Re: dia SIGSEGV when I move Class object

[Lars Clausen <lrclause cs uiuc edu>]

On 28 Aug 2003, Dmitry Perfilyev wrote:
> Hi!
> I have just compilled and run Dia. First impression
> was good, but when I try to create UML diagram Dia crashed.
> It crash (core dump) every time when I try to move "Class" or 
> "Template class" object. Other objects (in UML sheet) are moved 
> normal.

Class and Template Class are really the same object without and with
templates, so it'd appear to be one bug at most.

> When I try to run gdb on core I got next error:
> # gdb /usr/local/bin/dia core
> GNU gdb 5.3
> This GDB was configured as "sparc-sun-solaris2.8"...

Ah, a Solaris build.  They're only just starting to come out.  Now if I had
a Solaris machine...:)

> And stack:
>
> (gdb) where
> #0  0xfe50e71c in umlclass_update_data (umlclass=0x312640) at class.c:564
[...]

Scary, that line reads:

  umlclass->connections[0].pos = elem->corner;

which is no different than what the other objects do.  Can you check what
the other objects involved are, to see which one is the core dumper?
I.e. umlclass->connections, umlclass->connections[0], elem.

Of course the fact that GDB itself is broken by this point indicates some
runaway pointer bug earlier on.  I hate those things.  Maybe valgrind or
what's-it-called on Solaris would be the best to diagnose this.

I'd love to see Dia working on Solaris.  Hey, I'd even let Sun pay me to
work on it:)

-Lars

-- 
Lars Clausen (http://shasta.cs.uiuc.edu/~lrclause)| Hårdgrim of Numenor
"I do not agree with a word that you say, but I   |----------------------------
will defend to the death your right to say it."   | Where are we going, and
    --Evelyn Beatrice Hall paraphrasing Voltaire  | what's with the handbasket?