[PS] Script failed due to call depth overflow - Image1.jpg (0/1)

[Andrew Watt [MVP]]

I received the error message in the post's title when exploring the
$MaximumHistoryCount issue.

Here is the full transcript of the MSH session (except for commands 1
and 2):

<begin transcript>
PS C:\Documents and Settings\Andrew Watt> get-history

Id CommandLine
-- -----------
3 $MaximumHistoryCount = 1 ; clear-host

PS C:\Documents and Settings\Andrew Watt> function clear-host{
>> $MaximumHistoryCount = 1
>> clear-host
>> }
>>
PS C:\Documents and Settings\Andrew Watt> clear-host
The script failed due to call depth overflow. The call depth reached
101 and the maximum is 100.
PS C:\Documents and Settings\Andrew Watt>
<end transcript>

I have also attached a screen shot of a repro. On that occasion the
only previous command issued after opening a new PowerShell console
was $MaximumHistoryCount (result was 64).

Does anyone understand what is going on here?

Thanks

Andrew Watt MVP

[Andrew Watt [MVP]]

The original post related to a Windows XP SP2 machine.

I have since tried it on a Windows Server 2003 machine. Same result.

Also, on both machines after getting this error message the PowerShell
console responds to a cls or clear-host with the same message. I can
no longer clear the console's display!

Andrew Watt MVP

[Michael McMullen]

I think you meant to call the function clear-history didn't you? You're
calling clear-host from within a function with the same name, so it's
inifinite recursion being terminated when you go too deep. cls is just an
alias for clear-host, so it does the same thing.

[applepwc]

Last time,I want creat a set-location function with set-location cmdlet in it.
It doesn't work as you seen.
###

The script failed due to call depth overflow. The call depth reached
101 and the maximum is 100.

###

I questioned it,and Jouko answered me:

> How can I reference the cmdlet "set-location" in my script ?

Prefix the cmdlet name with name of the snap-in:

Microsoft.PowerShell.Management\Set-Location

You can determine the snap-in name with:

(Get-Command Set-Location).PSSnapIn.Name
####
So I suggest you use
Microsoft.PowerShell.Management\clear-host

in your function.
--
greeting

applepwc

[Andrew Watt [MVP]]

Thanks. You're right. It's *so* easy to see what you expect to be
there (clear-history) rather than what was there (clear-host).

In any case the function would have needed to be
function clear-history{
$global:MaximumHistoryCount = 1
clear-host
}
to work as I intended.

Andrew Watt MVP

[applepwc]

I'm sorry for my last post :
"So I suggest you useMicrosoft.PowerShell.Management\clear-host in your
function
"is actually wrong!

I made a big mistake:clear-host is function,not a cmdlet.

So I suggest another method to solve your problem:
1,rename original clear-host function to clear-host2.
Rename-Item function:\clear-host clear-host2

2,invoke clear-host2 in your new clear-host function.
--
greeting

applepwc

[Bruce Payette [MSFT]]

One way to control which command gets executed is to use the result returned
from get-command along with the & operator. For example, say I define a
function "dir".

PS (1) > function dir { "Foobar" }

Now I call get-command

PS (2) > get-command dir

CommandType Name Definition
----------- ---- ----------
Function dir "Foobar"
Alias dir Get-ChildItem

As you can see the token "dir" maps to two possible commands. If I just type
"dir", then I'll get the alias because aliases are resolved before
functions.

PS (3) > dir C:\CONFIG.SYS

Directory: Microsoft.PowerShell.Core\FileSystem::C:\

Mode LastWriteTime Length Name
---- ------------- ------ ----
-a--- 11/17/2004 3:32 AM 0 CONFIG.SYS

So how do I get the function? Well, the CommandInfo object for the function
is the first one returned from get-command command:

PS (4) > (get-command dir)[0]

CommandType Name Definition
----------- ---- ----------
Function dir "Foobar"

Now I call that function using &:

PS (5) > &(get-command dir)[0]
Foobar

This works with any type of command. Let's try it with sc:

PS (6) > get-command sc

CommandType Name Definition
----------- ---- ----------
Alias sc Set-Content
Application sc.exe C:\WINDOWS\system32\...
ExternalScript sc.ps1 C:\Documents and Set...

There are three commands that map to the token "sc". We want the one at
index 1: the exe:

PS (7) > (get-command sc)[1]

CommandType Name Definition
----------- ---- ----------
Application sc.exe C:\WINDOWS\system32\...

Now invoke that command, querying for the plugplay service...

PS (8) > &(get-command sc)[1] query plugplay

SERVICE_NAME: plugplay
TYPE : 20 WIN32_SHARE_PROCESS
STATE : 4 RUNNING
(NOT_STOPPABLE,NOT_PAUSABLE,ACCEP
TS_SHUTDOWN)
WIN32_EXIT_CODE : 0 (0x0)
SERVICE_EXIT_CODE : 0 (0x0)
CHECKPOINT : 0x0
WAIT_HINT : 0x0
PS (9) >

Since CommandInfo objects are just objects, we can store them into a
variable to use later:

PS (10) > &$sc query plugplay

SERVICE_NAME: plugplay
TYPE : 20 WIN32_SHARE_PROCESS
STATE : 4 RUNNING
(NOT_STOPPABLE,NOT_PAUSABLE,ACCEP
TS_SHUTDOWN)
WIN32_EXIT_CODE : 0 (0x0)
SERVICE_EXIT_CODE : 0 (0x0)
CHECKPOINT : 0x0
WAIT_HINT : 0x0

When invoked in this way, no lookup is done, the command is directly
executed. If you're a UNIX shell user, this roughly corresponds to the idea
of a "tracked alias" except that we don't currently allow you to bind a
CommandInfo object to an alias (yet). Next release perhaps :-)

-bruce

--
Bruce Payette [MSFT]
Windows PowerShell Technical Lead
Microsoft Corporation
This posting is provided "AS IS" with no warranties, and confers no rights.

"applepwc" <appl...@msn.com> wrote in message
news:8906342F-84AE-4E6F...@microsoft.com...