Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations derfloh on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!
AIX 4.3.1 to 5.1 upgrade
- Thread starter promotor
- Start date
![[noevil]](/data/assets/smilies/noevil.gif "[noevil] [noevil]")
There is a major gotcha in the way ksh works in 5.2 compared to 4.3.3 (and most other popular flavours of unix come to think of it). It comes down to variable substitution within brackets.
On 4.3.3 the following returns as you would expect and displays "ok"
valid_envs="bst|cid|cmb|per|sys"
this_env=per
[[ $this_env = @($valid_envs) ]] && print ok
On aix 5L (5.1 or 5.2) this returns nothing.
A little futher testing shows there is also a problem (related) where a shell variable is substituted in a !( ) construct
$ y='xxx|yyy'
$ [[ xxx = !($y) ]] && print huh
on 4.3 this does not print huh, which is the correct action. On 5.2 it does print huh which is wrong.
We have an on-going call with IBM where they are maintaining that you cannot have variable substitution within the brackets. I am aware of work-arounds but when porting an application that uses ksh from 4.3.3 or any other unix to 5L, the last thing you should have to do is to modify your shell scripts to cope with this glorious "backwards enhancement" that 5L has to offer.
The best of it is that IBM are maintaining it is a feature to behaive as it does, even though it deviates AIX 5L from some of the open standards of other unix environments!
Any one else hit this bug yet, or have any thoughts on the subject?
On 4.3.3 the following returns as you would expect and displays "ok"
valid_envs="bst|cid|cmb|per|sys"
this_env=per
[[ $this_env = @($valid_envs) ]] && print ok
On aix 5L (5.1 or 5.2) this returns nothing.
A little futher testing shows there is also a problem (related) where a shell variable is substituted in a !( ) construct
$ y='xxx|yyy'
$ [[ xxx = !($y) ]] && print huh
on 4.3 this does not print huh, which is the correct action. On 5.2 it does print huh which is wrong.
We have an on-going call with IBM where they are maintaining that you cannot have variable substitution within the brackets. I am aware of work-arounds but when porting an application that uses ksh from 4.3.3 or any other unix to 5L, the last thing you should have to do is to modify your shell scripts to cope with this glorious "backwards enhancement" that 5L has to offer.
The best of it is that IBM are maintaining it is a feature to behaive as it does, even though it deviates AIX 5L from some of the open standards of other unix environments!
Any one else hit this bug yet, or have any thoughts on the subject?
ANUK,
did some testing myself. your test works fine using /usr/bin/ksh93 on AIX 5.1.0-03, but not /usr/bin/ksh. not sure what would be the proper way to workaround here, since a workaround might break other things. just keep pestering IBM, they relented on the 99-fields limit on awk so maybe they will fix this too.
IBM Certified -- AIX 4.3 Obfuscation
did some testing myself. your test works fine using /usr/bin/ksh93 on AIX 5.1.0-03, but not /usr/bin/ksh. not sure what would be the proper way to workaround here, since a workaround might break other things. just keep pestering IBM, they relented on the 99-fields limit on awk so maybe they will fix this too.
IBM Certified -- AIX 4.3 Obfuscation
cryptogro4
MIS
This is a known problem with ksh and ksh93. Given that, I am not sure there is much IBM could do. I would tend to think that the problem would have to be fixed in the ksh code itself, which, of course, IBM doesn't maintain; they are a mere licensee.
Only if IBM put pressure on the maintainer would they fix the problem, and I am not sure IBM would consider this a large enough problem to expend that much effort.
Only if IBM put pressure on the maintainer would they fix the problem, and I am not sure IBM would consider this a large enough problem to expend that much effort.
Hi guys
I am goin to upgrade a IBM Model 7026-H70 from AIX 4.3 to 5.2. I have done testing in my labs and was successful with some issues which i managed to solve. The main upgrade im goin to do involves Oracle. Did anyone face any problems upgrading a AIX 4.3 server with Oracle to 5.2. Any help on this will be greatly appreciated.
Thanks in Advance
I am goin to upgrade a IBM Model 7026-H70 from AIX 4.3 to 5.2. I have done testing in my labs and was successful with some issues which i managed to solve. The main upgrade im goin to do involves Oracle. Did anyone face any problems upgrading a AIX 4.3 server with Oracle to 5.2. Any help on this will be greatly appreciated.
Thanks in Advance
- Status
Similar threads
- Locked
- Question