Carl Troein

Show:

What's happening?

  • Linux error/warning fixes

    Fix for errors/warnings under linux against bibo patch rollup 327317. Also, I've tried to sort out some of the worst inconsistencies in the indentation in rcx_comm.c as it was driving me nuts. It's still a mess, but at least now there exists a tab setting (4) where all indentation is positive and even mostly consistent within each block. Some of the changes: don't use the undefined fd_winux on...

    2009-05-17 17:15:07 UTC in BrickOS Alternative LEGO Mindstorms OS

  • Spurious 'polygon not closed'

    'set object polygon' sometimes emits the warning "Polygon is not closed - adding extra vertex" even though the polygon is closed. It appears to be due to the use of memcmp in the test (set.c, line 3660) as there may be non-zeroed padding in struct position. This is with gcc 4.3.2 on x86_64 against the CVS version from 2009-04-22.

    2009-04-22 08:43:28 UTC in gnuplot development

  • java bindings broken with gcc 4.3.1

    Compiling libsbml_wrap.cpp with g++ 4.3.1 with strict aliasing (as is currently done) leads to a broken libsbmlj. 'make check' in src/bindings/java very quickly leads to a segfault in Java_org_sbml_libsbml_libsbmlJNI_SBMLDocument_1getTypeCode. The culprit seems to be swig (tested with 1.3.36). One way to fix this would be to make sure that libsbml_wrap.cpp is compiled with -fno-strict-aliasing.

    2008-12-12 17:09:26 UTC in Systems Biology Markup Language (SBML)

  • Comment: Linux shared library problems

    Fair enough. It's easily worked around and probably very distribution-specific in any case. But at least now it's documented. //Carl.

    2008-11-28 11:56:09 UTC in Systems Biology Software Project

  • Comment: Loading model always fails

    Hi Frank, and thanks for the quick response. I'm trying it right now at home (on openSUSE 11.0, versus 10.3 at work). I get pretty same error after following your instructions. As this output from pstree shows, the broker and roadRunner are running happily (and I'm using bash, the only shell I really know my way around): │ └─bash,3877 │ ├─Broker,21716 │ │.

    2008-11-28 11:53:57 UTC in Systems Biology Software Project

  • Comment: Loading model always fails

    Killing the broker and restarting it, manually setting SBW_HOME and PATH had some effect. The error message is now different. $ SBW_HOME=/home/carl/SBW-2.7.8 PATH=$PATH:/home/carl/SBW-2.7.8/bin Broker Then from SBW Inspector Module -> start edu.kgi.roadRunner.sim Simulation Service -> open any model -> Exception : An exception was thrown by the type initializer for...

    2008-11-26 17:06:42 UTC in Systems Biology Software Project

  • Comment: Linux shared library problems

    After the workarounds described in the report, the installation only generates what appears to be a bogus error message: Error executing post installation script /home/carl/SBW-2.7.8/register.sh \child process exited abnormally.

    2008-11-26 16:36:36 UTC in Systems Biology Software Project

  • Loading model always fails

    As shown in the attached screenshot, attempting to open any SBML model (e.g. color.xml or brusselator.xml) invariably produces this error message: Exception : The currently selected simulator does not support this operation Detailed Exception :

    2008-11-26 16:33:22 UTC in Systems Biology Software Project

  • Linux shared library problems

    I've tried to SBW install on openSUSE 10.3 (x86_64), and some dependencies turned out to take a fair bit of work to resolve. Mono was apparently required - easy enough to figure out from the error messages, but it would've been nice if the installer had checked for it. Then there was this: Error executing post installation script /home/carl/SBW-2.7.8/register.sh...

    2008-11-26 16:11:41 UTC in Systems Biology Software Project

  • postscript PolyFill ghostscript rounding workaround

    Previously posted on the bugs list; reposted here since it includes a patch. I noticed that with filledcurves the filled area diverges from the drawn line when viewed in gv. The path drawing code has a workaround for rounding error in the postscript interpreter, but this is missing in the polygon drawing. This patch changes the polygon drawing code so that they both output the exact same...

    2008-03-11 21:07:26 UTC in gnuplot development

About Me

  • 2002-05-28 (7 years ago)
  • 554528
  • troein (My Site)
  • Carl Troein

Send me a message