Re: why does this seg fault?

Mike,

You did not send a Makefile. I have no problems with your scripts when doing:

  root > .L test.cc++
  root > test()
  root > .L macro.C++
  root > macro()

see the two files in attachement with some minor changes.

Rene Brun

On Thu, 20 Apr
2006, Mike Williams wrote:

>
> hi,
>
> i've written a small example to illustrate a problem i found. i've
> attached several files:
>
> HistoMaker.h: Class HistoMaker definition file.
> LinkDef.h: LinkDef i'm using
> macro.C: An example macro
> test.cc: A compilable version of macro (without the plotting)
>
> HistoMaker is a simple class with a function that returns a vector<TH1F*>
> of filled histos (it also allocates memory for the histos). if i call it
> once everything is ok. to avoid a memory leak, before i call it again i
> delete the histos i got the previous call (this is done in macro.C).
>
> the output i get when i run macro() is:
>
> Iteration: 0
> Ready to build hGaus_0...
> Build successfull.
> Ready to build hGaus_1...
> Build successfull.
> Ready to build hGaus_2...
> Build successfull.
> Ready to build hGaus_3...
> Build successfull.
> deleted hGaus_0
> deleted hGaus_1
> deleted hGaus_2
> deleted hGaus_3
> Iteration: 1
> Ready to build hGaus_0...
>
> *** Break *** segmentation violation
> Generating stack trace...
> 0x007d9c8d in THashTable::FindObject(char const*) const + 0x53 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCore.so.5.10
> 0x007d8f68 in THashList::FindObject(char const*) const + 0x1e from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCore.so.5.10
> 0x00eefd1a in TH1::Build() + 0x26a from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libHist.so.5.10
> 0x00eee826 in TH1::TH1(char const*, char const*, int, double, double) +
> 0x132 from /usr/local/root_v5.10.00.RHEL.gcc34/lib/libHist.so.5.10
> 0x00f0508c in TH1F::TH1F(char const*, char const*, int, double, double) +
> 0x48 from /usr/local/root_v5.10.00.RHEL.gcc34/lib/libHist.so.5.10
> 0x01c70d44 in <unknown> from
> /export/dsk20/williams/delete_problem/././libHistoMaker.so
> 0x001555b5 in G__ExceptionWrapper + 0x5d from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x00181577 in G__exec_asm + 0xaa7 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x002513a8 in G__exec_loop + 0x60e from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x00251956 in G__exec_for + 0x264 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x0025612b in G__exec_statement + 0x42fd from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x0021d5c8 in G__interpret_func + 0x26e0 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x0020128d in G__getfunction + 0x25e1 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x001f4be9 in G__getitem + 0x881 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x001f2a69 in G__getexpr + 0xdd69 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x0024be98 in G__exec_function + 0xd6 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x002564d6 in G__exec_statement + 0x46a8 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x001d1766 in G__exec_tempfile_core + 0x35a from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x001d1a13 in G__exec_tempfile_fp + 0x31 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x00260268 in G__process_cmd + 0x54ec from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCint.so.5.10
> 0x007e8b55 in TCint::ProcessLine(char const*, TInterpreter::EErrorCode*)
> + 0x141 from /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCore.so.5.10
> 0x00713c08 in TApplication::ProcessLine(char const*, bool, int*) + 0x68a
> from /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCore.so.5.10
> 0x0055c0d4 in TRint::HandleTermInput() + 0x23a from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libRint.so.5.10
> 0x0055a957 in TTermInputHandler::Notify() + 0x25 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libRint.so.5.10
> 0x0055c58e in TTermInputHandler::ReadNotify() + 0x14 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libRint.so.5.10
> 0x0088a72d in TUnixSystem::CheckDescriptors() + 0x163 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCore.so.5.10
> 0x0088949c in TUnixSystem::DispatchOneEvent(bool) + 0x188 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCore.so.5.10
> 0x007970d4 in TSystem::InnerLoop() + 0x20 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCore.so.5.10
> 0x0079706d in TSystem::Run() + 0x7d from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCore.so.5.10
> 0x00714483 in TApplication::Run(bool) + 0x37 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libCore.so.5.10
> 0x0055ba41 in TRint::Run(bool) + 0x453 from
> /usr/local/root_v5.10.00.RHEL.gcc34/lib/libRint.so.5.10
> 0x08048ea1 in main + 0x8d from
> /usr/local/root_v5.10.00.RHEL.gcc34/bin/root.exe 0x034e579a in
> __libc_start_main + 0xda from /lib/tls/libc.so.6
> 0x08048d81 in TApplicationImp::ShowMembers(TMemberInspector&, char*) +
> 0x35 from /usr/local/root_v5.10.00.RHEL.gcc34/bin/root.exe
> Root > Function macro() busy flag cleared
>
>
> if i comment out the delete histos[h] line, it runs the entire way through
> without seg faulting...but, of course, i get warning messages about
> replacing hGaus_0, etc., possible memory leak...
>
> i'm not sure what's wrong here...am i just missing something or is this a
> bug?
>
> also, test.cc is basically macro.C without the plotting, and it runs with
> no errors when compiled.
>
> thanks,
>
> mike williams
>
>
>
>