[LinuxCNC/linuxcnc Issue#162] `/tests/interp/compile` test creates junk file

$ runtests tests/interp/compile
Running test: tests/interp/compile
Runtest: 1 tests run, 1 successful, 0 failed + 0 expected
$ /bin/ls -a tests/interp/compile | grep '^[^Retu.]' | od --format=x1
0000000 c8 e3 53 da 7f 0a
0000006
$ head -3 $'tests/interp/compile/\xc8\xe3\x53\xda\x7f'
5161    0.000000
5162    0.000000
5163    0.000000

#1 – SebKuzminsky 于 2016-09-14

I’m not able to reproduce this problem. I’m testing with 88e22fa (which is just after 2.7.7) on Debian Stretch.

$ git clean -fdx tests/
$ runtests tests/interp/compile/
Running test: tests/interp/compile
Runtest: 1 tests run, 1 successful, 0 failed + 0 expected
$ /bin/ls -1a tests/interp/compile/
.
..
expected
.gitignore
README
rs274ngc.var
test.sh
use-rs274
use-rs274.cc
$ /bin/ls -1a tests/interp/compile/ | /bin/grep '^[^Retu.]'
rs274ngc.var

Is there any chance that the file with the strange name was created by some other means? If you remove it by hand, is it re-created when you run the test again?

#2 – jepler 于 2016-09-14

I am seeing something similar to @zultron . Debian Jessie, 2.7 branch at

70a9a36 Docs: fix example scrips so they work when copied and pasted.

#3 – jepler 于 2016-09-14

$ strace -ff -e open ./use-rs274  
....
open("h;\330j\177", OWRONLY|OCREAT|O_TRUNC, 0666) = 4

#4 – jepler 于 2016-09-14

valgrind:

Syscall param open(filename) points to uninitialised byte(s)
  at 0x5ABAC00: _opennocancel (syscall-template.S:81)
  by 0x5A53CA1: IOfile_open (fileops.c:227)
  by 0x5A53E19: IOfilefopen@@GLIBC2.2.5 (fileops.c:332)
  by 0x5A49063: _fopeninternal (iofopen.c:90)
  by 0x4F69708: Interp::saveparameters(char const, double const) (rs274ngcpre.cc:1860)
  by 0x4F6A33B: Interp::synch() (rs274ngc_pre.cc:1962)
  by 0x4F6CDB9: Interp::init() (rs274ngc_pre.cc:1163)
  by 0x4029DD: main (in /home/jepler/src/linuxcnc/tests/interp/compile/use-rs274)
Address 0xffeffdc81 is on thread 1's stack
in frame #5, created by Interp::synch() (rs274ngc_pre.cc:1928)

#5 – zultron 于 2016-09-14

I haven’t tested, but sweet! Looks like @jepler nailed it. Thanks fellas.