Re: [Libreoffice] calc: Re: com.sun.star.container.XNameContainer.removeByName() fails v3.5b0+ (linux)
Jan Holesovsky píše v Út 20. 12. 2011 v 16:26 +0100: Hi Tomas, On 2011-12-20 at 13:23 +0100, Tomas Hlavaty wrote: i guess 3.4.4 is this: libreoffice-3.4.4.2 beta0: libreoffice-3.5.0.0 beta1: libreoffice-3.4.99.1 yes, really :) Is that somewhere documented? Or wouldn't it be better if it was obvious from the tag names? Bugs happen - sorry for that :-( We should probably re-tag the beta0 as libreoffice-3.4.99.0, and dump libreoffice-3.5.0.0 from the repositories. I have created the libreoffice-3.4.99.0 tag today. I left the tag there libreoffice-3.5.0.0 because it is not easy to get rid of it. Anyone whoc have it, could push it back easily. I am sorry for the troubles. Best Regards, Petr ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] calc: Re: com.sun.star.container.XNameContainer.removeByName() fails v3.5b0+ (linux)
Michael Stahl píše v Po 19. 12. 2011 v 19:26 +0100: What are the git tags for v3.4.4, v3.5b0 and v3.5b1? i guess 3.4.4 is this: libreoffice-3.4.4.2 I have create the tags libreoffice-3.4.x-final based on the last release candidates for the give release. It should make it easier to get final sources. In this case, you might use libreoffice-3.4.4-final now. Best Regards, Petr ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] calc: Re: com.sun.star.container.XNameContainer.removeByName() fails v3.5b0+ (linux)
Hi Michael, Michael Stahl mst...@redhat.com writes: you are much more likely to get a reply for this kind of bug on the dev list. sorry I haven't got the feel for the right places yet;-) sounds like a bug to me. guess you should file it in bugzilla and set the regression keyword. ok, I'll do that when I'm confident it's a regression. in sc you should find some new C++ unit tests and old Java unit tests (in qa/complex). Thank you for the pointer. i guess 3.4.4 is this: libreoffice-3.4.4.2 beta0: libreoffice-3.5.0.0 beta1: libreoffice-3.4.99.1 yes, really :) Is that somewhere documented? Or wouldn't it be better if it was obvious from the tag names? Thank you, Tomas ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] calc: Re: com.sun.star.container.XNameContainer.removeByName() fails v3.5b0+ (linux)
Hi Bjoern, Bjoern Michaelsen bjoern.michael...@canonical.com writes: Any you could try bibisect: http://sweetshark.livejournal.com/7683.html to close in on the bug even more (and thus make it get fixed faster). bibisect is brilliant but unfortunatelly dosn't cover the range I would need for this (v3.4.4+). Regards, Tomas ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] calc: Re: com.sun.star.container.XNameContainer.removeByName() fails v3.5b0+ (linux)
Hi Michael, Michael Stahl mst...@redhat.com writes: i guess 3.4.4 is this: libreoffice-3.4.4.2 libreoffice-3.4.4.2 tag doesn't exist:-( Any other ideas? Are there any rules how tags are created and named? And branches? Thank you, Tomas ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] calc: Re: com.sun.star.container.XNameContainer.removeByName() fails v3.5b0+ (linux)
On 20/12/11 15:51, Tomas Hlavaty wrote: Hi Michael, Michael Stahl mst...@redhat.com writes: i guess 3.4.4 is this: libreoffice-3.4.4.2 libreoffice-3.4.4.2 tag doesn't exist:-( Any other ideas? Are there any rules how tags are created and named? And branches? the 3.4 stuff does not exist in the core repo but in the old bootstrap based zoo of repos. http://wiki.documentfoundation.org/Development/Native_Build_From_Bootstrap On 20/12/11 14:52, Tomas Hlavaty wrote: could you please advise me how to build and run the test case? I did: $ cd sc/ $ make and it built the tests. It seems to have run something called CHK from the console log, but where can I find the test results? the CUT (or JUT) lines announce that a single unit test is being run. the output is redirected to a log file, and if the test is successful no output is displayed on the terminal. if the test fails, then the log file is output to the terminal. How do I run this single test case? it seems you can't run this one yet because it isn't actually built with a makefile :) if it were it would be something like this: make -r /data/lo/core/workdir/unxlngx6/CppunitTest/sc_cellrangesbase.test ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] calc: Re: com.sun.star.container.XNameContainer.removeByName() fails v3.5b0+ (linux)
Hi Tomas, On 2011-12-20 at 13:23 +0100, Tomas Hlavaty wrote: i guess 3.4.4 is this: libreoffice-3.4.4.2 beta0: libreoffice-3.5.0.0 beta1: libreoffice-3.4.99.1 yes, really :) Is that somewhere documented? Or wouldn't it be better if it was obvious from the tag names? Bugs happen - sorry for that :-( We should probably re-tag the beta0 as libreoffice-3.4.99.0, and dump libreoffice-3.5.0.0 from the repositories. Regards, Kendy ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] calc: Re: com.sun.star.container.XNameContainer.removeByName() fails v3.5b0+ (linux)
Hello Tomas, 2011/12/20 Michael Stahl mst...@redhat.com: On 20/12/11 14:52, Tomas Hlavaty wrote: could you please advise me how to build and run the test case? I did: $ cd sc/ $ make and it built the tests. It seems to have run something called CHK from the console log, but where can I find the test results? the CUT (or JUT) lines announce that a single unit test is being run. the output is redirected to a log file, and if the test is successful no output is displayed on the terminal. if the test fails, then the log file is output to the terminal. How do I run this single test case? it seems you can't run this one yet because it isn't actually built with a makefile :) Yes and that was my intention. I made a assumption in the test that might not be true for every build. I need someone with an non en-US build and non en-US locale to check that forcing en-US in the test setup works for sheet names too. You can use the attached patch and run a normal make in sc. The test should then be executed as sc_spreadsheetobj. You can find the logfile for the test at workdir/platform/CppunitTest/sc_spreadsheetobj.test.log Regards, Markus ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] calc: Re: com.sun.star.container.XNameContainer.removeByName() fails v3.5b0+ (linux)
hi Tomas, you are much more likely to get a reply for this kind of bug on the dev list. On 19/12/11 16:32, Tomas Hlavaty wrote: Hi all, something has changed recently and it is not possible to remove sheets via uno calls anymore. It used to work with v3.4.4 but fails with v3.5b0 and v3.5b1. Steps to reproduce: 1) Create a new spreadsheet via uno 2) While trying to remove the second sheet com.sun.star.container.XNameContainer.removeByName(Sheet2) via uno fails (on an instance of com.sun.star.sheet.XSpreadsheets) v3.5b0 at least indicates an exception: type: com.sun.star.uno:runtime-exception message: C++ code threw St12out_of_range: vector::_M_range_check context: NULL v3.5b1 simply disconnects leaving me with a broken pipe. sounds like a bug to me. guess you should file it in bugzilla and set the regression keyword. I don't have an easy to run test case at the moment:-( Any pointers to a simple existing code that does something similar and is easy to set up so that I can modify it to document this bug with a running code? in sc you should find some new C++ unit tests and old Java unit tests (in qa/complex). What are the git tags for v3.4.4, v3.5b0 and v3.5b1? i guess 3.4.4 is this: libreoffice-3.4.4.2 3.5 tags: beta0: libreoffice-3.5.0.0 beta1: libreoffice-3.4.99.1 yes, really :) ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] calc: Re: com.sun.star.container.XNameContainer.removeByName() fails v3.5b0+ (linux)
On Mon, Dec 19, 2011 at 07:26:21PM +0100, Michael Stahl wrote: sounds like a bug to me. guess you should file it in bugzilla and set the regression keyword. Any you could try bibisect: http://sweetshark.livejournal.com/7683.html to close in on the bug even more (and thus make it get fixed faster). Best, Bjoern ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] calc: Re: com.sun.star.container.XNameContainer.removeByName() fails v3.5b0+ (linux)
Hey, let me check that. If it is a bug I'll create a test and fix it for beta3. Markus ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] calc: Re: com.sun.star.container.XNameContainer.removeByName() fails v3.5b0+ (linux)
Hello Tomas, Ok, I wrote a simple c++ based test for that and can't reproduce your problem. The test ( http://cgit.freedesktop.org/libreoffice/core/commit/?id=b77902d38d06df87694564ded859733bdb1d4fb2 ) creates an empty document, checks that it has a second sheet( bad assumption there from me) and removes the sheet. Eventually I even check that removing with the same name will throw an NoSuchElementException. I also had some quick looks at the implementation and I don't see any place that would throw an out of bounds exception. All of these parts pass without any problem, not even a suspicious line in a dbgutil build. Can you open a bug report for this with some more details and assign it to me? Regards, Markus ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice