On Tuesday 11 August 2009 03:44:39 Dan Langille wrote: > Dan Langille wrote: > > Dan Langille wrote: > >> Kern Sibbald wrote: > >>> On Monday 10 August 2009 03:14:26 Dan Langille wrote: > >>>> Kern Sibbald wrote: > >>>>> On Sunday 09 August 2009 20:44:03 Dan Langille wrote: > >>>>>> Kern Sibbald wrote: > >>>>>>> On Sunday 09 August 2009 17:31:04 Dan Langille wrote: > >>>>>>>> FYI: this came up on my first git run. No idea why. > >>>>>>> > >>>>>>> Well the first thing would be to find out why gdb says that your > >>>>>>> source and binary do not agree. Fixing that will probably fix the > >>>>>>> problem > >>>>>> > >>>>>> Did a fresh run. Same problem. > >>>>>> > >>>>>> Not sure why this occurs: 'Variable "exelen" is not available' > >>>>>> > >>>>>> You say it's source != binary. Given a fresh run, how can that > >>>>>> happen? > >>>>> > >>>>> I don't know the answer. You are doing the build. > >>>>> > >>>>> If the binary and source don't correspond as gdb says, then the > >>>>> "build" is broken or you are not executing the binary you think you > >>>>> are. > >>>> > >>>> These are the regression tests. That's all I'm running: > >>> > >>> Dan: > >>> The first thing is to update to the current git, then try again (there > >>> was for a short time a build error that crept into the repo). If it > >>> fails again, then you should manually build with "make setup" and > >>> ensure there are no build errors. If there are, I need to see them. > >>> If make setup works, then try manually running one of the tests. If it > >>> fails, you will need to manually run the FD under the debugger to get a > >>> good traceback -- traceback below is not very useful. It doesn't make > >>> a lot of sense, unless it is recursing -- i.e. it got a seg fault, then > >>> in the automatic traceback handler, it got another seg fault. Such > >>> cases are unusual, and the solution is to manually run the debugger so > >>> that you catch it at the first crash and > >> > >> manually produce the traceback. > >> > >> Done as instructed above. Note that I'm now running this. I think the > >> problem is that one test fails, generating the traceback. When that > >> happens, I'll do as suggested above. > > > > FYI I see that my regression test from last night had two failures: > > > > http://regress.bacula.org/viewTest.php?onlyfailed&buildid=8588 > > > > I'll report back after the current test runs. > > I think I found the cause: > > $ tests/multi-storage-test > mkdir: /usr/home/dan/src/BaculaRegressionTesting-TRUNK/regress/working2: > File exists > sed: 1: "s/Plugin/#Plugin/i": bad flag in substitute command: 'i'
Yes, I just saw that too. I suspect that the author expected it to do a case insensitive replacement -- however, I think it is not portable. Thanks for resolving the problem -- I'll fix the git repo this morning ... Kern > > > === Starting multi-storage-test at 19:31:17 === > > Starting the Bacula Storage daemon > Starting the Bacula Storage daemon > Starting the Bacula File daemon > 10-Aug 19:31 bacula-fd: Fatal Error at filed.c:320 because: > No File daemon resource defined in > /usr/home/dan/src/BaculaRegressionTesting-TRUNK/regress/bin/bacula-fd.conf > > etc... > > If I remove that trailing i, the test completes successfully. > > The FreeBSD sed page is here: > > http://www.freebsd.org/cgi/man.cgi?query=sed&apropos=0&sektion=0&manpath=Fr >eeBSD+7.2-RELEASE&format=html > > I looked at RedHat: > > http://www.freebsd.org/cgi/man.cgi?query=sed&apropos=0&sektion=0&manpath=Re >d+Hat+Linux%2Fi386+9&format=html > > But couldn't figure out what the i was supposed to be doing. > > --------------------------------------------------------------------------- >--- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 > 30-Day trial. Simplify your report design, integration and deployment - and > focus on what you do best, core application coding. Discover what's new > with Crystal Reports now. http://p.sf.net/sfu/bobj-july > _______________________________________________ > Bacula-devel mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/bacula-devel ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Bacula-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/bacula-devel
