bash evaluation change...

[Nathanael D. Noblet]

On 03/03/2011 10:18 AM, Nathanael D. Noblet wrote:
> On 03/02/2011 11:55 PM, Toshio Kuratomi wrote:
>
>> I'd say, report this as a bug and see where it leads.  It's certainly not
>> expected behaviour even if it is intentional.
>
>
> https://bugzilla.redhat.com/show_bug.cgi?id=681940
>
>
> as well I submitted to bash upstream with bashbug which submits to a
> mailling list of all things...


I can't submit using the bashbug-32 from mock... here's what I was 
*trying* to submit if anyone feels like submitting to bash upstream for 
me? I will also put this in the bug incase the bash maintainer has an 
upstream contact.

Subject: change in case evaluation handling

Configuration Information [Automatically generated, do not change]:
Machine: i386
OS: linux-gnu
Compiler: gcc
Compilation CFLAGS:  -DPROGRAM='bash' -DCONF_HOSTTYPE='i386' 
-DCONF_OSTYPE='linux-gnu' -DCONF_MACHTYPE='i386-redhat-linux-gnu' 
-DCONF_VENDOR='redhat' -DLOCALEDIR='/usr/share/locale' -DPACKAGE='bash' 
-DSHELL -DHAVE_CONFIG_H -DDEBUG -DMALLOC_DEBUG -I.  -I. -I./include 
-I./lib  -D_GNU_SOURCE -DRECYCLES_PIDS -D_LARGEFILE_SOURCE 
-D_FILE_OFFSET_BITS=64 -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 
-fexceptions -fstack-protector --param=ssp-buffer-size=4 -m32 
-march=i686 -mtune=atom -fasynchronous-unwind-tables
uname output: Linux iridium.gnat.ca 2.6.35.11-83.fc14.x86_64 #1 SMP Mon 
Feb 7 07:06:44 UTC 2011 i686 i686 i386 GNU/Linux
Machine Type: i386-redhat-linux-gnu

Bash Version: 4.2
Patch Level: 0
Release Status: rc2

Description:
	The following script illustrates a regression in handline case evaluations

case x"" in
   x) echo 'expected'
      ;;
   x??) echo 'this would be really odd'
      ;;
   x?) echo 'does this prove anything?'
       echo x""x
      ;;
   *) echo 'no match'
      ;;
esac

as output you get:

=>	does this prove anything
	xx

In bash 4.1.X you would get as output 'expected'. The issue was found by 
configure failing with lines like
case x"$enable_feature" in where enable_feature hadn't been passed to 
configure so the variable was undefined.
I should note that handling

if [ x"$some_undefined_var" = x ]; continues to function the same way as 
previous versions

Repeat-By:
	using the following script (named test_script for me):
======================================
#!/bin/sh

case x"$mytest" in
  xyes)
   echo "yes"
   ;;
  xno)
   echo "no"
   ;;
  x)
   echo "empty"
   ;;
  *)
   echo "not supported"
   ;;
esac

mytest=$1
case x"$mytest" in
  xyes)
   echo "yes"
   ;;
  xno)
   echo "no"
   ;;
  x)
   echo "empty"
   ;;
  *)
   echo "not supported"
   ;;
esac

exit;
=========================================

and call it with

sh test_script yes
sh test_script no
sh test_script whatever
sh test_script

4.1.X has the following output:

- sh test_script yes
empty
yes

- sh test_script no
empty
no

- sh test_script whatever
empty
not supported

- sh test_script
empty
empty

4.2.0 has the following output:
- sh test_script yes
not supported
yes

- sh test_script no
not supported
no

- sh test_script whatever
not supported
not supported

- sh test_script
not supported
not supported