i am writing a simple script in powershell using "read-host -prompt" so that i can input some text that will then be used to create some environment variables.
Is there a way that the subsequent times the script is run, the read-host message will show an existing value for the variable if it exists ? and if not what i want then accept my new input to change that variable ?
for example ...
$myvar = read-host -prompt "whats your value" and i enter 10 to set $myvar value to 10
next time the script is run, "whats your value" will show 10, and if i press enter without changing the value it will use the value 10 again .. if i enter a new value it will update $myvar to the new value
Thanks for any help
If I understand correctly, you're looking for two features that are not implemented in Read-Host as of PowerShell 7.1:
(a) Pre-filling the edit buffer with a default value, which the user can either accept as-is or modify.
The alternative is just to inform the user, via the -Prompt string, what value will be used if they don't enter a new one, but that way you won't be able to distinguish between the user choosing the default value or simply wanting to abort the prompt (which they could do with Ctrl-C, however).
(b) Keeping a persistent history of user-entered values that remembers (at least) the most recently entered value, across PowerShell sessions.
Note: Read-Host is currently bare-bones. The module that provides the rich interactive command-line editing experience for PowerShell itself is PSReadLine, and would be great if its features, which include a persistent history and modifying the edit buffer, could be made available to user code for general-purpose prompting - see GitHub proposal #881.
Surfacing such enhancements via Read-Host is probably the best option, or at least the ability to prefill the edit buffer could be implemented there: see GitHub proposal #14013.
See - limited - custom implementations of (a) and (b) below.
(a) is currently only possibly via a workaround, and only on Windows, both in regular console windows and Windows Terminal (it does not work in the obsolescent PowerShell ISE Thanks, CFou., and in Visual Studio Code's integrated terminal it only works if you place the focus on it by clicking immediately after launching a debug session):
# The (default) value to pre-fill the Read-Host buffer with.
$myVar = 'This is a default value.'
# Workaround: Send the edit-buffer contents as *keystrokes*
# !! This is not 100% reliable as characters may get dropped, so we send
# !! multiple no-op keys first (ESC), which usually works.
(New-Object -ComObject WScript.Shell).SendKeys(
'{ESC}' * 10 + ($myVar -replace '[+%^(){}]', '{$&}')
)
$myVar = Read-Host 'Enter a value' # Should display prompt with value of $myVar
Note: The -replace operation is necessary to escape characters in the default value that would otherwise have special meaning to .SendKeys().
(b) requires you to implement your own persistence mechanism, and the obvious choice is to use a file:
Here's a simplistic approach that only stores the most recent value entered.
Supporting multiple historic values per prompt would also support for recall in Read-Host, such as as using up-arrow and down-arrow to cycle through the history, which is not supported as of PowerShell 7.1.
# Choose a location for the history file.
$historyFile = "$HOME/.rhhistory"
# Read the history file (which uses JSON), if it exists yet.
$history = Get-Content -Raw -ErrorAction Ignore $historyFile | ConvertFrom-Json
$defaultValue = 'This is a default value.'
# Get the 'myVar' entry, if it exists, otherwise create it and use the default value.
$myVar =
if (-not $history) { # no file yet; create the object to serialize to JSON later
$history = [pscustomobject] #{ myVar = '' }
$defaultValue
} elseif (-not $history.myVar) { # file exists, but has no 'myVar' entry; add it.
$history | Add-Member -Force myVar ''
$defaultValue
} else { # return the most recently entered value.
$history.myVar
}
# Prompt the user.
(New-Object -ComObject WScript.Shell).SendKeys(
'{ESC}' * 10 + ($myVar -replace '[+%^(){}]', '{$&}')
)
$myVar = Read-Host 'Enter a value'
# Validate the value...
# Update the history file with the value just entered.
$history.myVar = $myVar
$history | ConvertTo-Json > $historyFile
You cannot do this straight :
####[int]$myVar = 10 Read too fast the question :)
#$myVar = Get-ItemProperty ...
#or
#myVar = Get-Content ...
$tempVar = Read-Host "Enter the value ($myVar is default)"
if ($tempVar)
{
$myVar = $tempVar
}
$myVar
# Store $myVar in the registry or in the file for next run
My bad, I answered without a good reading :) As #postanote said,
If you need to keep a track of the value from each run, you can store it in a file (IMHO, I would avoid the profile.ps1) or in a personal registry key.
Then read the file or the registry value to initialize your variable.
Related
I have a VB application, which starts several instances of a third party non-GUI application. To keep track of these multiple instances, I update their title, using the SetWindowText() function. This application however has the nasty habit of continuously updating the title, so each SetWindowText works only temporary. As soon as you click anywhere in the screen, the tile is changed back.
I found a way to update the title through PowerShell, using the following code:
$titletext = "My Title"
# Start a thread job to change the window title to $titletext
$null = Start-ThreadJob { param( $rawUI, $windowTitle )
Start-Sleep -s 2
if ( $rawUI.WindowTitle -ne $windowTitle ) {
$rawUI.WindowTitle = $windowTitle
}
}-ArgumentList $host.ui.RawUI, $titletext
& 'c:\Program Files\Application\Application.exe' '-id=userid -pass=password'
This works perfectly and the title change is permanent, so exactly what I want. The only problem is that everything is being logged in the Windows PowerShell log, including the parameters -id= and -pass=.
A solution would be if I can start application.exe through my VB application and do the rename through a PowerShell script, but I don't know if that is possible through a ThreadJob.
Is it possible to start a ThreadJob and rename another window, maybe through it's handle?
Changing the console title from inside that console is your best bet, which is what your PowerShell code does.
While it is possible to call the SetWindowText() API function to set another process' console-window title, this change isn't guaranteed to stay in effect, because any subsequent interaction with such a window causes the original window title to be restored (this behavior seems to be built into conhost.exe, the console host underlying regular console windows on Windows).
By contrast, setting the title of the console window associated with the current process, does stay in effect (unless overridden again later), which is what the SetConsoleWindow() WinAPI function does (which shell- and API-based mechanisms such as title in cmd.exe, and [Console]::Title / $hostUI.RawUI.WindowTitle in PowerShell presumably ultimately call).
Therefore, stick with your PowerShell approach and avoid the password-logging problem with the help of an environment variable, as detailed below.
Windows PowerShell's script-block logging - see about_Logging - logs the source code of code being created.
You can avoid argument values from being logged if you - instead of providing literal arguments - provide them indirectly, via variables that you set from outside PowerShell.
Therefore:
Make your VB.NET application (temporarily) set an environment variable that contains the password. (Perhaps needless to say, storing and passing plain-text passwords is best avoided).
In your PowerShell script, refer to that environment variable instead of passing a literal password - that way, the actual password will not be shown in the logs.
For example, assuming that your VB.NET application has created environment variable MYPWD containing the password, before launching the PowerShell script:
$titletext = "My Title"
# Start a thread job to change the window title to $titletext
$null = Start-ThreadJob { param( $rawUI, $windowTitle )
Start-Sleep -s 2
if ( $rawUI.WindowTitle -ne $windowTitle ) {
$rawUI.WindowTitle = $windowTitle
}
} -ArgumentList $host.ui.RawUI, $titletext
# Note:
# * Assumes that your VB.NET application has set env. var. "MYPWD".
# * The arguments must be passed *individually*, not inside a single string.
& 'c:\Program Files\Application\Application.exe' -id=userid "-pass=$env:MYPWD"
Disclaimer : I am the epitome of a scipting/Powershell rookie, so please bear with me.
I've written a script to return the Active Directory username of any user currently logged into a given workstation.
$input = Read-Host "Workstation Name"
$domain = ".*****.***.com"
$computer = $input + $domain
$list = gwmi win32_computersystem -comp $computer | select Username,Caption
Write-Output $list
However, if I run this from a pinned script in the taskbar, the Powershell window closes before I have a chance to view the results.
I have tried method 2 and 3 from this post, but to no avail. Method 2 prompts for user input before the results are displayed instead of after, even when the code for the prompt is added at the end of the script.
Any help would be greatly appreciated.
Method 2 from the linked post - i.e., waiting for the user to press a key before exiting the script - can be used, but it requires additional effort:
End your script as follows in order to see the value of $list before the pause command prompts:
$list | Out-Host # Force *synchronous* to-display output.
pause # Wait for the user to press Enter before exiting.
Note: pause in PowerShell is simply a function wrapper around Read-Host as follows: $null = Read-Host 'Press Enter to continue...' Therefore, if you want to customize the prompt string, call Read-Host directly.
This answer explains why the use of Out-Host (or Format-Table) is necessary in this case; in short:
In PSv5+, an implicitly applied Format-Table command asynchronously waits for up to 300 msecs. for additional pipeline input, in an effort to derive suitable column widths from the input data.
Because you use Write-Output output objects without predefined formatting data that have 2 properties (4 or fewer ), tabular output is implicitly chosen, and Format-Table is used behind the scenes, asynchronously.
Note: The asynchronous behavior applies only to output objects for whose types formatting instructions aren't predefined (as would be reported with Get-FormatData <fullOutputTypeName>); for instance, the output format for the System.Management.Automation.AliasInfo instances output by Get-Alias is predefined, so Get-Alias; pause does produce output in the expected sequence.
The pause command executes before that waiting period has elapsed, and only after you've answered the prompt does the table print, after which point the window closes right away.
The use of an explicit formatting command (Out-Host in the most generic case, but any Format-* cmdlet will do too) avoids that problem by producing display output synchronously, so that the output will be visible by the time pause displays its prompt.
I had the same problem for scripts that I'm executing "on demand". I tend to simply add a Read-Host at the end of the script like so
$str = "This text is hardly readable because the console closes instantly"
Write-Output $str
Read-Host "Script paused - press [ENTER] to exit"
Disclaimer : I am the epitome of a scipting/Powershell rookie, so please bear with me.
I've written a script to return the Active Directory username of any user currently logged into a given workstation.
$input = Read-Host "Workstation Name"
$domain = ".*****.***.com"
$computer = $input + $domain
$list = gwmi win32_computersystem -comp $computer | select Username,Caption
Write-Output $list
However, if I run this from a pinned script in the taskbar, the Powershell window closes before I have a chance to view the results.
I have tried method 2 and 3 from this post, but to no avail. Method 2 prompts for user input before the results are displayed instead of after, even when the code for the prompt is added at the end of the script.
Any help would be greatly appreciated.
Method 2 from the linked post - i.e., waiting for the user to press a key before exiting the script - can be used, but it requires additional effort:
End your script as follows in order to see the value of $list before the pause command prompts:
$list | Out-Host # Force *synchronous* to-display output.
pause # Wait for the user to press Enter before exiting.
Note: pause in PowerShell is simply a function wrapper around Read-Host as follows: $null = Read-Host 'Press Enter to continue...' Therefore, if you want to customize the prompt string, call Read-Host directly.
This answer explains why the use of Out-Host (or Format-Table) is necessary in this case; in short:
In PSv5+, an implicitly applied Format-Table command asynchronously waits for up to 300 msecs. for additional pipeline input, in an effort to derive suitable column widths from the input data.
Because you use Write-Output output objects without predefined formatting data that have 2 properties (4 or fewer ), tabular output is implicitly chosen, and Format-Table is used behind the scenes, asynchronously.
Note: The asynchronous behavior applies only to output objects for whose types formatting instructions aren't predefined (as would be reported with Get-FormatData <fullOutputTypeName>); for instance, the output format for the System.Management.Automation.AliasInfo instances output by Get-Alias is predefined, so Get-Alias; pause does produce output in the expected sequence.
The pause command executes before that waiting period has elapsed, and only after you've answered the prompt does the table print, after which point the window closes right away.
The use of an explicit formatting command (Out-Host in the most generic case, but any Format-* cmdlet will do too) avoids that problem by producing display output synchronously, so that the output will be visible by the time pause displays its prompt.
I had the same problem for scripts that I'm executing "on demand". I tend to simply add a Read-Host at the end of the script like so
$str = "This text is hardly readable because the console closes instantly"
Write-Output $str
Read-Host "Script paused - press [ENTER] to exit"
I am trying to print a PDF as XPS, the script opens the PDF Print Output As screen and enters the correct name before sending the ENTER command to the window to save the file.
How can I select the address bar to enter the desired path? Or how can I change the default save path?
EDIT: Thank you for the feedback. Here is the script:
function print_files($secure_pdf_dir){
#Retrieves the name for the .xps files
Get-ChildItem $secure_pdf_dir -Filter *.pdf -Recurse | Foreach-Object {
#For each .pdf file in that directory, continue
same_time $_.FullName
}
}
## The following function keeps checking for a new window called "Save Print Output As"
## When the window shows up, it enters the name of the file and press ENTER
function enter_my_names($xps_dir, $fullname){
$wshell = New-Object -ComObject wscript.shell;
while($wshell.AppActivate('Save Print Output As') -ne $true){
$wshell.AppActivate('Save Print Output As')
}
$basename = [io.path]::GetFileNameWithoutExtension($fullname)
#This is where the name is actually entered
$wshell.SendKeys($xps_dir\$basename)
$wshell.SendKeys("{ENTER}")
}
## The following function launches simultaneously a print job on the input file
## and a function waiting for the print job to show up to name the file
workflow same_time{
Param(
$fullname
)
parallel{
Start-Process -FilePath $fullname –Verb Print -PassThru
enter_my_names $xps_dir $fullname
}
}
#MAIN PROGRAM
#Here the script saves your current printer as default
$defprinter = Get-WmiObject -Query "Select * from Win32_Printer Where Default=$true"
#Queries for a XPS printer
$printer = Get-WmiObject -Query "Select * from Win32_Printer Where Name='Microsoft XPS Document Writer'"
#Sets the XPS printer as Default
$printer.SetDefaultPrinter()
#Starts the main job
print_files($secure_pdf_dir)
#Sets the old default printer back as default again
$defprinter.SetDefaultPrinter()
#This is a small delay to be sure everything is completed before closing Adobe Reader. You can probably shorten it a bit
sleep 2
#Finally, close Adobe Reader
Get-Process "acrord32" | Stop-Process
It seems that the $xps_dir variable is not being passed into the function properly.
Edit: Error I get when trying to add $xps_dir to the enter_my_names function:
Microsoft.PowerShell.Utility\Write-Error : Cannot validate argument on parameter
'FilePath'. The argument is null or empty. Provide an argument that is not null or
empty, and then try the command again.
At same_time:115 char:115
+
+ CategoryInfo : NotSpecified: (:) [Write-Error], ParameterBindingValidati
onException
+ FullyQualifiedErrorId : System.Management.Automation.ParameterBindingValidationEx
ception,Microsoft.PowerShell.Commands.WriteErrorCommand
+ PSComputerName : [localhost]
I'm not sure if you got this fully worked out or I misread your post, but I was able to implement some code from your initial post to finish my project [been beating my head against a wall as a non-traditional comp sci person]; so I wanted to share some things I did in hopes of helping you (in the event you're still at that crossroads).
Some points of interest:
-My overall process is a bulk report generator. A variable amount of .XML stylesheets are sent to IE because a javascript subroutine numbers the reports; to accomplish this I have visual basic doing some database parsing, passing values into custom scripts needed to form the .XML stylesheets, then VB creates PS .ps1 files with the various PS commands needed to create the IE com object. Finally, VB calls the .ps1 scripts, and the reports are generated by sending the stylesheet to IE, waiting for HTML render, send print command, and close the object.
-While your endgoal is to generate an XPS, mine was to fully render an HTML, then print it with a PDF printer; I assume we have a similar process in that we must engage a dialog box in the intermediate stage (in this case, using sendkeys to interact with it); this is specifically what I'm going to discuss below in hopes of helping out!
So, two points of discussion:
First, more of a general observation/query: I don't see where your $xps_dir variable is actually being defined; I mention this because there could be a string-related issue with the values being passed by the $xps_dir. One way to check your string to make sure its 'pretty' is to pass it to a .txt or something using the OUT-FILE command:
"$xps_dir" | Out-File -FilePath "C:\somefolder\sometext.txt" -Append
The text file itself doesn't need to exist, but the file does...the above command will create the text file.
Also, if you're ok with viewing the powershell prompt, a WRITE-HOST can be used.
Either way, I'm wondering if the final compiled string isn't compiled correctly, so its not recognizing it as a directory.
I think what more likely is happening is the SENDKEYS aren't being sent to the correct fields. For example, by using your code I was [FINALLY] able to set my 'Save As' dialog box as an object and interact with it. (When using my PDF printer, I get the 'Save As' dialog box).
At first, I tried to send the file name, then send enter. I assumed this would be fine, because during manual interaction the focus for the 'Save As' dialog box goes to the 'File Name' field; however, after trying:
$wshell.SendKeys($myfilename)
$wshell.SendKeys("{ENTER}")
I realized the SendKeys command was sending the keys to the 'Save in' drop down (or probably the first element of the Save As dialog box). So, I manually click inside of the 'Save in' field, then tabbed around until I got to 1. 'File Name' (wrote the # of tabs down), and 2. the 'Save' button (wrote # of tabs down). By adding SendKeys("{TAB}") to correspond with the number of tabs I observed, I was able to successfully enter my name string and close out the print dialog box:
$wshell.SendKeys("{TAB}")
$wshell.SendKeys("{TAB}")
$wshell.SendKeys("{TAB}")
$wshell.SendKeys("{TAB}")
$wshell.SendKeys("{TAB}")
$wshell.SendKeys($myfilename)
$wshell.SendKeys("{TAB}")
$wshell.SendKeys("{TAB}")
$wshell.SendKeys("{ENTER}")
NOTE: the number of tabs might be different for your dialog box, so I strongly recommend doing a manual run through to count how many tabs get you to the field you want to interact with.
In conclusion, I have one reflection for consideration:
I've used (to much success) VB and PS to use html elements to dynamically enhance my routines. Since we're able to set the dialog box as a Com object, my next goal will be to try to tap into the respective 'elements', similar to how html elements can be accessed (or more generally, object-oriented language allows). This way, I might be able to better interact with the dialog box without suffering the drawbacks of SendKeys:
Has to run as active process; user cannot engage any other windows else keys might be sent to them;
Keys are based on a count; any variation to that count will require the subroutine to be updated/edited, making it not elastic;
Also, I have not done any work regarding error windows, but one I've already noticed is the "your document has the same name blah blah", so just a heads up.
Even though we're doing two different things, here's my final code for this part of my routine in case it helps provide context:
#send command to internet explorer application with execWB; 6 for print, 2 for no user prompt on window
while ( $ie.busy ) { Start-Sleep -second 3 }
$ie.execWB(6,2)
#create dialog box object
$wshell = New-Object -ComObject wscript.shell;
while($wshell.AppActivate('Save As') -ne $true){ Start-Sleep -second 3 }{
$wshell.AppActivate('Save As')
}
#create file string from 3 strings: directory,basename returned, file extension
#the return from my output log is: "c:\test_folder\mybasenamestring.PDF"
$mydirectory = "c:\test_folder\"
$mybasename = [io.path]::GetFileNameWithoutExtension($fullname)
$myextension = ".PDF"
$myfilename = "$mydirectory$mybasename$myextension"
#Using tabs to navigate around the dialog window; send my string to the 'File Name' field, then tab to 'Save' and send enter
#This is where the name is actually entered
$wshell.SendKeys("{TAB}")
$wshell.SendKeys("{TAB}")
$wshell.SendKeys("{TAB}")
$wshell.SendKeys("{TAB}")
$wshell.SendKeys("{TAB}")
$wshell.SendKeys($myfilename)
$wshell.SendKeys("{TAB}")
$wshell.SendKeys("{TAB}")
$wshell.SendKeys("{ENTER}")
#going to add a command to report on the file size of the pdf; the pdf will stay 0kb until its fully generated. Once fully generated will add the $ie.quit or the kill command here for the $ie object
#this sends a stop process command to kill powershell since I'm running it as a ps1 file and it will remain open otherwise
stop-process -Id $PID
Last thing: I observed some weirdness with the capitalization, so I updated the final SendKeys to capitalize the string since I want all caps anyhow:
$wshell.SendKeys($myfilename.toupper())
Hope this helps! Thanks for posting; I was finally able to complete my entire process. As I clean and improve, if I find improvements for this area I'll try to remember to share. Thanks for your help!
Pass the variable containing the desired directory ($xps_dir in this case) to each function in the process until it gets to the enter_my_names function where it can then be sent to the window with the $wshell.SendKeys("$xps_dir\$basename").
function print_files($xps_dir, $secure_pdf_dir){
#Retrieves the name for the .xps files
Get-ChildItem $secure_pdf_dir -Filter *.pdf -Recurse | Foreach-Object {
#For each .pdf file in that directory, continue
same_time $xps_dir $_.FullName
}
}
## The following function keeps checking for a new window called "Save Print Output As"
## When the window shows up, it enters the name of the file and press ENTER
function enter_my_names{
param ($xps_dir, $fullname)
$wshell = New-Object -ComObject wscript.shell;
while($wshell.AppActivate('Save Print Output As') -ne $true){
$wshell.AppActivate('Save Print Output As')
}
$basename = [io.path]::GetFileNameWithoutExtension($fullname)
#This is where the name is actually entered
$wshell.SendKeys("$xps_dir\$basename")
$wshell.SendKeys("{ENTER}")
}
## The following function launches simultaneously a print job on the input file
## and a function waiting for the print job to show up to name the file
workflow same_time{
Param(
$xps_dir, $fullname
)
parallel{
Start-Process -FilePath $fullname –Verb Print -PassThru
enter_my_names $xps_dir $fullname
}
}
#MAIN PROGRAM
#Here the script saves your current printer as default
$defprinter = Get-WmiObject -Query "Select * from Win32_Printer Where Default=$true"
#Queries for a XPS printer
$printer = Get-WmiObject -Query "Select * from Win32_Printer Where Name='Microsoft XPS Document Writer'"
#Sets the XPS printer as Default
$printer.SetDefaultPrinter()
#Starts the main job
print_files $xps_dir $secure_pdf_dir
#Sets the old default printer back as default again
$defprinter.SetDefaultPrinter()
#This is a small delay to be sure everything is completed before closing Adobe Reader. You can probably shorten it a bit
sleep 2
#Finally, close Adobe Reader
Get-Process "acrord32" | Stop-Process
The file dialogue will accept a path in the file name field, so instead of sending the filename file.pdf send the full path to the file C:\folder\folder\file.pdf
EDIT:
You can send the folder path like so:
$wshell.SendKeys($xps_dir)
$wshell.SendKeys("{ENTER}")
$wshell.SendKeys($basename)
$wshell.SendKeys("{ENTER}")
Lets say I have a simple script with one parameter, which has a default value:
param(
[string] $logOutput = "C:\SomeFolder\File.txt"
)
# Script...
And lets say a user runs the script like so:
PS C:\> .\MyScript.ps1 -logOutput "C:\SomeFolder\File.txt"
Is there any way the script is able to know that the user explicitly entered a value (which happens to be the same as the default), rather than let the default be decided automatically?
In this example, the script is going to print some output to the given file. If the file was not specified by the user (i.e. the default got used automatically), then the script will not produce an error in the event it is unable to write to that location, and will just carry on silently. However, if the user did specify the location, and the script is unable to write to that location, then it should produce an error and stop, warning the user. How could I implement this kind of logic?
The simplest way to tell if a parameter was specified or not when you have a default is to look at $PSBoundParameters. For example:
if ($PSBoundParameters.ContainsKey('LogPath')) {
# User specified -LogPath
} else {
# User did not specify -LogPath
}
Would this work for you?
function Test-Param
{
[CmdletBinding()]
param(
[ValidateScript({Try {$null | Set-Content $_ -ErrorAction Stop
Remove-Item $_
$True}
Catch {$False}
})]
[string] $logOutput = 'C:\SomeFolder\File.txt'
)
}
The validate script will only run and throw an error if it is unable to write to the file location passed in using the -logOutput parameter. If the parameter is not specified, the test will not be run, and $logOutput will get set to the default value.