I try to do this by emulating the stlport4 peculiarities
with standard facilities using a shallow wrapper. E.g. stlport4's
hashconst char* functor is identified and replaced, hash_map can
mostly be emulated by an unordered_map and cases where this is not
possible are identified and replaced. This approach allows us to become
native by switching to the platform's/compiler's native STLs and later
drop the shallow emulation layer without becoming incompatible then.
There is also the interesting topic that e.g. libc++ doesn't allow
recursive declarations such as struct S{mapint,Sm;};. As using
libc++ with clang would be a good idea such declarations should be
replaced. A short-circuit/fallback for now could be to use boost's tr1
containers because they explicitly allow recursive declarations.
On the SDK question (3) I suggest to drop support for platforms that
even Apple doesn't support anymore such as OSX10.6. If Apple cannot
afford to support it anymore with regular releases we shouldn't strain
our infinitely smaller resources to achieve this.
On the API frontier (4) we need to support e.g. CoreText and drop ATSUI
(except for debugging layout changes). There are also other topics that
need to be looked into such as the native widget framework.
On Mac things already look quite good except for some crashes in
CoreText's cascade code, Writer services, etc. that need to be solved.
For Linux and Windows getting them to build and pass BVT using the TR1
compliant STL needs more work. When the emulation is cleaned up enough
I'll open a branch for the platform refresh 2013 = pr2013
Volunteers are very welcome... first some loose ends have to be tied though.
Herbert
-
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Problem with module VCL (libXinerama missing)
On 2013/03/25 8:30 PM, Alan Eduardo Puc Pech wrote: Hello team, I'm Alan, Now I have a problem about building AOO with the module VCL. The error this is: [...] /usr/bin/ld: cannot find -lXinerama Please make sure that you have a package libxinerama-dev (or similar, depending on the distribution you are using). When you have fixed the errors in that module you can resume the build by running: build --all:vcl I would like to know what this error is? because I went to this directory: /home/alan/aoo/main/solver/400/ unxlngi6.pro/workdir/LinkTarget/Library/libvclplug_gen.so and I did not find this file: libvclplug_gen.so. I need your help. The file libvclplug_gen.so would be the result library. It isn't build if one of its dependencies (in this case libXinerama) is missing. Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: w7snap build error
On 2013/03/26 8:23 AM, Ariel Constenla-Haile wrote: On Wed, Mar 20, 2013 at 12:08:38PM +0100, Oliver-Rainer Wittmann wrote: I have committed the suggested patch from Ariel. Herbert was so kind to trigger the buildbot. It seems this doesn't fix anything: http://ci.apache.org/projects/openoffice/buildlogs/winsnap/main/officecfg/wntmsci12.pro/misc/logs/registry.data.org.openoffice.Office.UI.txt Please note that this is an old log as build #41 on aoo-w7snap didn't even make it to the upload step, so the old logs were never overwritten. Now why the buildbot stopped and timed out after 23h is a mystery to me too. Someone with direct access into that machine should have a look at the machines log files. Herbert - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: What is the build environment?
Hi! On 2013/03/05 6:51 AM, Yi Xuan Liu wrote: I tried to install the latest build on my machine, on which the OS is Ubuntu 10.04. And I cannot start AOO after installation. The error is GLIBC_2.15 not found, which is due to the lower OS version. What is our build environment? Is there any limitation for OS version? The build platform depends on what you mean with latest build. E.g. for our nightly buildbot builds we are using the ASF provided machines dedicated for this service. http://apache.org/dev/machines shows that our nightly 64bit-Linux build is created on AMD64 Ubuntu 10.04.4 LTS and our nightly 32bit-Linux build is created on i386 Ubuntu 12.04 LTS. So if you were trying to install the 32bit nightly Linux build the problem can be easily explained: On Linux building forward compatible binaries is rather easy nowadays, but the same isn't true for building backwards compatible stuff. For our releases, betas etc. we (thanks Ariel!) is doing the build on rather old systems so that they have a very good binary compatibility. So in short: If you want to test the Linux-Nightly on Ubuntu 10.4 (or newer) please use the 64bit version. If you want to test the 32bit Linux-Nightly as provided by the current buildbot (bb-vm2_ubuntu_32bit) please use 32bit Ubuntu 12.4 or later. Hope that helps, Herbert By the way I just committed r1452693 so that the configure logs will provide more details of the build platform. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Enable-werror
On 2013/02/27 2:10 AM, jorge ivan poot diaz wrote: I want to know if it is advisable to work on the source code running: ./configure --enable-werror I'm working on Ubuntu 12.04.1 LTS. For code you are working on treating warnings as errors is a very good idea. On the other hand it becomes quite distracting if warnings happen in unrelated code and the werror option would cause the break the build there. On platforms like Windows or Mac where we are currently still require the compilers VC2008 or XCode3 no new warnings are expected to appear until we switch these requirements, so enabling werror is encouraged. Plenty of new helpful warnings appear on more recent compilers, e.g. on the platform you are working on (in general: platforms with regular updates to the development environment like Linux/FreeBSD). Unless you plan to fix them I'd advise against enabling werror there. Looking at the warnings and solving the most critical ones is always interesting and beneficial for the project though. Hope that helps, Herbert
Re: java 7 patch
On 2013/02/25 1:09 AM, Pedro Giffuni wrote: While here let me remind you that to solve the jdk7 issue you have to look at i121098. The fix is rather trivial but someone else has to do it. Thanks for the pointer. I adjusted the produced classfile format to our baseline as defined in the checks during configure which is Java 5. Even this version was released in 2004 so it is old enough to be upgraded should the need arise. Herbert
Re: Fwd: Tutorial Writer
On 2013/02/19 2:56 AM, Alan Eduardo Puc Pech wrote:
Hi team!
I'm Alan. I'm building the module writer sw
In one part of the tutorial I want insert the following code:
--- sw/source/ui/app/mn.src 2004-11-29 14:29:10.0 +0530
+++ sw/source/ui/app/mn.src 2004-12-21 20:28:27.0 +0530
@@ -945,6 +945,13 @@ Menu MN_GRF_POPUPMENU
SEPARATOR ;
MenuItem
{
+Identifier = FN_SAVE_GRAPHIC ;
+HelpID = FN_SAVE_GRAPHIC ;
+Text [ de ] = ~Grafik speichern... ;
+Text [ en-US ] = Save ~Graphic... ;
+};
+MenuItem
+{
Identifier = FN_FORMAT_GRAFIC_DLG ;
HelpID = FN_FORMAT_GRAFIC_DLG ;
Text [ de ] = ~Grafik... ;
and I did so:
Menu MN_GRF_POPUPMENU
{
ItemList =
{
MN_ALIGN_FRAME
MN_MOUSE_FRAME_WITH_CONTOUR
// -- OD 2009-07-14 #i73249#
SEPARATOR ;
MN_TITLE_DESCRIPTION_SHAPE
// --
SEPARATOR ;
MenuItem
{
Identifier = FN_FORMAT_GRAFIC_DLG ;
HelpId = CMD_FN_FORMAT_GRAFIC_DLG ;
Text [ en-US ] = ~Picture... ;
};
MenuItem
{
Identifier = FN_SAVE_GRAPHIC;
Identifier = FN_SAVE_GRAPHIC ;
HelpId = CMD_FN_SAVE_GRAPHIC;
HelpID = FN_SAVE_GRAPHIC ;
Text [ de ] = ~Grafik speichern... ;
Text [ en-US ] = Save Graphics... ;
};
MenuItem
{
Identifier = FN_FORMAT_GRAFIC_DLG ;
HelpID = FN_FORMAT_GRAFIC_DLG ;
Text [ de ] = ~Grafik... ;
};
MN_FRM_CAPTION_ITEM
MenuItem
{
ITEM_EDIT_IMAP
};
};
};
--
but when I building the module, I have the following error:
[...]
Apparently I'm not inserting the code well, I need you help. how inserting
the code in the archive mn.src very well.
Yes, the merge result you pasted above doesn't look right. In the patch
the lines marked with a + at the start mean that they are to be
inserted. So the resulting code for your example would look like this:
MenuItem
{
Identifier = FN_SAVE_GRAPHIC ;
HelpID = FN_SAVE_GRAPHIC ;
Text [ de ] = ~Grafik speichern... ;
Text [ en-US ] = Save ~Graphic... ;
};
MenuItem
{
Identifier = FN_FORMAT_GRAFIC_DLG ;
HelpID = FN_FORMAT_GRAFIC_DLG ;
Text [ de ] = ~Grafik... ;
The patch with the '+' is a so called unified diff [1]. Lines with a
'+' at the start are added, lines with a '-' at the start are removed.
Once you get used to it you'll see that they are quite easy to read.
[1] http://en.wikipedia.org/wiki/Unified_diff#Unified_format
Usually one doesn't need to apply a patch by hand. There are tools for
this that make it quite easy. The most prominent one is accidentally
called patch [2]
[2] http://en.wikipedia.org/wiki/Patch_%28Unix%29
Hope this helps,
Herbert
Re: [AOO 3.4/3.4.1] crashes due to upgraded user profile from former OOo installation
On 2013/02/15 2:37 PM, Oliver-Rainer Wittmann wrote: It is the root cause for both which are more or less the same. A user profile corruption in the extension databases causes crashes when checking for extension updates. Extension updates can be triggered manually or automatically as part of the checking for an application update. I'd rephrase that: These problematic bundled extensions disappeared when the older OpenOffice was removed and the new Apache OpenOffice didn't reinstall them at the same locations as they were before. All this time the user profile stays exactly the same, it is not corrupted. The problem was that there was a bug in the OpenOffice code which wasn't prepared for the bundled extensions to disappear. Oliver's fix to catch the exceptions thrown in that case fix that problem. An example from the real world exemplifies this scenario: you get a cloak room ticket, then somebody accidentally moves your coat in the cloak room and later the attendant dies from the distress of the coat not being where he expected it to be from the ticket's text. The ticket was never corrupted at all. The fix is simply to stay cool if stuff was moved around. The hackish way to solve the problem was to remove the user profile altogether so that OpenOffice never knew there were such extensions. In the cloak room analogy above this corresponds to dropping the ticket before the easily irritable attendant ever gets a reason to get over-excited. Herbert
Re: Proposal: How we should handle committer vetos and reverts in the future
Hi Pedro, I personally cannot work like this: my work (code and time) receives more respect in other projects. I understand your frustration. Please be assured that you and your work are most highly respected. You have done so much for the project and helped to improve its quality very considerably in a lot of areas, also in areas that were otherwise regrettably neglected. As the result of this situation I am reevaluating my role in this project. I currently have no space or motivation.to do real development work. I still deeply care: This project is important and I see there is a group of awesome people doing a great job. I still can and will help: I intend to do important updates and platform fixes but I will likely be taking a more passive role from now on. Please reconsider and don't give up. The project is so extensive and you made and can make a difference for an influential product used by many many people. There is lots of room for enjoyable development work and fruitful cooperation. Herbert
Re: Tutorial About
On 2013/02/14 2:27 AM, jorge ivan poot diaz wrote: How I can do a debug mode? Where I can find more information about the debug mode? http://wiki.openoffice.org/wiki/Non_Product_Build is a good start for the debug mode of AOO. Also http://wiki.openoffice.org/wiki/Debugging is a good start for debugging AOO using gdb. Herbert
Re: Ubuntu Build Instructions
On 31.01.2013 15:55, janI wrote: but it still highlight the problem I had in the beginning...our wiki has really much valuable information, but to an extent it is shadowed by identical information which are not maintained. I searched for ubuntu build instruction way back, and as you can imagine got confused. Thx. to the brilliant help from this list I got it up and running. I hope we in the short future can get our wiki a bit streamlined (I am not thinking about removing information, but simply mark it as outdated, with a link to the newer information. I totally agree and try to do it whenever I stumble over something like this. I'd also suggest to reuse pages even if they contain obsoleted content. If their title is general purpose, is linked to from many other places and easily findable then updating them is a better solution IMHO. If anyone needs to access their older outdated content it is good to know that it is still available via Wiki's wonderful page history feature. @herbert regarding buildbot, I can see your point and agree with it. However I still think we should document exactly how our binary distributables are made. I have actually not been able to produce an exact match yet where I have tried. When people want to play with the system it is nice to have a stable start like rebuilding the release and the same result. For the exact configuration switches please see the page [1] I linked to in my previous mail. To create an exact match on Linux you'd have to install the oldest common denominator system that is used to build releases. We can ask Ariel to provide all the glorious details of this system, AFAIK they are plain RHEL5 and RHEL6, right? These old systems are great for building maximum-compatibility releases, but IMHO they are not much fun For developing. I personally love having up-to-date versions of gdb, valgrind, perf-tools, git-svn, btrfs snapshots, KVM/virtualbox, python, etc. but maybe that's just my personal disposition. [1] https://cwiki.apache.org/confluence/display/OOOUSERS/Development+Snapshot+Builds#DevelopmentSnapshotBuilds-buildflags I'm looking forward to see and talk to you on FOSDEM this weekend! Herbert
Re: [IMPORTANT, DISCUSS]: no migration/use of former user profile with AOO 4.0
On 18.01.2013 04:05, Rob Weir wrote: On Thu, Jan 17, 2013 at 2:55 AM, Oliver-Rainer Wittmann orwittm...@googlemail.com wrote: I now have an user profile which causes constantly crashes. It had been created on a Windows 7 machine where I had OOo 3.3, AOO 3.4 and now AOO 3.4.1 installed. I did not remember what I did in step #2, but it was very little - I think. I am now ready to debug the crash. It happens somehow when trying to access certain installed extensions. This is just the first glance. Thus, do not nail me down on it, when I find out that it is something else ;-) I keep you informed. [...] Do you (or anyone else) have a Bugzilla issue you want to use for this? We have quite a few reports on this issue and I'd like to start closing them out as duplicates of a root issue. If no one has a good one, I can pick one at random. I suggest to use https://issues.apache.org/ooo/show_bug.cgi?id=121625 Tracking issue for AOO 3.4 startup related crashes for this. Please beware that we are not yet really sure that all these crashes have the same root cause, so I suggest to wait with closing them as duplicates until we have a better understanding of the underlying problems. Herbert
Re: is a bug ?
Hi Galileo, On 2012/12/20 9:44 PM, you wrote: to execute command soffice --headless --acept=socket,port=8100;urp; i get fontconfig warning /usr/lib/libreoffice/share/fonts/truetype/fc_local.conf line 13: having multiple family in alias ins't supported and may not works as expected This is just a warning about unimplemented syntax in fontconfig: http://lists.freedesktop.org/archives/fontconfig/2012-February/003788.html The syntax would make sense but as the library doesn't follow it it is good that it at least emits a warning about that. Until this becomes supported we should replace the occurences, I opened an issue for that [1]. Having briefly looked into it it is unlikely that anyone is hit by the problem though. [1] https://issues.apache.org/ooo/show_bug.cgi?id=121535 Herbert
