|Anonymous | Login||2017-11-19 09:09 EST|
|Main | My View | View Issues | Change Log | Roadmap | My Account|
|View Issue Details|
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0001926||graphviz||Neato||public||2010-07-22 04:53||2011-04-28 04:03|
|Summary||0001926: crash/garbage from neato/circo/twopi/fdp from small graph with ports|
The attached graph causes most of the layout programs (but not dot)
to either segfault or emit output with long strings of garbage attached
to text. In all cases I'm running them as "fdp -Tsvg <foo.dot" (though
the segfaults also happen with e.g. PNG output).
Bugs 1630 and 1965 look vaguely similar to this one.
Typical backtrace (note the garbled title element
on the last edge emitted before the crash):
(gdb) run -Tsvg < foo.dot
Starting program: /opt/local/bin/neato -Tsvg < foo.dot
[... snipping the beginning of the SVG output ...]
<!-- n1->n2 -->
<g id="edge2" class="edge"><title>n1:foD$foL$ foT$0fo\[email protected]?L$X?T$\?Å#D$T??h?,$?U???(?]?u??}??->n2</title>
<path fill="none" stroke="black" d="M37.5,-29.5C37.5,-29.5 49.1463,-29.7548 62.3607,-30.0439"/>
<polygon fill="black" stroke="black" points="62.5304,-33.5483 72.6045,-30.2679 62.6835,-26.55 62.5304,-33.5483"/>
<!-- n1->n2 -->
Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_INVALID_ADDRESS at address: 0xfffffff0
0x96f87a90 in strlen ()
#0 0x96f87a90 in strlen ()
0000001 0x0002979a in strdup_and_subst_obj ()
0000002 0x0001ae65 in emit_edge ()
0000003 0x0001c80d in emit_page ()
0000004 0x0001d205 in emit_graph ()
0000005 0x0001ef01 in gvRenderJobs ()
0000006 0x00001dab in main ()
|Steps To Reproduce|
n1 [ label=<<TABLE><TR><TD PORT="p1">A</TD></TR><TR><TD PORT="p2">B</TD></TR></TABLE>> ]
n1:p2 -> n2
n1:p1 -> n2
[wiml] I reported a bug via the web interface a few minutes ago (a crash in
neato/fdp/twopi/etc.). Looking through the bug list again, I noticed bug
0001923 ("Memory corruption when using neato and svg to plot nodes with
ports"), which sounds like the same bug as my bug. Conveniently, that
bug is already fixed in CVS.
I haven't tried testing against CVS head; I think I will have to wait
until macports gets an updated version of graphviz to test that.
Indeed, I think you hit the same problem reported in 0001923, which has been fixed.
I certainly can't replicate the problem. Obviously, if you still have problems when macports
is update, let us know.
|Tags||No tags attached.|
|2011-04-28 04:03||user1||New Issue|
|2011-04-28 04:03||user1||Assigned To||=> erg|
|MantisBT 1.2.5[^] Copyright © 2000 - 2011 MantisBT Group|