*Synopsis*: ksh93 crash with static IFS
Due to a change requested by <User 1-5Q-1267>,
this CR is being redispatched:
CR 6990608 changed on Oct 12 2010 by <User 1-5Q-1267>
=== Field ============ === New Value ============= === Old Value =============
Public Comments New Note
Status 1-Dispatched 11-Closed
SubStatus Not Reproducible
====================== =========================== ===========================
*Change Request ID*: 6990608
*Synopsis*: ksh93 crash with static IFS
Product: solaris
Category: shell
Subcategory: korn93
Type: Defect
Subtype:
Status: 1-Dispatched
Substatus:
Priority: 3-Medium
Introduced In Release:
Introduced In Build:
Responsible Engineer:
Keywords: opensolaris
=== *Description* ============================================================
Category
shell
Sub-Category
korn93
Description
The test case here crashes /bin/ksh93:
=================================
function printarray
{
typeset mode=$1
nameref ar=$2
typeset -S IFS # static variable
if [[ $mode == 'set' ]] ; then
IFS=';'
elif [[ $mode == 'unset' ]] ; then
unset IFS
fi
print -- "${ar[*]}"
}
function run
{
typeset -a data=(
'aaa'
'bbb'
'ccc'
'ddd'
)
printarray set data
printarray print data
printarray unset data
print -- "${data[*]}"
printarray set data
printarray print data
printarray unset data
print -- "${data[*]}"
}
run
=================================
gdb is less than useful with optimized code:
(gdb) bt
#0 0x00000001001199f0 in ?? ()
#1 0xffffffff7d5535ec in sh_macexpand () from /usr/lib/64/libshell.so.1
#2 0xffffffff7d538370 in sh_argbuild () from /usr/lib/64/libshell.so.1
#3 0xffffffff7d582488 in sh_exec () from /usr/lib/64/libshell.so.1
#4 0xffffffff7d582404 in sh_exec () from /usr/lib/64/libshell.so.1
#5 0xffffffff7d588894 in sh_funscope () from /usr/lib/64/libshell.so.1
#6 0xffffffff7d588cfc in sh_funct () from /usr/lib/64/libshell.so.1
#7 0xffffffff7d583ea4 in sh_exec () from /usr/lib/64/libshell.so.1
#8 0xffffffff7d5823d8 in sh_exec () from /usr/lib/64/libshell.so.1
#9 0xffffffff7d588894 in sh_funscope () from /usr/lib/64/libshell.so.1
#10 0xffffffff7d588cfc in sh_funct () from /usr/lib/64/libshell.so.1
#11 0xffffffff7d583ea4 in sh_exec () from /usr/lib/64/libshell.so.1
#12 0xffffffff7d55c87c in exfile () from /usr/lib/64/libshell.so.1
#13 0xffffffff7d55bc44 in sh_main () from /usr/lib/64/libshell.so.1
#14 0x0000000100000d60 in main ()
Frequency
Always
Regression
no
Steps to Reproduce
Run testcase
Expected Result
aaa;bbb;ccc;ddd
aaa;bbb;ccc;ddd
aaa bbb ccc ddd
aaa bbb ccc ddd
aaa;bbb;ccc;ddd
aaa;bbb;ccc;ddd
aaa bbb ccc ddd
aaa bbb ccc ddd
Actual Result
crash, coredump
Error Message(s)
coredump
Test Case
function printarray
{
typeset mode=$1
nameref ar=$2
typeset -S IFS # static variable
if [[ $mode == 'set' ]] ; then
IFS=';'
elif [[ $mode == 'unset' ]] ; then
unset IFS
fi
print -- "${ar[*]}"
}
function run
{
typeset -a data=(
'aaa'
'bbb'
'ccc'
'ddd'
)
printarray set data
printarray print data
printarray unset data
print -- "${data[*]}"
printarray set data
printarray print data
printarray unset data
print -- "${data[*]}"
}
run
Workaround
Make typeset -S a global variable
Additional configuration information
Solaris 11, tosiba laptop
*** (#1 of 2): 2010-10-08 14:00:59 GMT+00:00 <User 1-9HOB9N>
I don't see this on a system running build 150.
*** (#2 of 2): 2010-10-08 14:22:36 GMT+00:00 <User 1-5Q-4611>
=== *Public Comments* ========================================================
Sent a mail to the submitter asking which build he saw this problem on.
Haven't heard from him.
I am closing this CR as not reproducible.
As AlanC pointed out build 134 was the last build community have access
to, so I suggest, please check if you see this issue with 134, if so
discuss with ksh93 community to figure which CR fixed this issue
and make those chanes in your workpace.
*** (#1 of 2): 2010-10-12 06:21:05 GMT+00:00 <User 1-5Q-4611>
ksh93 project lead Roland Mainz reports:
I can reproduce the bug, but it is not easy (understatement) to hit it.
The bug here is a memory corruption which depends on precise memory
layout and allocation/deallocation sequence. Depending on what's in
the environment variables and other things like layout of the script
you may hit it or not.
The good news is that newer ksh93 versions no longer have that, this
was likely killed when we tightened the testing for static variables
(that's why I wrote the test module in
http://svn.genunix.org/repos/on/branches/ksh93/gisburn/scripts/tests/sun_solaris_staticvariables.sh
and later enhanced it a lot to drive out all the crazy bugs hiding
there...).
*** (#2 of 2): 2010-10-12 15:13:07 GMT+00:00 <User 1-5Q-1267>
=== *Workaround* =============================================================
=== *Additional Details* =====================================================
Targeted Release:
Commit To Fix In Build:
Fixed In Build:
Integrated In Build:
Verified In Build:
See Also:
Duplicate of:
Hooks:
Hook1:
Hook2:
Hook3:
Hook4:
Hook5:
Hook6: <email address omitted>
Program Management:
Root Cause:
Fix Affects Documentation: No
Fix Affects Localization: No
=== *History* ================================================================
Date Submitted: 2010-10-08 14:00:58 GMT+00:00
Submitted By: <User 1-9HOB9N>
Status Changed Date Updated Updated By
11-Closed 2010-10-12 06:21:05 GMT+00:00 <User 1-5Q-4611>
1-Dispatched 2010-10-12 15:13:06 GMT+00:00 <User 1-5Q-1267>
=== *Service Request* ========================================================
Impact: Critical
Functionality: Primary
Severity: 1
Product Name: solaris
Product Release: solaris_nevada
Product Build: snv_01
Operating System: solaris_nevada
Hardware: generic
Submitted Date: 2010-10-08 14:00:59 GMT+00:00
=== *Multiple Release (MR) Cluster* - 0 ======================================
_______________________________________________
ksh93-integration-discuss mailing list
ksh93-integration-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/ksh93-integration-discuss
