Roderich:
> As your experiments show the most likely cause is the custom
> perl interpreter used by the packed executable or the few
> "essential" modules that it uses during bootstrap. Did you
> build PAR::Packer yourself or installed it vis ppm?
> If the latter: can you try to build PAR::Packer from source
> (including running "dmake test"), install it, then re-pack bug2.pl
> and try the resulting executable?
I had before reporting this problem, installed PAR packer from source
in the following manner.
Installation of Module-ScanDeps-0.98
===========================================================
Log: truncated:
...
t/14-static_functional_cached.t ................. 1/? 0
Couldn't retrieve data from file deps_cache.dat. Building new cache.
...
All tests successful.
Files=18, Tests=42203, 177 wallclock secs ( 5.27 usr + 0.42 sys = 5.69 CPU)
Result: PASS
===========================================================
Installation of PAR packer 1.008
=======================
During installation, compilation failed during linking of static.c in
myldr complaining
about a non-existent manifest.
So, I manually removed the use of manifest by modifying myldr/Makefile.PL
# Embed the manifest file for VC 2005 (aka VC8) or higher, but not for the
# 64-bit Platform SDK compiler
if( $Config{ivsize} == 4 and $Config{ccversion} =~ /^(\d+)/ and $1 >= 14 ) {
#$mt_cmd = 'mt -nologo -manifest [email protected] -outputresource:$@;1';
$mt_cmd = '-$(NOOP)';
} else {
$mt_cmd = '-$(NOOP)';
}
Based on information in
http://msdn.microsoft.com/en-us/library/dd293574.aspx, I figured out
that manifests are not required. So, I commented out the above line
(mt -nologo -manifest ...).
Testing:
======
t/00-pod.t ............... ok
t/10-parl-generation.t ... ok
t/20-pp.t ................ 31/34 Out of memory!
# Looks like you planned 34 tests but ran 31.
# Looks like your test exited with 1 just after 31.
t/20-pp.t ................ Dubious, test returned 1 (wstat 256, 0x100)
Failed 3/34 subtests
t/30-current_exec.t ...... # Please wait
t/30-current_exec.t ...... ok
t/40-packer_cd_option.t .. ok
t/90-rt59710.t ........... ok
Test Summary Report
-------------------
t/20-pp.t (Wstat: 256 Tests: 31 Failed: 0)
Non-zero exit status: 1
Parse errors: Bad plan. You planned 34 tests but ran 31.
Files=6, Tests=89, 1471 wallclock secs ( 0.02 usr + 0.11 sys = 0.13 CPU)
Result: FAIL
Failed 1/6 test programs. 0/89 subtests failed.
NMAKE : fatal error U1077: 'C:\Perl\bin\perl.exe' : return code '0xff'
Since one test failed, I just did
nmake install
to try it out.
I repeated the same steps again and verify that the problem (crash)
continues to occur.
Thanks,
-Ani
