Re: [libredwg] [bug #28858] Segfaults on DWG file

[Felipe Sanches]

some patches here... not perfect yet :-P
but feel free to push these to master, Pitanga. Because I think it is very close to the right way of doing these things. I just cant figure out what's wrong with it right now. Perhaps some of you guys can help me figure it out.

I am too sleepy. I need to go to bed now.

Happy Hacking!
Felipe "Juca" Sanches

On Thu, Feb 11, 2010 at 12:22 AM, Felipe Sanches <address@hidden> wrote:

On Wed, Feb 10, 2010 at 9:41 PM, Wookey <address@hidden> wrote:

+++ Felipe Sanches [2010-02-10 16:25 -0200]:

> hi, guys,
>
> The file attached to this bug report caught my attention to 2 bugs in our
> description of the dwg spec. The patch attached fixes these 2 bugs and thus
> results in the file being properly parsed by LibreDWG.

Thanx for the marvellously quick fix.

> As you can see in this part of the output I've got, all 14.778 objects of the
> dwg file were properly parsed, but now we have another issue: a segfault in the
> application itself (testSVG.c) instead of in the lib:

Yep - I get the same here after applying the patch.

Wookey

actually I have another patch here that tries to solve this, but I am facing some strange issues here. It SegFaults in a hard to predict pattern. Sometimes it SegFaults, sometimes it dont.

I have even got the tool to run once without segfaulting, which generated a valid svg file here. But the file is not a good conversion of your dwg. It seems that our SVG generator is still very poor. 

here it is:
http://bighead.poli.usp.br/~juca/tmp/23Nightingale.svg

If you look at it in a text editor you'll see some interesting strings in there :-)

--Felipe Sanches

0356-check-if-some-pointers-are-not-null.patch
Description: Text Data

0357-resolve-object-references-after-parsing-r2004-and-r2.patch
Description: Text Data

0358-implement-version-dependant-iterator-for-owned-obect.patch
Description: Text Data