[GRASS-dev] [GRASS GIS] #1451: r.li.mps (mean patch size): SIGPIPE
at ipc.c:31
GRASS GIS
trac at osgeo.org
Mon Sep 19 01:46:53 EDT 2011
#1451: r.li.mps (mean patch size): SIGPIPE at ipc.c:31
----------------------+-----------------------------------------------------
Reporter: hamish | Owner: grass-dev@…
Type: defect | Status: new
Priority: normal | Milestone: 6.4.2
Component: Raster | Version: svn-develbranch6
Keywords: r.li.mps | Platform: Linux
Cpu: x86-64 |
----------------------+-----------------------------------------------------
Hi,
as reported on the ML, r.li.mps falls over with a SIGPIPE. Running outside
of gdb results in a "double free or corruption (fasttop)" fault.
to reproduce: (North Carolina dataset)
{{{
g.region rast=landcover_1m
r.mapcalc "forest.test = if(landcover_1m == 2, 1, null())"
mkdir -p ~/.r.li/history/
cat << EOF > ~/.r.li/history/movwindow7
SAMPLINGFRAME 0|0|1|1
SAMPLEAREA -1|-1|0.0147679324895|0.0133079847909
MOVINGWINDOW
EOF
r.li.mps map=forest.test conf=movwindow7 out=mps.test
...
Program received signal SIGPIPE, Broken pipe.
}}}
gdb session extracts:
{{{
(gdb) frame 2
#2 0x00007ffff754edc9 in calculateIndex (file=0x605270 "movwindow7",
f=0x4013a7 <meanPatchSize>, parameters=0x0,
raster=0x605250 "forest1996.maybe", output=0x605590 "test.out") at
daemon.c:198
198 send(child[j].channel, &m);
(gdb) list
193 j = 0;
194
195
196 while (j < WORKERS && donePid != child[j].pid)
197 j++;
198 send(child[j].channel, &m);
199
200 }
201
202 /*kill childs */
(gdb) frame 1
#1 0x00007ffff7551365 in send (pipe=5, m=0x7fffffffaa80) at ipc.c:31
31 check = write(pipe, m, sizeof(msg));
(gdb) list
26 int send(int pipe, msg * m)
27 {
28 int check;
29
30 /* write on pipe */
31 check = write(pipe, m, sizeof(msg));
32 if (check > 0)
33 return 1;
34 else
35 return 0;
}}}
Hamish
--
Ticket URL: <https://trac.osgeo.org/grass/ticket/1451>
GRASS GIS <http://grass.osgeo.org>
More information about the grass-dev
mailing list