[gdal-dev] JP2 Segfault

[Yves Jacolin]

Hello,

I am working on some JP2 data (total 262 Go) . I started creating a vrt file 
and then "gdal_translate" them.

I always have a segfault. I tried to use a GDAL_SKIP config to use another 
driver but it seems that I can successfully do this:

>> Stack trace in bt <<

Starting program: /usr/local/bin/gdal_translate --config GDAL_SKIP JP2ECW --
config GDAL_SKIP ECW -a_srs EPSG:2154 -co BIGTIFF=IF_SAFER -co COMPRESS=JPEG -
co TILED=YES -co PHOTOMETRIC=YCBCR ORTHO_EXPRESS_2015.vrt 
ORTHO_EXPRESS_2015.tif
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Input file size is 470000, 640000
0[New Thread 0x7fffd0278700 (LWP 14580)]
[New Thread 0x7fffcfa77700 (LWP 14581)]
[New Thread 0x7fffcf276700 (LWP 14582)]
[New Thread 0x7fffcea75700 (LWP 14583)]

Program received signal SIGSEGV, Segmentation fault.
0x00007ffff10ba0f9 in CNCSJPCPacket::ParseHeader(CNCSJPC&, CNCSJPCIOStream&, 
CNCSJPCProgression*, bool, bool*) () from /usr/local/lib/libNCSEcw.so.0
(gdb) bt
#0  0x00007ffff10ba0f9 in CNCSJPCPacket::ParseHeader(CNCSJPC&, CNCSJPCIOStream&, 
CNCSJPCProgression*, bool, bool*) () from /usr/local/lib/libNCSEcw.so.0
#1  0x00007ffff10c20b1 in CNCSJPCPrecinct::ReadPackets() () from 
/usr/local/lib/libNCSEcw.so.0
#2  0x00007ffff10c1c5b in CNCSJPCPrecinct::CreateSubBands(bool) () from 
/usr/local/lib/libNCSEcw.so.0

>> end of bt <<

Using  --config GDAL_SKIP ECW doesn't change anything.

I can't understand two things:

* first: why libNCSEcw is used instead of openjpeg200 lib?
* second: what kind of problem could trigger this kind of segfault?

I am using GDAL 2.1.1 :)

Thanks for any help,

Y.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 490 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.osgeo.org/pipermail/gdal-dev/attachments/20160721/01814748/attachment.sig>