← Back to team overview

kicad-developers team mailing list archive

Re: SIGTRAP crash with ngspice


I wanted to close the loop on the SIGTRAP crash that I reported a few days ago.

The issue is probably unique to Fedora, and stems from a huge quantity of "percent complete" messages that the ngspice library passes back to KiCad.  I'm not kidding when I say huge - it amounts to over 1 megabyte of text, including many backspace characters.  When I try to select the text, it probably blows up the clipboard and causes the crash.

The ideal fix will be to have the Fedora ngspice maintainer remove a flag, which in turn will disable the "percent complete" messages.

In the interim, I have a patch that filters out any line containing a string of backspaces.  I don't know if there is any interest in applying this patch to the official KiCad sources, but I've attached it here in case there is interest.  I see it as defensive programming. :-)

I'll wait a bit to see what happens with the ngspice library.  If it looks like there will be a substantial delay in having the flag removed, I can always apply my patch and push that as the next official Fedora build.


On 1/11/19 2:25 PM, Steven A. Falco wrote:
> On 1/11/19 1:51 PM, Steven A. Falco wrote:
>> On 1/11/19 1:21 PM, Seth Hillbrand wrote:
>>> Am 2019-01-11 12:18, schrieb Steven A. Falco:
>>>> I tried another ngspice experiment.  I ran a simulation, and it
>>>> worked, albeit with the data size warning I asked about earlier.  So
>>>> far so good.
>>> That warning was saying that it ran out of memory and had overwritten unplanned space.  I would expect a crash after that.  I'm not sure that KiCad can trap that kind of error as it occurred inside the library.
>> Ok, but I'm concerned about the bug reports I'll get if I were to enable ngspice and push the build to the Fedora community.
>> I'll wait to see if the forum has any thoughts on the warning.  I'll only enable ngspice in the official Fedora builds if I can get to where the examples run cleanly.
> I saved the netlist from the simple example I've been running:
> .title KiCad schematic
> V101 in 0 PULSE (0 5 1u 1u 1u 1 1)
> C101 Net-_C101-Pad1_ 0 1u
> C102 out 0 100n
> R101 Net-_C101-Pad1_ in 10k
> R102 out Net-_C101-Pad1_ 1k
> .save @v101[i]
> .save @c101[i]
> .save @c102[i]
> .save @r101[i]
> .save @r102[i]
> .save V(0)
> .save V(GND)
> .save V(Net-_C101-Pad1_)
> .save V(in)
> .save V(out)
> .tran 1u 100m
> .end
> When I run this circuit in stand-alone ngspice, it does a lot of overprinting, apparently to indicate the run-time of the job.  Here is a small example of the character stream, where there are tons of backspaces as part of the overprinting:
> saf$ od -c foo
> 0000000   %   0   .   0   0  \b  \b  \b  \b  \b   %   0   .   0   0  \b
> 0000020  \b  \b  \b  \b   %   0   .   0   0  \b  \b  \b  \b  \b   %   0
> 0000040   .   0   0  \b  \b  \b  \n
> 0000047
> The stand-alone captured output from this simple circuit is around 1 Mbyte!  I'm guessing that this huge amount of data is blowing up the window and showing all of the strange characters.
> I'm not sure of the division of labor here.  Is that something that a KiCad developer should address, or is it a library issue?  Hopefully there is a flag that can be passed to the library to suppress the overprinted text.
> 	Steve

--- a/eeschema/sim/sim_plot_frame.cpp	2019-01-13 15:57:18.798891439 -0500
+++ b/eeschema/sim/sim_plot_frame.cpp	2019-01-13 15:58:03.648978524 -0500
@@ -1282,8 +1282,14 @@
 void SIM_PLOT_FRAME::onSimReport( wxCommandEvent& aEvent )
-    m_simConsole->AppendText( aEvent.GetString() + "\n" );
-    m_simConsole->SetInsertionPointEnd();
+    wxString t = aEvent.GetString();
+    // Toss any text with a string of embedded backspaces, as these are
+    // progress status lines that don't print properly.
+    if( t.find("\b\b\b") == wxString::npos ) {
+        m_simConsole->AppendText( t + "\n" );
+        m_simConsole->SetInsertionPointEnd();
+    }

Follow ups