Re: Request to Create a Branch for the Implement the Loading of TOC and Improve TOC Fidelity with MS Word Binary Document
On 6/29/12 3:59 AM, chengjh wrote: Hi, We have proposed to implement the loading of TOC and improve TOC fidelity with MS Word binary document..And now,we have finished the loading implementation part and delivered patch for review in https://issues.apache.org/ooo/show_bug.cgi?id=119963. Because this is just the stage I code implementation, and more improvements within stage II/III/ code implementation will be followed,moreover,special qe efforts are needed to cover the whole TOC function area and the impact areas,in order to decrease the negative impacts on the main stream,we request to create a branch and deliver our code implementation to the branch first, and then integrate the final qualified code to main..How about your comments?Thanks. I don't see a problem here, as you can see under http://svn.apache.org/viewvc/incubator/ooo/branches/ we have already 2 branches. One AOO34 branch to maintain our 3.4 code line and a working branch from Armin (alg/aw080) where he is working on long term refactoring stuff etc. Whereas I don't see demand for using a user specific separation of branches (eg. alg) it's fine to create a development branch for bigger changes that takes longer or which would break the master for some time. It makes of course sens to do it on a branch. Juergen Reference: [1]Candidate Proposal: https://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.0+Feature+Planning [2]Wiki with FS and SDD: http://wiki.services.**openoffice.org/wiki/Writer/ TOC http://wiki.services.openoffice.org/wiki/Writer/TOC Best Regards,Jianhong Cheng
Re: [IMPORTANT][Call for UX review] [Windows 8 certification]Test for Section 11 Apps must support multi-user sessions is not tested by Windows App Certification Kit
On 6/28/12 6:23 PM, zhangjf wrote: On Fri, Jun 29, 2012 at 12:15 AM, Jürgen Schmidt jogischm...@googlemail.com wrote: On 6/28/12 6:12 PM, Jürgen Schmidt wrote: On 6/28/12 5:33 PM, zhangjf wrote: If it still needs more time for discussion, I think it is also one option to only commit the new string change at first to catch up translation. It should have no impacts on function without committing the code. In this way, please review the new dialog and string first. Is it acceptable? sure, the way how it works is to check in the src file running localize to create a new sdf, convert it, update pootle, doing the translation on Pootle (to speed up and simplify the process) and finally merge it back in svn. I thunk the proposed solution is good and fulfill the requirements. Can we make a screenshot with the warning box and the English strings for review? Yuanlin's original first post in this mail thread contains the dialog snapshot url at https://issues.apache.org/ooo/attachment.cgi?id=78482. ok thanks, I have overseen this. I have 2 questions: 1. dialog title shows Fatal Error, is it really a Fatal Error? I don't think so, we detect a running instance and close the application or better don't continue to start. I think it's more a warning, isn't it? 2. in case of error I think we have a better error icon, in case of a warning the used icon is ok from my pov. Juergen In general I would support the proposed solution with a clear +1 to move forward immediately. Juergen Juergen zhangjf On Thu, Jun 28, 2012 at 10:48 PM, Jürgen Schmidt jogischm...@googlemail.com wrote: Hi, sorry for my top posting but I think this is very urgent and important. When we want to integrate this in 3.4.1 we have to do it immediately, means by the end of this week. The warning messages have to translated!!! Any opinions Juergen On 6/27/12 11:13 AM, Jürgen Schmidt wrote: On 6/27/12 3:23 AM, Lin Yuan wrote: Currently in AOO, only part of the data in user profile is locked and can not access by mutiple instances. So as tested on Windows Server 2008, AOO will crash in such situation. The patch is not to really support one user to launch multiple instances on mutiple sessions case. According to the suggestion in Windows 8 Certification below: *Note*: If an app does not support multiple user sessions or remote access, it must clearly state this when launched from this kind of session. With the patch, AOO will popup a warning dialog and exit in this case. So it will still not support mutiple user sessions for one user but the UX is more frendly than the current crash issue. we have to define fast if we want include it for 3.4.1 or not. It will require some translation effort that we have to organize in time (e.g. updating Pootle etc.) Juergen Thanks, Lin Yuan 2012/6/27 Joost Andrae joost.and...@gmx.de Hi Rob, but in a Windows Terminal Server session you have user profiles for each user. This patch is for if you connect with Terminal Services twice using the same user account. I just wanted to make sure that there is no real problem to get OpenOffice configured so it can be used within a multi user environment (MS TS, Citrix, Sun SGD, or UNIX profiles). If the same user connects a second time then there might be a locking problem with his profile data. If you want to fix this then it's OK but in my opinion it's not really needed because usually it should be prevented that one user accesses the same user profile from another terminal (RDP, X11) session. Just my two € Cents, Joost
Re: [IMPORTANT][Call for UX review] [Windows 8 certification]Test for Section 11 Apps must support multi-user sessions is not tested by Windows App Certification Kit
How about just simply remove - Fatal Error from the dialog title string? it won't add one more string for translation. zhangjf On Fri, Jun 29, 2012 at 2:19 PM, Jürgen Schmidt jogischm...@googlemail.com wrote: On 6/28/12 6:23 PM, zhangjf wrote: On Fri, Jun 29, 2012 at 12:15 AM, Jürgen Schmidt jogischm...@googlemail.com wrote: On 6/28/12 6:12 PM, Jürgen Schmidt wrote: On 6/28/12 5:33 PM, zhangjf wrote: If it still needs more time for discussion, I think it is also one option to only commit the new string change at first to catch up translation. It should have no impacts on function without committing the code. In this way, please review the new dialog and string first. Is it acceptable? sure, the way how it works is to check in the src file running localize to create a new sdf, convert it, update pootle, doing the translation on Pootle (to speed up and simplify the process) and finally merge it back in svn. I thunk the proposed solution is good and fulfill the requirements. Can we make a screenshot with the warning box and the English strings for review? Yuanlin's original first post in this mail thread contains the dialog snapshot url at https://issues.apache.org/ooo/attachment.cgi?id=78482. ok thanks, I have overseen this. I have 2 questions: 1. dialog title shows Fatal Error, is it really a Fatal Error? I don't think so, we detect a running instance and close the application or better don't continue to start. I think it's more a warning, isn't it? 2. in case of error I think we have a better error icon, in case of a warning the used icon is ok from my pov. Juergen In general I would support the proposed solution with a clear +1 to move forward immediately. Juergen Juergen zhangjf On Thu, Jun 28, 2012 at 10:48 PM, Jürgen Schmidt jogischm...@googlemail.com wrote: Hi, sorry for my top posting but I think this is very urgent and important. When we want to integrate this in 3.4.1 we have to do it immediately, means by the end of this week. The warning messages have to translated!!! Any opinions Juergen On 6/27/12 11:13 AM, Jürgen Schmidt wrote: On 6/27/12 3:23 AM, Lin Yuan wrote: Currently in AOO, only part of the data in user profile is locked and can not access by mutiple instances. So as tested on Windows Server 2008, AOO will crash in such situation. The patch is not to really support one user to launch multiple instances on mutiple sessions case. According to the suggestion in Windows 8 Certification below: *Note*: If an app does not support multiple user sessions or remote access, it must clearly state this when launched from this kind of session. With the patch, AOO will popup a warning dialog and exit in this case. So it will still not support mutiple user sessions for one user but the UX is more frendly than the current crash issue. we have to define fast if we want include it for 3.4.1 or not. It will require some translation effort that we have to organize in time (e.g. updating Pootle etc.) Juergen Thanks, Lin Yuan 2012/6/27 Joost Andrae joost.and...@gmx.de Hi Rob, but in a Windows Terminal Server session you have user profiles for each user. This patch is for if you connect with Terminal Services twice using the same user account. I just wanted to make sure that there is no real problem to get OpenOffice configured so it can be used within a multi user environment (MS TS, Citrix, Sun SGD, or UNIX profiles). If the same user connects a second time then there might be a locking problem with his profile data. If you want to fix this then it's OK but in my opinion it's not really needed because usually it should be prevented that one user accesses the same user profile from another terminal (RDP, X11) session. Just my two € Cents, Joost
Re: [BUILDBOT] build breakers
A week has passed and the build bots still fail with the same messages (well, thanks to Herbert the Windows7 bot shows a slightly more helpful message for the same problem.) Five minutes ago I found and fixed a typo in the area of the Linux problem. That will not fix the failing test but hopefully will show more diagnostics to better pinpoint the exact line where the test is run. -Andre On 22.06.2012 11:37, Andre Fischer wrote: Hi, there are build breakers on all three platforms - Windows7 in shell_2: has ./configure: line 7952: test: : integer expression expected ./configure: line 7957: test: : integer expression expected ./configure: line 7962: test: : integer expression expected configure: error: Compiler too old. Use Microsoft C/C++ .NET 2003/2005 compiler. could be caused by missing DLLs when the compiler is called to detect its version. In that case cl does not print anything. I have seen something similar when I tried to use the 2010 dev studio. - Linux64 in compile_2 there is the line ./configure: line 11931: test: : integer expression expected @Andrew: can you post the config.log and configure files so that we can better analyze the cause of this bug? - Linux32 the build breaker seems to be caused by my recent changes in ext_libraries/serf. It may be sufficient to make a clean build. @Andrew: can you trigger that? Regards, Andre
Re: [BUILDBOT] build breakers
I just check-out the latest revision from svn server. I will start to make a full build on Mac/Suse/Windows. I will report the details if I meet a build break. 2012/6/29 Andre Fischer a...@a-w-f.de A week has passed and the build bots still fail with the same messages (well, thanks to Herbert the Windows7 bot shows a slightly more helpful message for the same problem.) Five minutes ago I found and fixed a typo in the area of the Linux problem. That will not fix the failing test but hopefully will show more diagnostics to better pinpoint the exact line where the test is run. -Andre On 22.06.2012 11:37, Andre Fischer wrote: Hi, there are build breakers on all three platforms - Windows7 in shell_2: has ./configure: line 7952: test: : integer expression expected ./configure: line 7957: test: : integer expression expected ./configure: line 7962: test: : integer expression expected configure: error: Compiler too old. Use Microsoft C/C++ .NET 2003/2005 compiler. could be caused by missing DLLs when the compiler is called to detect its version. In that case cl does not print anything. I have seen something similar when I tried to use the 2010 dev studio. - Linux64 in compile_2 there is the line ./configure: line 11931: test: : integer expression expected @Andrew: can you post the config.log and configure files so that we can better analyze the cause of this bug? - Linux32 the build breaker seems to be caused by my recent changes in ext_libraries/serf. It may be sufficient to make a clean build. @Andrew: can you trigger that? Regards, Andre -- Best regards, Chao Huang
Re: Request to Create a Branch for the Implement the Loading of TOC and Improve TOC Fidelity with MS Word Binary Document
Hi, On 29.06.2012 03:59, chengjh wrote: Hi, We have proposed to implement the loading of TOC and improve TOC fidelity with MS Word binary document..And now,we have finished the loading implementation part and delivered patch for review in https://issues.apache.org/ooo/show_bug.cgi?id=119963. Because this is just the stage I code implementation, and more improvements within stage II/III/ code implementation will be followed,moreover,special qe efforts are needed to cover the whole TOC function area and the impact areas,in order to decrease the negative impacts on the main stream,we request to create a branch and deliver our code implementation to the branch first, and then integrate the final qualified code to main..How about your comments?Thanks. Reference: [1]Candidate Proposal: https://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.0+Feature+Planning [2]Wiki with FS and SDD: http://wiki.services.**openoffice.org/wiki/Writer/ TOC http://wiki.services.openoffice.org/wiki/Writer/TOC In general I think it makes completely sense to work on a branch for a certain feature which takes more implementation, testing etc. efforts. In this special case I am not sure, if it is needed. I am currently reviewing the patch for issue 119963. It works fine from my point of view. There are no open ends, it is complete and does not cause any problems as far as I can see. Thus, I am planning to apply this patch to trunk today or on Monday. Thus, from my point of view this feature milestone and the following ones are still small enough to handle them without an additional branch. Best regards, Oliver.
Re: Touching base with the mirrorbrain operators
Hi Drew, Back in March I sent an email to 102 contact references for that mirror network, it would be a good idea IMO to drop another note to these folks, let them know we recognize the traffic they are still carrying, say thanks and keep in touch as the projects looks to, and plans for the next release - 3.4.1 and beyond. Sound like a good idea? how about to mention their contribution within an official press announcement ? They deserve it. +1 Kind regards, Joost
Re: [IMPORTANT][Call for UX review] [Windows 8 certification]Test for Section 11 Apps must support multi-user sessions is not tested by Windows App Certification Kit
I am reviewing yuanlin's updated patch for the new dialog message only, https://issues.apache.org/ooo/attachment.cgi?id=78521action=diff. And I suppose it will remove - Fatal Error from the error dialog title string late, so this will not introduce more strings for translation. If there is no more concerns in 4 hours from now, I will commit this string patch to 3.4.1 at first. thanks, zhangjf On Fri, Jun 29, 2012 at 3:14 PM, zhangjf zhan...@apache.org wrote: How about just simply remove - Fatal Error from the dialog title string? it won't add one more string for translation. zhangjf On Fri, Jun 29, 2012 at 2:19 PM, Jürgen Schmidt jogischm...@googlemail.com wrote: On 6/28/12 6:23 PM, zhangjf wrote: On Fri, Jun 29, 2012 at 12:15 AM, Jürgen Schmidt jogischm...@googlemail.com wrote: On 6/28/12 6:12 PM, Jürgen Schmidt wrote: On 6/28/12 5:33 PM, zhangjf wrote: If it still needs more time for discussion, I think it is also one option to only commit the new string change at first to catch up translation. It should have no impacts on function without committing the code. In this way, please review the new dialog and string first. Is it acceptable? sure, the way how it works is to check in the src file running localize to create a new sdf, convert it, update pootle, doing the translation on Pootle (to speed up and simplify the process) and finally merge it back in svn. I thunk the proposed solution is good and fulfill the requirements. Can we make a screenshot with the warning box and the English strings for review? Yuanlin's original first post in this mail thread contains the dialog snapshot url at https://issues.apache.org/ooo/attachment.cgi?id=78482. ok thanks, I have overseen this. I have 2 questions: 1. dialog title shows Fatal Error, is it really a Fatal Error? I don't think so, we detect a running instance and close the application or better don't continue to start. I think it's more a warning, isn't it? 2. in case of error I think we have a better error icon, in case of a warning the used icon is ok from my pov. Juergen In general I would support the proposed solution with a clear +1 to move forward immediately. Juergen Juergen zhangjf On Thu, Jun 28, 2012 at 10:48 PM, Jürgen Schmidt jogischm...@googlemail.com wrote: Hi, sorry for my top posting but I think this is very urgent and important. When we want to integrate this in 3.4.1 we have to do it immediately, means by the end of this week. The warning messages have to translated!!! Any opinions Juergen On 6/27/12 11:13 AM, Jürgen Schmidt wrote: On 6/27/12 3:23 AM, Lin Yuan wrote: Currently in AOO, only part of the data in user profile is locked and can not access by mutiple instances. So as tested on Windows Server 2008, AOO will crash in such situation. The patch is not to really support one user to launch multiple instances on mutiple sessions case. According to the suggestion in Windows 8 Certification below: *Note*: If an app does not support multiple user sessions or remote access, it must clearly state this when launched from this kind of session. With the patch, AOO will popup a warning dialog and exit in this case. So it will still not support mutiple user sessions for one user but the UX is more frendly than the current crash issue. we have to define fast if we want include it for 3.4.1 or not. It will require some translation effort that we have to organize in time (e.g. updating Pootle etc.) Juergen Thanks, Lin Yuan 2012/6/27 Joost Andrae joost.and...@gmx.de Hi Rob, but in a Windows Terminal Server session you have user profiles for each user. This patch is for if you connect with Terminal Services twice using the same user account. I just wanted to make sure that there is no real problem to get OpenOffice configured so it can be used within a multi user environment (MS TS, Citrix, Sun SGD, or UNIX profiles). If the same user connects a second time then there might be a locking problem with his profile data. If you want to fix this then it's OK but in my opinion it's not really needed because usually it should be prevented that one user accesses the same user profile from another terminal (RDP, X11) session. Just my two € Cents, Joost
Re: [BUILDBOT] build breakers
On 29.06.2012 09:39, Chao Huang wrote: I just check-out the latest revision from svn server. I will start to make a full build on Mac/Suse/Windows. I will report the details if I meet a build break. That would be great. Thanks. -Andre 2012/6/29 Andre Fischer a...@a-w-f.de A week has passed and the build bots still fail with the same messages (well, thanks to Herbert the Windows7 bot shows a slightly more helpful message for the same problem.) Five minutes ago I found and fixed a typo in the area of the Linux problem. That will not fix the failing test but hopefully will show more diagnostics to better pinpoint the exact line where the test is run. -Andre On 22.06.2012 11:37, Andre Fischer wrote: Hi, there are build breakers on all three platforms - Windows7 in shell_2: has ./configure: line 7952: test: : integer expression expected ./configure: line 7957: test: : integer expression expected ./configure: line 7962: test: : integer expression expected configure: error: Compiler too old. Use Microsoft C/C++ .NET 2003/2005 compiler. could be caused by missing DLLs when the compiler is called to detect its version. In that case cl does not print anything. I have seen something similar when I tried to use the 2010 dev studio. - Linux64 in compile_2 there is the line ./configure: line 11931: test: : integer expression expected @Andrew: can you post the config.log and configure files so that we can better analyze the cause of this bug? - Linux32 the build breaker seems to be caused by my recent changes in ext_libraries/serf. It may be sufficient to make a clean build. @Andrew: can you trigger that? Regards, Andre
Re: Request to Create a Branch for the Implement the Loading of TOC and Improve TOC Fidelity with MS Word Binary Document
Thanks to Juergen and Oliver's comments..I got points as followed: a)If the code changes of an improvement are too many,effort is also big and development has to be went on for long,moreover,special testing work will be taken to cover the given function areas and impacted areas, that's ok to create a branch for the improvement..It is better to do so. b)If the risk and impact areas are under control,and we are confident to ensure the quality along with that the finished scenarios are clear and expected, it is better for us to deliver the code changes to main stream( trunk ) directly even the deliverable is not complete.Thus QE volunteers can help us to find out regression defects as early as possible,and also,better feedback can be got in time. So,the selection can be decided according to the actual project's situation and evaluation.To the TOC Loading,we are able to deliver the code changes to main and finish the development work stage by stage in main directly..Any misunderstanding,please correct me. On Fri, Jun 29, 2012 at 3:49 PM, Oliver-Rainer Wittmann orwittm...@googlemail.com wrote: Hi, On 29.06.2012 03:59, chengjh wrote: Hi, We have proposed to implement the loading of TOC and improve TOC fidelity with MS Word binary document..And now,we have finished the loading implementation part and delivered patch for review in https://issues.apache.org/ooo/**show_bug.cgi?id=119963https://issues.apache.org/ooo/show_bug.cgi?id=119963. Because this is just the stage I code implementation, and more improvements within stage II/III/ code implementation will be followed,moreover,special qe efforts are needed to cover the whole TOC function area and the impact areas,in order to decrease the negative impacts on the main stream,we request to create a branch and deliver our code implementation to the branch first, and then integrate the final qualified code to main..How about your comments?Thanks. Reference: [1]Candidate Proposal: https://cwiki.apache.org/**confluence/display/OOOUSERS/** AOO+4.0+Feature+Planninghttps://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.0+Feature+Planning [2]Wiki with FS and SDD: http://wiki.services.**openoff** ice.org/wiki/Writer/ http://openoffice.org/wiki/Writer/ TOC http://wiki.services.**openoffice.org/wiki/Writer/TOChttp://wiki.services.openoffice.org/wiki/Writer/TOC ** In general I think it makes completely sense to work on a branch for a certain feature which takes more implementation, testing etc. efforts. In this special case I am not sure, if it is needed. I am currently reviewing the patch for issue 119963. It works fine from my point of view. There are no open ends, it is complete and does not cause any problems as far as I can see. Thus, I am planning to apply this patch to trunk today or on Monday. Thus, from my point of view this feature milestone and the following ones are still small enough to handle them without an additional branch. Best regards, Oliver. -- Best Regards,Jianhong Cheng
Re: [BUILDBOT] build breakers
I don't know here which distribution linux32 refers to. Our build on SuSe 11.1 breaks on serf 1.1 because it has a new dependency on libexpat.so, we resolved it by installing the libexpat-devel rpm package. zhangjf On Fri, Jun 22, 2012 at 5:37 PM, Andre Fischer a...@a-w-f.de wrote: Hi, there are build breakers on all three platforms - Windows7 in shell_2: has ./configure: line 7952: test: : integer expression expected ./configure: line 7957: test: : integer expression expected ./configure: line 7962: test: : integer expression expected configure: error: Compiler too old. Use Microsoft C/C++ .NET 2003/2005 compiler. could be caused by missing DLLs when the compiler is called to detect its version. In that case cl does not print anything. I have seen something similar when I tried to use the 2010 dev studio. - Linux64 in compile_2 there is the line ./configure: line 11931: test: : integer expression expected @Andrew: can you post the config.log and configure files so that we can better analyze the cause of this bug? - Linux32 the build breaker seems to be caused by my recent changes in ext_libraries/serf. It may be sufficient to make a clean build. @Andrew: can you trigger that? Regards, Andre
Re: [BUILDBOT] build breakers
On 29.06.2012 10:42, zhangjf wrote: I don't know here which distribution linux32 refers to. Our build on SuSe 11.1 breaks on serf 1.1 because it has a new dependency on libexpat.so, we resolved it by installing the libexpat-devel rpm package. That should only be necessary if you are using --with-system-serf. Otherwise the external version should be used. I probably broke that when I added the --with-system-serf switch. -Andre zhangjf On Fri, Jun 22, 2012 at 5:37 PM, Andre Fischer a...@a-w-f.de wrote: Hi, there are build breakers on all three platforms - Windows7 in shell_2: has ./configure: line 7952: test: : integer expression expected ./configure: line 7957: test: : integer expression expected ./configure: line 7962: test: : integer expression expected configure: error: Compiler too old. Use Microsoft C/C++ .NET 2003/2005 compiler. could be caused by missing DLLs when the compiler is called to detect its version. In that case cl does not print anything. I have seen something similar when I tried to use the 2010 dev studio. - Linux64 in compile_2 there is the line ./configure: line 11931: test: : integer expression expected @Andrew: can you post the config.log and configure files so that we can better analyze the cause of this bug? - Linux32 the build breaker seems to be caused by my recent changes in ext_libraries/serf. It may be sufficient to make a clean build. @Andrew: can you trigger that? Regards, Andre
Re: [BUILDBOT] build breakers
Hi Andre, On Fri, Jun 29, 2012 at 5:50 AM, Andre Fischer a...@a-w-f.de wrote: On 29.06.2012 10:42, zhangjf wrote: I don't know here which distribution linux32 refers to. Our build on SuSe 11.1 breaks on serf 1.1 because it has a new dependency on libexpat.so, we resolved it by installing the libexpat-devel rpm package. That should only be necessary if you are using --with-system-serf. Otherwise the external version should be used. I probably broke that when I added the --with-system-serf switch. Revision 1351981 broke rev. 1341713, now no patches are applied at all: make writeable... no patch needed... touch ./unxlngx6/misc/build/so_patched_so_serf mkdir ./unxlngx6/misc/build/serf-1.1.0/ Regards
Re: [BUILDBOT] build breakers
Except that, we don't have more problems. Actual we already had success builds (r1355082) on Windows XP (vs2008 pro w/ cygwin), Suse 11.1, Ubuntu 11.04, MacOS 10.6 today. From the above linux 64 and Win7 error message, I am wondering if it needs a fresh autoconf. We saw similar error messages in an updated Mac source build environment before, but I am not sure. zhangjf On Fri, Jun 29, 2012 at 4:50 PM, Andre Fischer a...@a-w-f.de wrote: On 29.06.2012 10:42, zhangjf wrote: I don't know here which distribution linux32 refers to. Our build on SuSe 11.1 breaks on serf 1.1 because it has a new dependency on libexpat.so, we resolved it by installing the libexpat-devel rpm package. That should only be necessary if you are using --with-system-serf. Otherwise the external version should be used. I probably broke that when I added the --with-system-serf switch. -Andre zhangjf On Fri, Jun 22, 2012 at 5:37 PM, Andre Fischer a...@a-w-f.de wrote: Hi, there are build breakers on all three platforms - Windows7 in shell_2: has ./configure: line 7952: test: : integer expression expected ./configure: line 7957: test: : integer expression expected ./configure: line 7962: test: : integer expression expected configure: error: Compiler too old. Use Microsoft C/C++ .NET 2003/2005 compiler. could be caused by missing DLLs when the compiler is called to detect its version. In that case cl does not print anything. I have seen something similar when I tried to use the 2010 dev studio. - Linux64 in compile_2 there is the line ./configure: line 11931: test: : integer expression expected @Andrew: can you post the config.log and configure files so that we can better analyze the cause of this bug? - Linux32 the build breaker seems to be caused by my recent changes in ext_libraries/serf. It may be sufficient to make a clean build. @Andrew: can you trigger that? Regards, Andre
Re: [BUILDBOT] build breakers
The full build broke on Mac OS X 10.6.8. MacOSX 10.6.8 gcc version 4.2.1 (Apple Inc. build 5666) (dot 3) java version 1.6.0_33 configure switchers ./configure --with-dmake-url= http://dmake.apache-extras.org.codespot.com/files/dmake-4.12.tar.bz2; --with-epm-url=http://ftp.easysw.com/pub/epm/3.7/epm-3.7-source.tar.gz; --disable-mozilla --disable-build-mozilla --enable-category-b --enable-minimizer --enable-presenter-console --enable-wiki-publisher Break Info ignoring silgraphite-2.3.1.tar.gz because its prerequisites are not fulfilled bsh-2.0b1-src.tar.gz exists, MD5 checksum is OK downloading 1 missing tar balls to /Users/hchao/Apache/build/aoo.340.20120629/code/ext_sources downloading to /Users/hchao/Apache/build/aoo.340.20120629/code/ext_sources/377a60170e5185eb63d3ed2fae98e621-README_silgraphite-2.3.1.txt.part Can't locate object method show_progress via package LWP::UserAgent at /Users/hchao/Apache/build/aoo.340.20120629/code/main/solenv/bin/ download_external_dependencies.pl line 506, $in line 353. dmake present in /Users/hchao/Apache/build/aoo.340.20120629/code/main/solenv/ unxmacxi.pro/bin/dmake bundling of dictionaries is disabled. HuangChaos-iMac-106:main hchao$ 2012/6/22 Andre Fischer a...@a-w-f.de Hi, there are build breakers on all three platforms - Windows7 in shell_2: has ./configure: line 7952: test: : integer expression expected ./configure: line 7957: test: : integer expression expected ./configure: line 7962: test: : integer expression expected configure: error: Compiler too old. Use Microsoft C/C++ .NET 2003/2005 compiler. could be caused by missing DLLs when the compiler is called to detect its version. In that case cl does not print anything. I have seen something similar when I tried to use the 2010 dev studio. - Linux64 in compile_2 there is the line ./configure: line 11931: test: : integer expression expected @Andrew: can you post the config.log and configure files so that we can better analyze the cause of this bug? - Linux32 the build breaker seems to be caused by my recent changes in ext_libraries/serf. It may be sufficient to make a clean build. @Andrew: can you trigger that? Regards, Andre -- Best regards, Chao Huang
Re: [BUILDBOT] build breakers
On Fri, Jun 29, 2012 at 5:03 PM, Ariel Constenla-Haile arie...@apache.org wrote: Hi Andre, On Fri, Jun 29, 2012 at 5:50 AM, Andre Fischer a...@a-w-f.de wrote: On 29.06.2012 10:42, zhangjf wrote: I don't know here which distribution linux32 refers to. Our build on SuSe 11.1 breaks on serf 1.1 because it has a new dependency on libexpat.so, we resolved it by installing the libexpat-devel rpm package. That should only be necessary if you are using --with-system-serf. Otherwise the external version should be used. I probably broke that when I added the --with-system-serf switch. Revision 1351981 broke rev. 1341713, now no patches are applied at all: make writeable... no patch needed... touch ./unxlngx6/misc/build/so_patched_so_serf mkdir ./unxlngx6/misc/build/serf-1.1.0/ Regards Thanks for the update. So it will be fixed, right? zhangjf
Re: [BUILDBOT] build breakers
Hi Chao Huang, On Fri, Jun 29, 2012 at 6:07 AM, Chao Huang chao.de...@gmail.com wrote: The full build broke on Mac OS X 10.6.8. MacOSX 10.6.8 gcc version 4.2.1 (Apple Inc. build 5666) (dot 3) java version 1.6.0_33 configure switchers ./configure --with-dmake-url= http://dmake.apache-extras.org.codespot.com/files/dmake-4.12.tar.bz2; --with-epm-url=http://ftp.easysw.com/pub/epm/3.7/epm-3.7-source.tar.gz; --disable-mozilla --disable-build-mozilla --enable-category-b --enable-minimizer --enable-presenter-console --enable-wiki-publisher Break Info ignoring silgraphite-2.3.1.tar.gz because its prerequisites are not fulfilled bsh-2.0b1-src.tar.gz exists, MD5 checksum is OK downloading 1 missing tar balls to /Users/hchao/Apache/build/aoo.340.20120629/code/ext_sources downloading to /Users/hchao/Apache/build/aoo.340.20120629/code/ext_sources/377a60170e5185eb63d3ed2fae98e621-README_silgraphite-2.3.1.txt.part Can't locate object method show_progress via package LWP::UserAgent at Your Perl is too old, I've seen this too on CentOS 5: Can't locate object method show_progress via package LWP::UserAgent at /build/aoo/src/clean/trunk/main/solenv/bin/modules/ExtensionsLst.pm line 484. CentOS Perl: ]$ /usr/bin/perl --version This is perl, v5.8.8 built for i386-linux-thread-multi CentOS perl-libwww-perl-5.805-1.1.1.noarch ]$ /usr/bin/perl -MLWP::UserAgent -e 'print $LWP::UserAgent::VERSION\n' 2.033 As you see, the UserAgent version is too old, and does contain that method show_progress On Fedora 17: ]$ /usr/bin/perl -MLWP::UserAgent -e 'print $LWP::UserAgent::VERSION\n' 6.03 It looks like we need to ckeck not only Perl version but also LWP::UserAgent::VERSION You can fix this by not using system perl (I've done this on CentOS, but needs more tweaking). Regards
Re: [BUILDBOT] build breakers
On 29.06.2012 11:07, Chao Huang wrote: /Users/hchao/Apache/build/aoo.340.20120629/code/ext_sources/377a60170e5185eb63d3ed2fae98e621-README_silgraphite-2.3.1.txt.part Can't locate object method show_progress via package LWP::UserAgent at This sounds like the problem mentioned in [1] and in https://groups.google.com/forum/#!topic/pulledpork-users/Gq6MT27wp6w Maybe running perl -MCPAN -e upgrade LWP::UserAgent helps. Herbert [1] http://markmail.org/message/mnqv3ncast7754zw
Re: [BUILDBOT] build breakers
Hi, 2012/6/29 Ariel Constenla-Haile arie...@apache.org: Hi Andre, On Fri, Jun 29, 2012 at 5:50 AM, Andre Fischer a...@a-w-f.de wrote: On 29.06.2012 10:42, zhangjf wrote: I don't know here which distribution linux32 refers to. Our build on SuSe 11.1 breaks on serf 1.1 because it has a new dependency on libexpat.so, we resolved it by installing the libexpat-devel rpm package. That should only be necessary if you are using --with-system-serf. Otherwise the external version should be used. I probably broke that when I added the --with-system-serf switch. Revision 1351981 broke rev. 1341713, now no patches are applied at all: make writeable... no patch needed... touch ./unxlngx6/misc/build/so_patched_so_serf mkdir ./unxlngx6/misc/build/serf-1.1.0/ I needed to enable sef-1.0.0.ldl.patch applied on Linux x86. Regards Regards
Re: [BUILDBOT] build breakers
hi, Ariel Thanks for your reminder. I will update my local dev env, which is OK to build R1343441 on Windows/Mac/Ubuntu/Suse. 2012/6/29 Ariel Constenla-Haile arie...@apache.org Hi Chao Huang, On Fri, Jun 29, 2012 at 6:07 AM, Chao Huang chao.de...@gmail.com wrote: The full build broke on Mac OS X 10.6.8. MacOSX 10.6.8 gcc version 4.2.1 (Apple Inc. build 5666) (dot 3) java version 1.6.0_33 configure switchers ./configure --with-dmake-url= http://dmake.apache-extras.org.codespot.com/files/dmake-4.12.tar.bz2; --with-epm-url=http://ftp.easysw.com/pub/epm/3.7/epm-3.7-source.tar.gz; --disable-mozilla --disable-build-mozilla --enable-category-b --enable-minimizer --enable-presenter-console --enable-wiki-publisher Break Info ignoring silgraphite-2.3.1.tar.gz because its prerequisites are not fulfilled bsh-2.0b1-src.tar.gz exists, MD5 checksum is OK downloading 1 missing tar balls to /Users/hchao/Apache/build/aoo.340.20120629/code/ext_sources downloading to /Users/hchao/Apache/build/aoo.340.20120629/code/ext_sources/377a60170e5185eb63d3ed2fae98e621-README_silgraphite-2.3.1.txt.part Can't locate object method show_progress via package LWP::UserAgent at Your Perl is too old, I've seen this too on CentOS 5: Can't locate object method show_progress via package LWP::UserAgent at /build/aoo/src/clean/trunk/main/solenv/bin/modules/ExtensionsLst.pm line 484. CentOS Perl: ]$ /usr/bin/perl --version This is perl, v5.8.8 built for i386-linux-thread-multi CentOS perl-libwww-perl-5.805-1.1.1.noarch ]$ /usr/bin/perl -MLWP::UserAgent -e 'print $LWP::UserAgent::VERSION\n' 2.033 As you see, the UserAgent version is too old, and does contain that method show_progress On Fedora 17: ]$ /usr/bin/perl -MLWP::UserAgent -e 'print $LWP::UserAgent::VERSION\n' 6.03 It looks like we need to ckeck not only Perl version but also LWP::UserAgent::VERSION You can fix this by not using system perl (I've done this on CentOS, but needs more tweaking). Regards -- Best regards, Chao Huang
Re: Touching base with the mirrorbrain operators
On Thu, Jun 28, 2012 at 11:40 PM, drew jensen drewjensen.in...@gmail.com wrote: Howdy, It is wonderful to see the millions of downloads for the current 3.4 release, utilizing the sf.net resources. Also in this case one picture worths 1000 words, Rob keeps doing a fantastic job using SourceForge's APIs to show the AOO 3.4 total so far http://people.apache.org/~robweir/aoo-downloads.html By the way we're about to provide an easier way to detect when a day is over, so that is clear if queries for a given day are complete or partial. It's also encouraging to see the half million downloads per month still going out from the mirrorbrain servers - this 500,000 p/mos figure is comprised of older releases, many of which include language packs not yet available for 3.4 or the upcoming 3.4.1 - so this is a valuable resource to the project still, IMO. It's good to see we find a way to use both mirror networks. Beyond the big numbers at stake, I think it we were right by setting a clear rule about who serves which downloads (AOO 3.4, legacy). Doing so actually we eliminated all potential problems with which mirror network is having a problem kinds of debugging. Talking about which, let me state that so far we didn't experience any issue, people reported an average time of downloading of about 2 mins that's pretty good. Back in March I sent an email to 102 contact references for that mirror network, it would be a good idea IMO to drop another note to these folks, let them know we recognize the traffic they are still carrying, say thanks and keep in touch as the projects looks to, and plans for the next release - 3.4.1 and beyond. Sound like a good idea? Of course! Roberto Thanks //drew -- This e- mail message is intended only for the named recipient(s) above. It may contain confidential and privileged information. If you are not the intended recipient you are hereby notified that any dissemination, distribution or copying of this e-mail and any attachment(s) is strictly prohibited. If you have received this e-mail in error, please immediately notify the sender by replying to this e-mail and delete the message and any attachment(s) from your system. Thank you.
Re: [BUILDBOT] build breakers
Hi Tsutomu, On Fri, Jun 29, 2012 at 6:18 AM, Tsutomu Uchino hanya.r...@gmail.com wrote: Revision 1351981 broke rev. 1341713, now no patches are applied at all: make writeable... no patch needed... touch ./unxlngx6/misc/build/so_patched_so_serf mkdir ./unxlngx6/misc/build/serf-1.1.0/ I needed to enable sef-1.0.0.ldl.patch applied on Linux x86. yes, both patches are needed and were commented. .IF $(OS)==MACOSX || $(OS)==FREEBSD || $(OS)==LINUX # Do not link against expat. It is not necessary (apr-util is already linked against it) # and does not work (we use a different expat library schema.) #PATCH_FILES+=$(TARFILE_NAME).mac.patch .ENDIF .IF $(OS)==LINUX # Add -ldl as last library so that the linker has no trouble resolving dependencies. #PATCH_FILES+=$(TARFILE_NAME).ldl.patch .ENDIF Regards
Re: Touching base with the mirrorbrain operators
On Fri, Jun 29, 2012 at 6:22 AM, Roberto Galoppini rgalopp...@geek.net wrote: It's good to see we find a way to use both mirror networks. Beyond the big numbers at stake, I think it we were right by setting a clear rule about who serves which downloads (AOO 3.4, legacy). MirrorBrain has torrent support http://mirrorbrain.org/docs/configuration/#configuring-torrent-generation something users have been asking and SF lacks of. drew wrote: Back in March I sent an email to 102 contact references for that mirror network, it would be a good idea IMO to drop another note to these folks, let them know we recognize the traffic they are still carrying, say thanks and keep in touch as the projects looks to, and plans for the next release - 3.4.1 and beyond. What does it mean keep in touch? It looks like the current instance is running on an Oracle server at Santa Clara, CA, USA: download.services.openoffice.org (192.9.173.38) Are going to have an own mirrorbrain instance here with Apache infrastructure? Otherwise it's like keeping the mirrors net without the underlying technology that serves them. Regards
RE: Propose for 3.4.1: Can't remove password from file (119366)
Test again on trunk r1355082 which include modules nss and moz. From the result I think we can close this bug. Test with 3 scenario. A. Saved with password, then remove password with same revision B. Saved with password with OOo3.3, then remove password with r1355082 C. Saved with password with MS 2003, then remove password with r1355082 Result: Scenario A: ods/xls/odt/odp work fine doc can saved with password successfully, but when reopen the saved file, it is under read-only mode. This also repro on OOo 3.3 ppt is weird, in OOo 3.3, In saved dialog, the password check box can check, input password and saved the file then reopen it, No need password, so saved with password in OOo3.3 for ppt format is not work. In trunk r1355082, In saved dialog, saveing with password checkbox is disabled If we saved ppt file with password protect by OpenOffice, then open in MS Office, no need input password, so it is totally not work. Scenario B: ods/xls/odt/odp work fine doc file with password open under read-only mode, we can saved to another file to remove password. ppt file can't saved with password protect Scenario C: xls work fine doc file with password open under read-only mode, we can saved to another file to remove password. ppt file with password protect by MS office can't open in Open Office, it said Read Error. the loading of password-encrypted Microsoft PowerPint presentations is not supported. For PPT issue, there have 2 bugs about saved(Bug 39527) and loading(Bug 46307) For doc open with read-only mode issue, i think it is a know issue, but i can't find the bug in bugzilla, I will double check in bugzilla. Date: Thu, 28 Jun 2012 11:20:54 +0200 From: orwittm...@googlemail.com To: ooo-dev@incubator.apache.org Subject: Re: Propose for 3.4.1: Can't remove password from file (119366) Hi, On 28.06.2012 11:08, Oliver-Rainer Wittmann wrote: Hi, sorry for top posting - it is reply to the complete discussion regarding save with password This functionality for ODF documents is available when the build includes modules nss and moz. I am not sure, if our buildbots include these modules. I will check it. win7 buildbot includes nss and moz linux32 buildbot includes nss and moz linux64 buildbot includes nss and moz But, as you may have notice since a couple of days our buildbots have problems to create new packages. last sucessful build for win7 was on 2012-06-19 last sucessful build for linux32 was on 2012-06-20 last sucessful build for linxu64 was on 2012-06-17 none of these contains the fix for issue 119366 Best regards, Oliver. Best regards, Oliver. On 28.06.2012 10:21, YangTerry wrote: Just confirm with our build owner, moz package is not in the build. Will verify and update the result after download the trunk build from BuildBot. Thanks Fan Zheng help (*^__^*) From: polo8...@hotmail.com To: ooo-dev@incubator.apache.org Subject: RE: Propose for 3.4.1: Can't remove password from file (119366) Date: Thu, 28 Jun 2012 16:02:31 +0800 Will confirm with our local build owner. Also download the trunk build from http://ci.apache.org/projects/openoffice/ Thanks for your investgate. Date: Thu, 28 Jun 2012 15:19:39 +0800 Subject: Re: Propose for 3.4.1: Can't remove password from file (119366) From: zheng.easy...@gmail.com To: ooo-dev@incubator.apache.org Hold a second, you guys are using the download build on verifying, right? If so, that means some problems there. Maybe the release build env broken? 2012/6/28 Fan Zheng zheng.easy...@gmail.com Muhaha, lucky for me that I have 3 build enviroment on verification this, all of them are under Windows XP. The difference are: In build env A, there is no moz package involved; In build env B, there is moz package involved, but without my solution of issue 119366; In build env C, there is moz package involved and with my solution of issue 119366; And I did the following test cases: In env A: 1.1 Save into ODT, without password, passed; 1.2 Save into ODT, with password, failed, with error message General Error: Generral input/output error; 1.3 Save into DOC, without password, passed; 1.4 Save into DOC, with password, passed; But in continual case: Save As without password, failed; 2.1 Save into ODP, without password, passed; 2.2 Save into ODP, with password, failed, with error message General Error: Generral input/output error; 2.3 Save into PPT, without password, passed; 2.4 Save into PPT, the saveing with password checkbox is disabled... Weird. 3.1 Save into ODS, without password, passed; 3.2 Save into ODS, with password, failed, with error message General Error: Generral input/output error; 3.3 Save into XLS, without password, passed; 3.4 Save into XLS, with password, passed; But in continual case: Save As without password, failed; In env B: 1.1 Save into ODT,
Re: [BUILDBOT] build breakers
On 29.06.2012 11:08, zhangjf wrote: On Fri, Jun 29, 2012 at 5:03 PM, Ariel Constenla-Haile arie...@apache.org wrote: Hi Andre, On Fri, Jun 29, 2012 at 5:50 AM, Andre Fischer a...@a-w-f.de wrote: On 29.06.2012 10:42, zhangjf wrote: I don't know here which distribution linux32 refers to. Our build on SuSe 11.1 breaks on serf 1.1 because it has a new dependency on libexpat.so, we resolved it by installing the libexpat-devel rpm package. That should only be necessary if you are using --with-system-serf. Otherwise the external version should be used. I probably broke that when I added the --with-system-serf switch. Revision 1351981 broke rev. 1341713, now no patches are applied at all: make writeable... no patch needed... touch ./unxlngx6/misc/build/so_patched_so_serf mkdir ./unxlngx6/misc/build/serf-1.1.0/ Regards Thanks for the update. So it will be fixed, right? I am working on it. I am still figuring out whether the serf patches are still really needed. -Andre zhangjf
Re: Request to Create a Branch for the Implement the Loading of TOC and Improve TOC Fidelity with MS Word Binary Document
Hi, On 29.06.2012 10:37, chengjh wrote: Thanks to Juergen and Oliver's comments..I got points as followed: a)If the code changes of an improvement are too many,effort is also big and development has to be went on for long,moreover,special testing work will be taken to cover the given function areas and impacted areas, that's ok to create a branch for the improvement..It is better to do so. b)If the risk and impact areas are under control,and we are confident to ensure the quality along with that the finished scenarios are clear and expected, it is better for us to deliver the code changes to main stream( trunk ) directly even the deliverable is not complete.Thus QE volunteers can help us to find out regression defects as early as possible,and also,better feedback can be got in time. So,the selection can be decided according to the actual project's situation and evaluation.To the TOC Loading,we are able to deliver the code changes to main and finish the development work stage by stage in main directly..Any misunderstanding,please correct me. Yes. The team that is working on a feature can decide what is better - working on trunk or working on a branch. May be we can use the TOC enhancements as an opportunity to figure out a good and accepted way to work on branches. Thus, I am fine with both ways for the TOC enhancements. Best regards, Oliver. P.S.: I have just finished my review on the recent patch for issue 119963 On Fri, Jun 29, 2012 at 3:49 PM, Oliver-Rainer Wittmann orwittm...@googlemail.com wrote: Hi, On 29.06.2012 03:59, chengjh wrote: Hi, We have proposed to implement the loading of TOC and improve TOC fidelity with MS Word binary document..And now,we have finished the loading implementation part and delivered patch for review in https://issues.apache.org/ooo/**show_bug.cgi?id=119963https://issues.apache.org/ooo/show_bug.cgi?id=119963. Because this is just the stage I code implementation, and more improvements within stage II/III/ code implementation will be followed,moreover,special qe efforts are needed to cover the whole TOC function area and the impact areas,in order to decrease the negative impacts on the main stream,we request to create a branch and deliver our code implementation to the branch first, and then integrate the final qualified code to main..How about your comments?Thanks. Reference: [1]Candidate Proposal: https://cwiki.apache.org/**confluence/display/OOOUSERS/** AOO+4.0+Feature+Planninghttps://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.0+Feature+Planning [2]Wiki with FS and SDD: http://wiki.services.**openoff** ice.org/wiki/Writer/ http://openoffice.org/wiki/Writer/ TOC http://wiki.services.**openoffice.org/wiki/Writer/TOChttp://wiki.services.openoffice.org/wiki/Writer/TOC ** In general I think it makes completely sense to work on a branch for a certain feature which takes more implementation, testing etc. efforts. In this special case I am not sure, if it is needed. I am currently reviewing the patch for issue 119963. It works fine from my point of view. There are no open ends, it is complete and does not cause any problems as far as I can see. Thus, I am planning to apply this patch to trunk today or on Monday. Thus, from my point of view this feature milestone and the following ones are still small enough to handle them without an additional branch. Best regards, Oliver.
Re: RE: Open the UOF2.0 source codes
On Thu, Jun 28, 2012 at 7:49 PM, hongyun.an hongyun...@cs2c.com.cn wrote: Dear everyone: Thanks to all your help,so far I have finished merging the UOF v2.0 source code to the Aoo3.4(svn version1354914) both on the Windows and Linux. This is great news! Thank you Hongyun. (Forgive the delay, I am a little slow catching up on emails due to travels) Today, I delivered a presentation on Apache OpenOffice here in Beijing at the Open Source China Open Source World Summit conference where I made a special point to recognize the achievements of CS2C and your work. I was fortunate to meet your colleagues Michael Lee and Shi Zhoubo. We discussed a variety of topics concerning the project and how we can improve participation and awareness of Apache OpenOffice in China. I hope that Michael introduces himself soon on this list. The source code is 33 files and about 2M,mainly in filter/source/config,filter/source/xmlfilterdetect, filter/source/xsltfilter,scp2/source/ooo. Now it can work very well in sw,sc,sd. It could load as the uof v2.0 format files(uot,uos,uop),also could open them. thank you,Dennis. But I have forgotten my password. I go to the http://id.apache.org and enter my ID number,then click the Forgot your password. In the new window I enter my ID in the frame which is like this Apache User ID:___...@apache.org, With that I get the message Email sent,Message-ID= 20120628100728.462c...@id.apache.org. I don't know whether I have the apache.org email and I could not log in this email. Could you tell me how to reset my password? Another question: could I commit the source code to the Aoo3.4 trunk? best wishes! -- hongyun.an @hongyun.an Yes, you are Apache committer hyan@ a.o. You may be required to reset your password in order to make commits to SVN. There was a general requirement to change passwords as the result of a recent security issue. Simply go to http://id.apache.org and follow the change-password procedure. You may want to update your e-mail forwarding for hyan@ at the same time if you have not already done so. +1 for how you propose to account for the approach and then carry it out. - Dennis PS: Ensure that the SVN-client configuration options specify UTF-8 encoding on the committed files, assuming that is needed for the Chinese-language characters included in XML NCNames and their counterparts in the UOF source code. There may also need to be some experimentation to ensure that any use of Chinese-language characters in source code work through all of the build cases. I recommend subversion configuration file entries along the following lines: *.html = svn:mime-type=text/html;;charset=UTF-8 *.htm = svn:mime-type=text/html;;charset=UTF-8 *.xml = svn:mime-type=application/xml;;charset=UTF-8 *.c = svn:mime-type=text/plain;;charset=UTF-8;svn:eol-style=native *.cpp = svn:mime-type=text/plain;;charset=UTF-8;svn:eol-style=native (I have not tried the combination of svn:mime-type and svn:eol-style. Experiment is required.) -Original Message- From: hongyun.an [mailto:hongyun...@cs2c.com.cn] Sent: Monday, June 11, 2012 05:38 To: ooo-dev Subject: Re: Re: Open the UOF2.0 source codes The size of the source code is about 2M,30 files. As far as I know, I have the the commit right in incubator ooo Project. According to your suggestions,I think we should do two things firstly: 1:Check whether the code could work in Aoo3.4 based on the Windows and Linux. 2:Introduce the work we have done on wiki http://wiki.services.openoffice.org/wiki/Documentation. As soon as I complete these works ,I will message to the mail list. Thank you. best wishes. -- hongyun.an On Mon, Jun 11, 2012 at 1:44 AM, Pedro Giffuni p...@apache.org wrote: --- Lun 11/6/12, hongyun.an hongyun...@cs2c.com.cn ha scritto: ... Data: Lunedì 11 giugno 2012, 00:20 You are welcome.We hope it will be more compatible between the ODF and UOF2.0. Could anybody tell me how submit the code? -- I guess it depends on the size. One option would be to put up a tarball somewhere, but perhaps its better to create a branch in SVN. The UOF support is in 30 files, mainly in filter/source/config, filter/source/xmlfilterdetect, filter/source/xsltfilter If a committer can verify that it works, it should be small enough to check into the trunk. Some test documents would help as well. -Rob It would be great if one of our Chinese commiters takes the lead on this. Pedro. hongyun.an This is excellent news! Thank you Hongyun An, we are delighted to have this huge development for/from the Chinese community! Pedro. --- Dom 10/6/12, hongyun.an hongyun...@cs2c.com.cn ha scritto: Hi,everyone: My name is Hongyun An and I am serving in CS2C.Since 2008,I have worked for the document format interoperability. Now I am serving for
Re: [RELEASE][3.4.1]:propose bug 120096 (wipe up and wipe down effect works fail in Aoo 3.4,but ok in OO3.3) as release blocker
On 6/26/12 11:17 AM, YangTerry wrote: Change the mail title Regression issue, and the slide transition is basic transition. +1 +1 because of regression. Do we have somebody working on it already? Can we assign issues when somebody takes care of an issue, that makes the tracking easier. Juergen Date: Tue, 26 Jun 2012 17:07:34 +0800 Subject: Re: [regression]the three basic slide transition effect:wipe up,wipe down and box out lost From: bjduj...@gmail.com To: ooo-dev@incubator.apache.org sorry,in OO3.3,it is ok On Tue, Jun 26, 2012 at 5:07 PM, Du Jing bjduj...@gmail.com wrote: the three slide transition effect can't preview,and set on slide,no any play effect.but in OO3.4,it is ok. suggest fix it as soon as possible. I have submitted a bug in bugzilla to track the defect: https://issues.apache.org/ooo/show_bug.cgi?id=120096
Re: Touching base with the mirrorbrain operators
Hi, It looks like the current instance is running on an Oracle server at Santa Clara, CA, USA: download.services.openoffice.org (192.9.173.38) Are going to have an own mirrorbrain instance here with Apache infrastructure? Otherwise it's like keeping the mirrors net without the underlying technology that serves them. Peter Poeml is the author of MirrorBrain. He's been very supportive to implement the instance running at download.services and he provided his own instance in case download.services was down for whatever reason. MirrorBrain doesn't need much compute power to serve this amazingly reliable service. If you're interested into this piece of software then have a look at http://www.mirrorbrain.org/ btw. Peter is subscribed to this mailing list or at least he reads this list. His last posting is from mid of may. Kind regards, Joost
Re: Propose for 3.4.1: Can't remove password from file (119366)
Good news and thanks Terry. So the bug would be closed automatically or, need I do anything further? 2012/6/29 YangTerry polo8...@hotmail.com Test again on trunk r1355082 which include modules nss and moz. From the result I think we can close this bug. Test with 3 scenario. A. Saved with password, then remove password with same revision B. Saved with password with OOo3.3, then remove password with r1355082 C. Saved with password with MS 2003, then remove password with r1355082 Result: Scenario A: ods/xls/odt/odp work fine doc can saved with password successfully, but when reopen the saved file, it is under read-only mode. This also repro on OOo 3.3 ppt is weird, in OOo 3.3, In saved dialog, the password check box can check, input password and saved the file then reopen it, No need password, so saved with password in OOo3.3 for ppt format is not work. In trunk r1355082, In saved dialog, saveing with password checkbox is disabled If we saved ppt file with password protect by OpenOffice, then open in MS Office, no need input password, so it is totally not work. Scenario B: ods/xls/odt/odp work fine doc file with password open under read-only mode, we can saved to another file to remove password. ppt file can't saved with password protect Scenario C: xls work fine doc file with password open under read-only mode, we can saved to another file to remove password. ppt file with password protect by MS office can't open in Open Office, it said Read Error. the loading of password-encrypted Microsoft PowerPint presentations is not supported. For PPT issue, there have 2 bugs about saved(Bug 39527) and loading(Bug 46307) For doc open with read-only mode issue, i think it is a know issue, but i can't find the bug in bugzilla, I will double check in bugzilla. Date: Thu, 28 Jun 2012 11:20:54 +0200 From: orwittm...@googlemail.com To: ooo-dev@incubator.apache.org Subject: Re: Propose for 3.4.1: Can't remove password from file (119366) Hi, On 28.06.2012 11:08, Oliver-Rainer Wittmann wrote: Hi, sorry for top posting - it is reply to the complete discussion regarding save with password This functionality for ODF documents is available when the build includes modules nss and moz. I am not sure, if our buildbots include these modules. I will check it. win7 buildbot includes nss and moz linux32 buildbot includes nss and moz linux64 buildbot includes nss and moz But, as you may have notice since a couple of days our buildbots have problems to create new packages. last sucessful build for win7 was on 2012-06-19 last sucessful build for linux32 was on 2012-06-20 last sucessful build for linxu64 was on 2012-06-17 none of these contains the fix for issue 119366 Best regards, Oliver. Best regards, Oliver. On 28.06.2012 10:21, YangTerry wrote: Just confirm with our build owner, moz package is not in the build. Will verify and update the result after download the trunk build from BuildBot. Thanks Fan Zheng help (*^__^*) From: polo8...@hotmail.com To: ooo-dev@incubator.apache.org Subject: RE: Propose for 3.4.1: Can't remove password from file (119366) Date: Thu, 28 Jun 2012 16:02:31 +0800 Will confirm with our local build owner. Also download the trunk build from http://ci.apache.org/projects/openoffice/ Thanks for your investgate. Date: Thu, 28 Jun 2012 15:19:39 +0800 Subject: Re: Propose for 3.4.1: Can't remove password from file (119366) From: zheng.easy...@gmail.com To: ooo-dev@incubator.apache.org Hold a second, you guys are using the download build on verifying, right? If so, that means some problems there. Maybe the release build env broken? 2012/6/28 Fan Zheng zheng.easy...@gmail.com Muhaha, lucky for me that I have 3 build enviroment on verification this, all of them are under Windows XP. The difference are: In build env A, there is no moz package involved; In build env B, there is moz package involved, but without my solution of issue 119366; In build env C, there is moz package involved and with my solution of issue 119366; And I did the following test cases: In env A: 1.1 Save into ODT, without password, passed; 1.2 Save into ODT, with password, failed, with error message General Error: Generral input/output error; 1.3 Save into DOC, without password, passed; 1.4 Save into DOC, with password, passed; But in continual case: Save As without password, failed; 2.1 Save into ODP, without password, passed; 2.2 Save into ODP, with password, failed, with error message General Error: Generral input/output error; 2.3 Save into PPT, without password, passed; 2.4 Save into PPT, the saveing with password checkbox is disabled... Weird. 3.1 Save into ODS, without password, passed; 3.2 Save into ODS, with password, failed, with error
Re: [BUILDBOT] build breakers
On 29.06.2012 11:23, Ariel Constenla-Haile wrote: Hi Tsutomu, On Fri, Jun 29, 2012 at 6:18 AM, Tsutomu Uchino hanya.r...@gmail.com wrote: Revision 1351981 broke rev. 1341713, now no patches are applied at all: make writeable... no patch needed... touch ./unxlngx6/misc/build/so_patched_so_serf mkdir ./unxlngx6/misc/build/serf-1.1.0/ I needed to enable sef-1.0.0.ldl.patch applied on Linux x86. yes, both patches are needed and were commented. .IF $(OS)==MACOSX || $(OS)==FREEBSD || $(OS)==LINUX # Do not link against expat. It is not necessary (apr-util is already linked against it) # and does not work (we use a different expat library schema.) #PATCH_FILES+=$(TARFILE_NAME).mac.patch .ENDIF .IF $(OS)==LINUX # Add -ldl as last library so that the linker has no trouble resolving dependencies. #PATCH_FILES+=$(TARFILE_NAME).ldl.patch .ENDIF I have reactivated the patches. I used the opportunity to rename the patches: serf-1.0.0.ld.patch - serf-1.1.0.ld.patch serf-1.1.0.mac.patch - serf-1.1.0.libs.patch -Andre
Re: [IMPORTANT][Call for UX review] [Windows 8 certification]Test for Section 11 Apps must support multi-user sessions is not tested by Windows App Certification Kit
Since no further comments, I have committed the string patch to AOO3.4.1 by revision r1355344 at first. The code part is still open for review. Regards, Zhangjf On Fri, Jun 29, 2012 at 4:26 PM, zhangjf zhan...@apache.org wrote: I am reviewing yuanlin's updated patch for the new dialog message only, https://issues.apache.org/ooo/attachment.cgi?id=78521action=diff. And I suppose it will remove - Fatal Error from the error dialog title string late, so this will not introduce more strings for translation. If there is no more concerns in 4 hours from now, I will commit this string patch to 3.4.1 at first. thanks, zhangjf On Fri, Jun 29, 2012 at 3:14 PM, zhangjf zhan...@apache.org wrote: How about just simply remove - Fatal Error from the dialog title string? it won't add one more string for translation. zhangjf On Fri, Jun 29, 2012 at 2:19 PM, Jürgen Schmidt jogischm...@googlemail.com wrote: On 6/28/12 6:23 PM, zhangjf wrote: On Fri, Jun 29, 2012 at 12:15 AM, Jürgen Schmidt jogischm...@googlemail.com wrote: On 6/28/12 6:12 PM, Jürgen Schmidt wrote: On 6/28/12 5:33 PM, zhangjf wrote: If it still needs more time for discussion, I think it is also one option to only commit the new string change at first to catch up translation. It should have no impacts on function without committing the code. In this way, please review the new dialog and string first. Is it acceptable? sure, the way how it works is to check in the src file running localize to create a new sdf, convert it, update pootle, doing the translation on Pootle (to speed up and simplify the process) and finally merge it back in svn. I thunk the proposed solution is good and fulfill the requirements. Can we make a screenshot with the warning box and the English strings for review? Yuanlin's original first post in this mail thread contains the dialog snapshot url at https://issues.apache.org/ooo/attachment.cgi?id=78482. ok thanks, I have overseen this. I have 2 questions: 1. dialog title shows Fatal Error, is it really a Fatal Error? I don't think so, we detect a running instance and close the application or better don't continue to start. I think it's more a warning, isn't it? 2. in case of error I think we have a better error icon, in case of a warning the used icon is ok from my pov. Juergen In general I would support the proposed solution with a clear +1 to move forward immediately. Juergen Juergen zhangjf On Thu, Jun 28, 2012 at 10:48 PM, Jürgen Schmidt jogischm...@googlemail.com wrote: Hi, sorry for my top posting but I think this is very urgent and important. When we want to integrate this in 3.4.1 we have to do it immediately, means by the end of this week. The warning messages have to translated!!! Any opinions Juergen On 6/27/12 11:13 AM, Jürgen Schmidt wrote: On 6/27/12 3:23 AM, Lin Yuan wrote: Currently in AOO, only part of the data in user profile is locked and can not access by mutiple instances. So as tested on Windows Server 2008, AOO will crash in such situation. The patch is not to really support one user to launch multiple instances on mutiple sessions case. According to the suggestion in Windows 8 Certification below: *Note*: If an app does not support multiple user sessions or remote access, it must clearly state this when launched from this kind of session. With the patch, AOO will popup a warning dialog and exit in this case. So it will still not support mutiple user sessions for one user but the UX is more frendly than the current crash issue. we have to define fast if we want include it for 3.4.1 or not. It will require some translation effort that we have to organize in time (e.g. updating Pootle etc.) Juergen Thanks, Lin Yuan 2012/6/27 Joost Andrae joost.and...@gmx.de Hi Rob, but in a Windows Terminal Server session you have user profiles for each user. This patch is for if you connect with Terminal Services twice using the same user account. I just wanted to make sure that there is no real problem to get OpenOffice configured so it can be used within a multi user environment (MS TS, Citrix, Sun SGD, or UNIX profiles). If the same user connects a second time then there might be a locking problem with his profile data. If you want to fix this then it's OK but in my opinion it's not really needed because usually it should be prevented that one user accesses the same user profile from another terminal (RDP, X11) session. Just my two € Cents, Joost
Re: Touching base with the mirrorbrain operators
On Thu, Jun 28, 2012 at 5:40 PM, drew jensen drewjensen.in...@gmail.com wrote: Howdy, It is wonderful to see the millions of downloads for the current 3.4 release, utilizing the sf.net resources. It's also encouraging to see the half million downloads per month still going out from the mirrorbrain servers - this 500,000 p/mos figure is comprised of older releases, many of which include language packs not yet available for 3.4 or the upcoming 3.4.1 - so this is a valuable resource to the project still, IMO. Back in March I sent an email to 102 contact references for that mirror network, it would be a good idea IMO to drop another note to these folks, let them know we recognize the traffic they are still carrying, say thanks and keep in touch as the projects looks to, and plans for the next release - 3.4.1 and beyond. Sound like a good idea? If we want to use MirrorBrain longer term it would probably be good to have a ooo-mirrors list, or something like that, so we can have bidirectional communications. It would be a low-traffic list, but it would be better for the operators to join a specialized list than have them sign up for ooo-dev. -Rob Thanks //drew
Re: Touching base with the mirrorbrain operators
n Fri, Jun 29, 2012 at 9:26 AM, Rob Weir robw...@apache.org wrote: On Thu, Jun 28, 2012 at 5:40 PM, drew jensen drewjensen.in...@gmail.com wrote: Howdy, It is wonderful to see the millions of downloads for the current 3.4 release, utilizing the sf.net resources. It's also encouraging to see the half million downloads per month still going out from the mirrorbrain servers - this 500,000 p/mos figure is comprised of older releases, many of which include language packs not yet available for 3.4 or the upcoming 3.4.1 - so this is a valuable resource to the project still, IMO. Back in March I sent an email to 102 contact references for that mirror network, it would be a good idea IMO to drop another note to these folks, let them know we recognize the traffic they are still carrying, say thanks and keep in touch as the projects looks to, and plans for the next release - 3.4.1 and beyond. Sound like a good idea? If we want to use MirrorBrain longer term it would probably be good to have a ooo-mirrors list, or something like that, so we can have bidirectional communications. It would be a low-traffic list, but it would be better for the operators to join a specialized list than have them sign up for ooo-dev. Or maybe broaden it to an ooo-distributors list, to include the those distributing CD's, copies of AOO on their own website, as well as mirror operators. It benefits us if we have an easy way to communicate with them on things such as release plans (they should know things are coming well in advance), changes in branding, etc. Although the technologies differ, they all have the common interest in wanting to synch up with us on releases. But not enough to suffer through ooo-dev. -Rob -Rob Thanks //drew
Re: Touching base with the mirrorbrain operators
On Fri, 2012-06-29 at 09:26 -0400, Rob Weir wrote: On Thu, Jun 28, 2012 at 5:40 PM, drew jensen drewjensen.in...@gmail.com wrote: Howdy, It is wonderful to see the millions of downloads for the current 3.4 release, utilizing the sf.net resources. It's also encouraging to see the half million downloads per month still going out from the mirrorbrain servers - this 500,000 p/mos figure is comprised of older releases, many of which include language packs not yet available for 3.4 or the upcoming 3.4.1 - so this is a valuable resource to the project still, IMO. Back in March I sent an email to 102 contact references for that mirror network, it would be a good idea IMO to drop another note to these folks, let them know we recognize the traffic they are still carrying, say thanks and keep in touch as the projects looks to, and plans for the next release - 3.4.1 and beyond. Sound like a good idea? If we want to use MirrorBrain longer term it would probably be good to have a ooo-mirrors list, or something like that, so we can have bidirectional communications. It would be a low-traffic list, but it would be better for the operators to join a specialized list than have them sign up for ooo-dev. I would agree, it seems worthwhile at this point to setup such a list. Also heard from the Infra team that Peter Pöml has supplied a list of mirror contacts for such a list... //drew -Rob Thanks //drew
colours in Calc
I have recently installed the latest Open Office and it has totally ruined some of my documents! I had a calendar/planner using several colours which I had been able to devise for myself as well as lots from the palate provided. How on earth can I get my old document back as it took several hours work and I really do not want to have to spend hours and hours doing it all again. I have searched the pages on the web page, but I have given up trying to work my way through it all in the hope of a solution. I have used Open Office for years and have recommended it to everyone I know, but if this sort of thing happens with no warning . To quote the robot in Short Circuit - 'I am really pissed off' Pam Tuthill
[PROPOSAL] Create ML for mirrorbrain server operators
Hi, It is clear the mirrorbrain server network is still dishing up a significant number of downloads and the services are likely to be beneficial to the project for some period of time to come. Henk with the infra team reports he has a list of contact emails now for this group of mirror operators. The suggestion is therefore to create a oo-mirror mailing list to aid in communicating with this group. Henk has offered to act as a moderator for the list, and I offer to do the same. I'd like therefor to ask for lazy consensus to go ahead and create this oo-mirror list. Thanks //drew
Re: [PROPOSAL] Create ML for mirrorbrain server operators
On Fri, Jun 29, 2012 at 10:19 AM, drew jensen drewjensen.in...@gmail.com wrote: Hi, It is clear the mirrorbrain server network is still dishing up a significant number of downloads and the services are likely to be beneficial to the project for some period of time to come. Henk with the infra team reports he has a list of contact emails now for this group of mirror operators. The suggestion is therefore to create a oo-mirror mailing list to aid in communicating with this group. It looks like you did see my earlier post. I don't think we should have a dedicated list for mirrorbrain operators. But we should have a list of OpenOffice distributors in general, whether mirror operators, those seeding torrents, distributing CD's or even just hosting their own copy of AOO (CNet, FileHippo, etc.). There is a number of topics that all of these parties have in common: 1) When is the next release coming? 2) What will the matrix of languages and OS's be, so I can prepare? 3) What is the latest branding and artwork associated with this release so I can start updating my website? 4) Is there some urgent news that I need to be aware of, like a flawed release file that should be immediately withdraw? -Rob Henk has offered to act as a moderator for the list, and I offer to do the same. I'd like therefor to ask for lazy consensus to go ahead and create this oo-mirror list. Thanks //drew
Re: [PROPOSAL] Create ML for mirrorbrain server operators
On Fri, 2012-06-29 at 10:31 -0400, Rob Weir wrote: On Fri, Jun 29, 2012 at 10:19 AM, drew jensen drewjensen.in...@gmail.com wrote: Hi, It is clear the mirrorbrain server network is still dishing up a significant number of downloads and the services are likely to be beneficial to the project for some period of time to come. Henk with the infra team reports he has a list of contact emails now for this group of mirror operators. The suggestion is therefore to create a oo-mirror mailing list to aid in communicating with this group. It looks like you did see my earlier post. I don't think we should have a dedicated list for mirrorbrain operators. But we should have a list of OpenOffice distributors in general, whether mirror operators, those seeding torrents, distributing CD's or even just hosting their own copy of AOO (CNet, FileHippo, etc.). There is a number of topics that all of these parties have in common: 1) When is the next release coming? 2) What will the matrix of languages and OS's be, so I can prepare? 3) What is the latest branding and artwork associated with this release so I can start updating my website? 4) Is there some urgent news that I need to be aware of, like a flawed release file that should be immediately withdraw? -Rob Hi, That all sounds good to me, no arguments here to the idea of expanding this to a general purpose distributor ML. //drew snip
Re: Touching base with the mirrorbrain operators
2012/6/29 Rob Weir robw...@apache.org: If we want to use MirrorBrain longer term it would probably be good to have a ooo-mirrors list, or something like that, so we can have bidirectional communications. It would be a low-traffic list, but it would be better for the operators to join a specialized list than have them sign up for ooo-dev. Or maybe broaden it to an ooo-distributors list, to include the those distributing CD's, copies of AOO on their own website, as well as mirror operators. It benefits us if we have an easy way to communicate with them on things such as release plans (they should know things are coming well in advance), changes in branding, etc. Although the technologies differ, they all have the common interest in wanting to synch up with us on releases. But not enough to suffer through ooo-dev. -Rob +1. I think a mailing list for mirrors and other distributors is a really good idea Ricardo
Re: [RELEASE][3.4.1]: Include only one en-US dictionary extension
Done in SVN revision 1355436. -Andre On 25.06.2012 16:54, Andre Fischer wrote: On 19.06.2012 09:17, Andre Fischer wrote: On 19.06.2012 05:07, Ariel Constenla-Haile wrote: Hi there, there have been some reports of users complaining that the Thesaurus does not work. [...] As this bug has its root in the dictionary extensions, the only thing we can do to fix it is just provide only one extension, in this case dict-en.oxt. Dropping the other english dictionaries is a good idea for other reasons, too. Issue 119272 (https://issues.apache.org/ooo/show_bug.cgi?id=119272) describes the problem of all dictionaries using more than 160MB, most of this are the large thesaurus files. Including only one english dictionary would reduce this number considerably. Besides, it contains support for most variants of English anyway. I would like to go forward on this but am not sure that we reached consensus. Are there still concerns about this? If nobody complains in the next 72 hours then I will remove all English extensions except dict-en.oxt for both the standard English version and the British English version. See also issue 120033 (https://issues.apache.org/ooo/show_bug.cgi?id=120033) for details. -Andre
RE: [IMPORTANT][Call for UX review] [Windows 8 certification]Test for Section 11 Apps must support multi-user sessions is not tested by Windows App Certification Kit
4 hours is too short. This is an international project with contributors in different time zones around the globe. For example, when you posted this message, it was 01:26 where I am (utc-0700). I suggest that you either do CTR (commit it and be prepared for it to be rolled back, however unlikely) or do an RTC (review, then commit) that provides adequate time for interested parties to review and respond). If you want to ensure that CTR does receive review, report that you are doing so; also use a commit message that suggests review is desired. - Dennis -Original Message- From: zhangjf [mailto:zhan...@apache.org] Sent: Friday, June 29, 2012 01:26 To: ooo-dev@incubator.apache.org Subject: Re: [IMPORTANT][Call for UX review] [Windows 8 certification]Test for Section 11 Apps must support multi-user sessions is not tested by Windows App Certification Kit I am reviewing yuanlin's updated patch for the new dialog message only, https://issues.apache.org/ooo/attachment.cgi?id=78521action=diff. And I suppose it will remove - Fatal Error from the error dialog title string late, so this will not introduce more strings for translation. If there is no more concerns in 4 hours from now, I will commit this string patch to 3.4.1 at first. thanks, zhangjf On Fri, Jun 29, 2012 at 3:14 PM, zhangjf zhan...@apache.org wrote: How about just simply remove - Fatal Error from the dialog title string? it won't add one more string for translation. zhangjf On Fri, Jun 29, 2012 at 2:19 PM, Jürgen Schmidt jogischm...@googlemail.com wrote: On 6/28/12 6:23 PM, zhangjf wrote: On Fri, Jun 29, 2012 at 12:15 AM, Jürgen Schmidt jogischm...@googlemail.com wrote: On 6/28/12 6:12 PM, Jürgen Schmidt wrote: On 6/28/12 5:33 PM, zhangjf wrote: If it still needs more time for discussion, I think it is also one option to only commit the new string change at first to catch up translation. It should have no impacts on function without committing the code. In this way, please review the new dialog and string first. Is it acceptable? sure, the way how it works is to check in the src file running localize to create a new sdf, convert it, update pootle, doing the translation on Pootle (to speed up and simplify the process) and finally merge it back in svn. I thunk the proposed solution is good and fulfill the requirements. Can we make a screenshot with the warning box and the English strings for review? Yuanlin's original first post in this mail thread contains the dialog snapshot url at https://issues.apache.org/ooo/attachment.cgi?id=78482. ok thanks, I have overseen this. I have 2 questions: 1. dialog title shows Fatal Error, is it really a Fatal Error? I don't think so, we detect a running instance and close the application or better don't continue to start. I think it's more a warning, isn't it? 2. in case of error I think we have a better error icon, in case of a warning the used icon is ok from my pov. Juergen In general I would support the proposed solution with a clear +1 to move forward immediately. Juergen Juergen zhangjf On Thu, Jun 28, 2012 at 10:48 PM, Jürgen Schmidt jogischm...@googlemail.com wrote: Hi, sorry for my top posting but I think this is very urgent and important. When we want to integrate this in 3.4.1 we have to do it immediately, means by the end of this week. The warning messages have to translated!!! Any opinions Juergen On 6/27/12 11:13 AM, Jürgen Schmidt wrote: On 6/27/12 3:23 AM, Lin Yuan wrote: Currently in AOO, only part of the data in user profile is locked and can not access by mutiple instances. So as tested on Windows Server 2008, AOO will crash in such situation. The patch is not to really support one user to launch multiple instances on mutiple sessions case. According to the suggestion in Windows 8 Certification below: *Note*: If an app does not support multiple user sessions or remote access, it must clearly state this when launched from this kind of session. With the patch, AOO will popup a warning dialog and exit in this case. So it will still not support mutiple user sessions for one user but the UX is more frendly than the current crash issue. we have to define fast if we want include it for 3.4.1 or not. It will require some translation effort that we have to organize in time (e.g. updating Pootle etc.) Juergen Thanks, Lin Yuan 2012/6/27 Joost Andrae joost.and...@gmx.de Hi Rob, but in a Windows Terminal Server session you have user profiles for each user. This patch is for if you connect with Terminal Services twice using the same user account. I just wanted to make sure that there is no real problem to get OpenOffice configured so it can be used within a multi user environment (MS TS, Citrix, Sun SGD, or UNIX profiles). If the same user connects a second time then there might be a locking problem with his profile data. If you want to fix this
Re: Word cloud for ooo-dev post subjects
On Thu, Jun 28, 2012 at 11:46 PM, Kevin Grignon kevingrignon...@gmail.com wrote: KG01 - See comments inline. On Thu, Jun 28, 2012 at 6:08 AM, Rob Weir robw...@apache.org wrote: On Wed, Jun 27, 2012 at 5:36 PM, Donald Whytock dwhyt...@gmail.com wrote: On Wed, Jun 27, 2012 at 4:54 PM, Rob Weir robw...@apache.org wrote: http://people.apache.org/~robweir/ooo-dev-cloud.png This looks at the top 1000 terms used in ooo-dev post subjects since this project moved to Apache in June 2011. The only thing I removed was Re:, since that would have dominated the cloud and is machine, not user written KG01 - Great stuff Rob. These simple analytics are really interesting. *Would you be open to harvesting and sharing the extracted post titles for both ooo-dev and ooo-users and sending along to me*. The Python bit is unfamiliar to me. The text files, like you did for the twitter feeds would be great. Then I can go into wordle and tweak, using a variety of filters. Also, I am exploring other analytic tools to parse the data. Thanks. Post titles on the users list tend to be short and not very informative. We see a lot that just say Bug, or OpenOffice problem, or Help. I can certainly provide the subjects. But I wonder what would happen if we took the entire text of each root post (skipping responses)? That might give greater context. Twitter tweets are somewhat intermediate, longer than a typical post title, but shorter than a post. -Rob In this particular cloud, I used all posts, including responses. So if a term was used in a thread that had many responses, it would have additional weight in this chart. Technologies used: Python's mailbox API to extract the post titles. Could have done this with any number of command line text tools as well, but it is trivial in Python as well: import mailbox box = mailbox.mbox(fileName) for message in box: print message['Subject'] Then I used Wordle.net to generate the graphic. Based on the reaction given to the previous word cloud, I know that some list subscribers are curious to see how often we write about LibreOffice. So I'll help you find it in this graphic. Look for the big AOO, then under that see the COMMIT. Under COMMIT you can make out LIBREOFFICE, to the left of USERS. Regards, -Rob Somehow not as stylish in this font. Bug is visible in this one. No one tweets about bugs? I think this is a user/developer difference. Users talk in more direct terms, about how bugs impact them. So very few mention a bug. But 18 mentions on Twitter of some form of crash/crashed/crashing. On the ooo-dev list we call these bugs or issues. Users lose all their work. We debug an exception. The army pacifies the village, etc. It is good to remember the difference in impact our work (good or bad) has on others, even though we use more clinical terms on this list. -Rob Don
Re: [IMPORTANT][Call for UX review] [Windows 8 certification]Test for Section 11 Apps must support multi-user sessions is not tested by Windows App Certification Kit
On Fri, Jun 29, 2012 at 9:57 AM, zhangjf zhan...@apache.org wrote: If there is no more concerns in 4 hours from now, I will commit this string patch to 3.4.1 at first. aren't we in string freeze mode? Translation deadline is by end of June (this Sunday); and deadline is for fixing current strings, not introducing new ones. Regards
SVN mistake, Sorry
I just made mess in Apache-SVN. Just try to copy source code to my PC but mixed tortoiseSVN import command for downloading. So I can see that there is new ooo directory with some files from my PC. For me it is good to delete but I dont know if there is some other things to avoid bigger mistake. I started first downloading file aoo-3.4.1-incubating-src.zip from page: https://cwiki.apache.org/OOOUSERS/development-snapshot-builds.html but this zip-file seems to consist mostly empty directories, so ... Sorry Risto
Re: SVN mistake, Sorry
Risto Jääskeläinen [rjaas...@saunalahti.fi] kirjoitti: I just made mess in Apache-SVN. Just try to copy source code to my PC but mixed tortoiseSVN import command for downloading. So I can see that there is new ooo directory with some files from my PC. For me it is good to delete but I dont know if there is some other things to avoid bigger mistake. I started first downloading file aoo-3.4.1-incubating-src.zip from page: https://cwiki.apache.org/OOOUSERS/development-snapshot-builds.html but this zip-file seems to consist mostly empty directories, so ... Sorry Risto More precisely there is trunk ooo/ -directory in ooo directory which is made by me by mistake. Sorry Risto
Re: Word cloud for ooo-dev post subjects
On Thu, Jun 28, 2012 at 11:46 PM, Kevin Grignon kevingrignon...@gmail.com wrote: snip KG01 - Great stuff Rob. These simple analytics are really interesting. *Would you be open to harvesting and sharing the extracted post titles for both ooo-dev and ooo-users and sending along to me*. The Python bit is unfamiliar to me. The text files, like you did for the twitter feeds would be great. Then I can go into wordle and tweak, using a variety of filters. Also, I am exploring other analytic tools to parse the data. Thanks. Here they are: http://people.apache.org/~robweir/ooo-dev-subjects.zip http://people.apache.org/~robweir/ooo-users-subjects.zip -Rob In this particular cloud, I used all posts, including responses. So if a term was used in a thread that had many responses, it would have additional weight in this chart. Technologies used: Python's mailbox API to extract the post titles. Could have done this with any number of command line text tools as well, but it is trivial in Python as well: import mailbox box = mailbox.mbox(fileName) for message in box: print message['Subject'] Then I used Wordle.net to generate the graphic. Based on the reaction given to the previous word cloud, I know that some list subscribers are curious to see how often we write about LibreOffice. So I'll help you find it in this graphic. Look for the big AOO, then under that see the COMMIT. Under COMMIT you can make out LIBREOFFICE, to the left of USERS. Regards, -Rob Somehow not as stylish in this font. Bug is visible in this one. No one tweets about bugs? I think this is a user/developer difference. Users talk in more direct terms, about how bugs impact them. So very few mention a bug. But 18 mentions on Twitter of some form of crash/crashed/crashing. On the ooo-dev list we call these bugs or issues. Users lose all their work. We debug an exception. The army pacifies the village, etc. It is good to remember the difference in impact our work (good or bad) has on others, even though we use more clinical terms on this list. -Rob Don
Re: [UPDATE SERVICE] proposal to activate the update service for OOo 3.2 and OOo 3.2.1
On Thu, Jun 28, 2012 at 5:27 AM, Oliver-Rainer Wittmann orwittm...@googlemail.com wrote: Hi, On 27.06.2012 14:58, Rob Weir wrote: On Wed, Jun 27, 2012 at 5:30 AM, Oliver-Rainer Wittmann orwittm...@googlemail.com wrote: Hi, I am sorry - again some stuff regarding our update service. As it has been pointed out in [1] there are some more OOo versions out there for which we can activate the update service. Thus, I am proposing to activate the update service for installed OOo 3.2 and OOo 3.2.1 instances for which we have an AOO 3.4 installation package. These would be OOo 3.2/3.2.1 versions of the following languages on Windows, MacOS X, Linux 32bit and Linux 64bit: - English (US) - Arabic - Chinese (simplified) - Chinese (traditional) - Czech - Dutch - French - Galician - German - Hungarian - Italian - Japanese - Portuguese (Brazilian) - Russian - Spanish If nobody objects, I will work on the following tasks to make this happen: Although our website was pushing OOo 3.3 until May, when AOO 3.4 was ready, there were other places pushing 3.2 releases. So we still see OOo 3.2 downloads via MirrorBrain. Anyone who downloaded OOo 3.2 after the Oracle-hosted update server was decommissioned has not seen the upgrade notifications. But one question: Has anyone actually tested the upgrade experience from OOo 3.2 to AOO 3.4? As far as I know, it was not tested yet. But, I think we need to assure that is works. I tested on WinXP, 3.2.1 version (en_US with JRE). Upgrade to AOO 3.4 went smoothly. No observed problems. -Rob Best regards, Oliver.
RE: SVN mistake, Sorry
I deleted the accidental /incubator/ooo/trunk%20ooo/ commit. One way to avoid accidental commits is to use the http: (not https:) URL for the repository. This will allow SVN Checkout and SVN Update but SVN Commit will not be allowed. I'm checking the 3.4.1 src.zip now. - Dennis PS: The way I deleted the folder was using the Tortoise SVN context menu repo-browser selection to view the ooo/ level. I deleted trunk ooo in the repo-browser window by right-clicking and selecting the delete function. -Original Message- From: Risto Jääskeläinen [mailto:rjaas...@saunalahti.fi] Sent: Friday, June 29, 2012 11:38 To: AOOkehitys Subject: SVN mistake, Sorry I just made mess in Apache-SVN. Just try to copy source code to my PC but mixed tortoiseSVN import command for downloading. So I can see that there is new ooo directory with some files from my PC. For me it is good to delete but I dont know if there is some other things to avoid bigger mistake. I started first downloading file aoo-3.4.1-incubating-src.zip from page: https://cwiki.apache.org/OOOUSERS/development-snapshot-builds.html but this zip-file seems to consist mostly empty directories, so ... Sorry Risto
RE: SVN mistake, Sorry
Thank you! Risto Dennis E. Hamilton [dennis.hamil...@acm.org] kirjoitti: I deleted the accidental /incubator/ooo/trunk%20ooo/ commit. One way to avoid accidental commits is to use the http: (not https:) URL for the repository. This will allow SVN Checkout and SVN Update but SVN Commit will not be allowed. I'm checking the 3.4.1 src.zip now. - Dennis PS: The way I deleted the folder was using the Tortoise SVN context menu repo-browser selection to view the ooo/ level. I deleted trunk ooo in the repo-browser window by right-clicking and selecting the delete function. -Original Message- From: Risto Jääskeläinen [mailto:rjaas...@saunalahti.fi] Sent: Friday, June 29, 2012 11:38 To: AOOkehitys Subject: SVN mistake, Sorry I just made mess in Apache-SVN. Just try to copy source code to my PC but mixed tortoiseSVN import command for downloading. So I can see that there is new ooo directory with some files from my PC. For me it is good to delete but I dont know if there is some other things to avoid bigger mistake. I started first downloading file aoo-3.4.1-incubating-src.zip from page: https://cwiki.apache.org/OOOUSERS/development-snapshot-builds.html but this zip-file seems to consist mostly empty directories, so ... Sorry Risto
RE: SVN mistake, Sorry
@Risto, I confirm your observation about the aoo-3.4.1 src.zip. I downloaded a00-3.4.1-incubating-src.zip and there is something peculiar. WinZip reports only 948 files for a total of 5,483kB. But the zip is 357MB. I did a WinZip test and received a warning about the central directory not having an end-of-directory signature. Also, in the test output I see many pointless entries for directories, most of which have no files. I am assuming that the .zip is defective in some way. I am now downloading the tar.gz to see if that is any better. (WinZip will open these) - Dennis -Original Message- From: Risto Jääskeläinen [mailto:rjaas...@saunalahti.fi] Sent: Friday, June 29, 2012 11:38 To: AOOkehitys Subject: SVN mistake, Sorry [ ... ] I started first downloading file aoo-3.4.1-incubating-src.zip from page: https://cwiki.apache.org/OOOUSERS/development-snapshot-builds.html but this zip-file seems to consist mostly empty directories, so ... Sorry Risto
Re: [RELEASE][3.4.1][DEV-BUILD]: propose next dev snapshot for 3.4.1 based on revision 1354891
Hi. I had problems in installation AOO 3.4.1, package .deb 86-64. He broke my apt-get with the package: ooobasis-dev3.4-images_3.4.1-1_amd64.deb Solution: I had downgrade with same package 3.4.0. Albino
RE: SVN mistake, Sorry
@Risto, I recommend using a00-3.4.1-incubating-src.tar.gz if you have a Windows utility that will open it. There I see a total of 66,365 files requiring 1.6GB (uncompressed). There is apparently not a separate AOO341 branch or tag. But if the .tar.gz doesn't work for you, it appears that there is work going on at http://svn.apache.org/repos/asf/incubator/ooo/branches/AOO34 as an alternative, depending what you are interested in. - Dennis -Original Message- From: Dennis E. Hamilton [mailto:dennis.hamil...@acm.org] Sent: Friday, June 29, 2012 12:32 To: ooo-dev@incubator.apache.org Subject: RE: SVN mistake, Sorry @Risto, I confirm your observation about the aoo-3.4.1 src.zip. I downloaded a00-3.4.1-incubating-src.zip and there is something peculiar. WinZip reports only 948 files for a total of 5,483kB. But the zip is 357MB. I did a WinZip test and received a warning about the central directory not having an end-of-directory signature. Also, in the test output I see many pointless entries for directories, most of which have no files. I am assuming that the .zip is defective in some way. I am now downloading the tar.gz to see if that is any better. (WinZip will open these) - Dennis -Original Message- From: Risto Jääskeläinen [mailto:rjaas...@saunalahti.fi] Sent: Friday, June 29, 2012 11:38 To: AOOkehitys Subject: SVN mistake, Sorry [ ... ] I started first downloading file aoo-3.4.1-incubating-src.zip from page: https://cwiki.apache.org/OOOUSERS/development-snapshot-builds.html but this zip-file seems to consist mostly empty directories, so ... Sorry Risto
Re: Touching base with the mirrorbrain operators
Rob Weir wrote: If we want to use MirrorBrain longer term it would probably be good to have a ooo-mirrors list ... Or maybe broaden it to an ooo-distributors list Good idea, and this ooo-distributors list would have a clear focus, so there's no risk it will overlap with ooo-dev. And I also agree that the MirrorBrain network should be kept alive, at least to serve the same purposes it's serving now, i.e., distributing the pre-Apache, LGPL-licensed OpenOffice.org binaries, including the many localizations not yet available in 3.4.x; we cannot rely on the Apache archive for this since the traffic is still huge. Regards, Andrea.
Re: [RELEASE][3.4.1][DEV-BUILD]: propose next dev snapshot for 3.4.1 based on revision 1354891
Hi Albino, On Fri, Jun 29, 2012 at 4:44 PM, Albino B Neto bino...@gmail.com wrote: I had problems in installation AOO 3.4.1, package .deb 86-64. He broke my apt-get with the package: ooobasis-dev3.4-images_3.4.1-1_amd64.deb I haven't faced this problem. Anyway, you are downloading last week's revision. Look at this mail subject, revision should be 1354891, and it's not yet uploaded (only MacOS right now). Regards
New committer: De Bin Lei
The Apache OpenOffice PPMC announces the addition of committer De Bin Lei, leidb@ apache.org The list of all current podling committers is at: http://people.apache.org/committers-by-project.html#ooo. Committers have a defined role in the workings of the Apache Software Foundation: http://www.apache.org/foundation/how-it-works.html. For new committers: First, please read through the Guide for New Committers: http://www.apache.org/dev/new-committers-guide.html There is some useful information there related to email, security, etc. Secondly, you can control your profile and settings by using your Apache User ID and password, supplied for your Apache account, to log on to http://id.apache.org. There you can change your password to one of your choosing. In addition, please add the other e-mail addresses that you want to be known by. If you want to change the forwarding of leidb@ apache.org in the future, use the profile to accomplish that. Next, please add an entry for yourself on this project page: http://incubator.apache.org/openofficeorg/people.html Now that you are a committer, this is a good opportunity for you to make your first commit, editing that page. The simplest way is to use the Apache CMS to edit this page via the web: http://incubator.apache.org/openofficeorg/website-local.html#using-the-apache-cms-bookmarklet-simpler-method - the Apache OpenOffice PPMC
Re: [IMPORTANT][Call for UX review] [Windows 8 certification]Test for Section 11 Apps must support multi-user sessions is not tested by Windows App Certification Kit
On Sat, Jun 30, 2012 at 12:07 AM, Dennis E. Hamilton dennis.hamil...@acm.org wrote: 4 hours is too short. This is an international project with contributors in different time zones around the globe. For example, when you posted this message, it was 01:26 where I am (utc-0700). I suggest that you either do CTR (commit it and be prepared for it to be rolled back, however unlikely) or do an RTC (review, then commit) that provides adequate time for interested parties to review and respond). If you want to ensure that CTR does receive review, report that you are doing so; also use a commit message that suggests review is desired. - Dennis Thanks for the reminder. I am always willing to rollback the commit if there is any objective to the committed new string appears. I will monitor the discussion for a few more time. zhangjf -Original Message- From: zhangjf [mailto:zhan...@apache.org] Sent: Friday, June 29, 2012 01:26 To: ooo-dev@incubator.apache.org Subject: Re: [IMPORTANT][Call for UX review] [Windows 8 certification]Test for Section 11 Apps must support multi-user sessions is not tested by Windows App Certification Kit I am reviewing yuanlin's updated patch for the new dialog message only, https://issues.apache.org/ooo/attachment.cgi?id=78521action=diff. And I suppose it will remove - Fatal Error from the error dialog title string late, so this will not introduce more strings for translation. If there is no more concerns in 4 hours from now, I will commit this string patch to 3.4.1 at first. thanks, zhangjf On Fri, Jun 29, 2012 at 3:14 PM, zhangjf zhan...@apache.org wrote: How about just simply remove - Fatal Error from the dialog title string? it won't add one more string for translation. zhangjf On Fri, Jun 29, 2012 at 2:19 PM, Jürgen Schmidt jogischm...@googlemail.com wrote: On 6/28/12 6:23 PM, zhangjf wrote: On Fri, Jun 29, 2012 at 12:15 AM, Jürgen Schmidt jogischm...@googlemail.com wrote: On 6/28/12 6:12 PM, Jürgen Schmidt wrote: On 6/28/12 5:33 PM, zhangjf wrote: If it still needs more time for discussion, I think it is also one option to only commit the new string change at first to catch up translation. It should have no impacts on function without committing the code. In this way, please review the new dialog and string first. Is it acceptable? sure, the way how it works is to check in the src file running localize to create a new sdf, convert it, update pootle, doing the translation on Pootle (to speed up and simplify the process) and finally merge it back in svn. I thunk the proposed solution is good and fulfill the requirements. Can we make a screenshot with the warning box and the English strings for review? Yuanlin's original first post in this mail thread contains the dialog snapshot url at https://issues.apache.org/ooo/attachment.cgi?id=78482. ok thanks, I have overseen this. I have 2 questions: 1. dialog title shows Fatal Error, is it really a Fatal Error? I don't think so, we detect a running instance and close the application or better don't continue to start. I think it's more a warning, isn't it? 2. in case of error I think we have a better error icon, in case of a warning the used icon is ok from my pov. Juergen In general I would support the proposed solution with a clear +1 to move forward immediately. Juergen Juergen zhangjf On Thu, Jun 28, 2012 at 10:48 PM, Jürgen Schmidt jogischm...@googlemail.com wrote: Hi, sorry for my top posting but I think this is very urgent and important. When we want to integrate this in 3.4.1 we have to do it immediately, means by the end of this week. The warning messages have to translated!!! Any opinions Juergen On 6/27/12 11:13 AM, Jürgen Schmidt wrote: On 6/27/12 3:23 AM, Lin Yuan wrote: Currently in AOO, only part of the data in user profile is locked and can not access by mutiple instances. So as tested on Windows Server 2008, AOO will crash in such situation. The patch is not to really support one user to launch multiple instances on mutiple sessions case. According to the suggestion in Windows 8 Certification below: *Note*: If an app does not support multiple user sessions or remote access, it must clearly state this when launched from this kind of session. With the patch, AOO will popup a warning dialog and exit in this case. So it will still not support mutiple user sessions for one user but the UX is more frendly than the current crash issue. we have to define fast if we want include it for 3.4.1 or not. It will require some translation effort that we have to organize in time (e.g. updating Pootle etc.) Juergen Thanks, Lin Yuan 2012/6/27 Joost Andrae joost.and...@gmx.de Hi Rob, but in a Windows Terminal Server session you have user profiles for each user. This patch is for if you connect with Terminal Services twice using the same user account. I just wanted