I have a batch script on my computer called cs.bat. When I enter cs in the command prompt, pushd takes me to a certain directory and leaves me there. In PowerShell, the command does the same thing but then brings me back into the starting directory.
Why is this the case? How can I make it so that I stay in the directory after typing 'cs' into power shell?
Powershell includes aliases for Pushd and Popd.
Get-Alias Pushd : pushd -> Push-Location
Get-Alias Popd : popd -> Pop-Location
You can then use Get-Help Push-Location -Full -Online to get the latest help for that cmdlet.
Then just make a script and test this behavior.
#Sample.ps1 script
#Get current DIR
dir
#push location to some location and DIR there.
Push-Location C:\Scripts\
dir
#at this point, your console will still be in the Push-Location directory
#simply run the Pop-Location cmdlet to switch back.
This is happening because your "cs.bat" runs in a different process (running cmd.exe) spawned by PowerShell (whereas batch files execute in the same instance when run from cmd). Current directory is a per-process concept, so changing it in one process has no effect on another.
Probably the simplest way to get around it is to write a "cs.ps1" script (or function), that would run in the PowerShell process.
Related
I have quite strange behavior happening on my powershell code. I pinned it down to these simple reproduceable steps.
Launch Powershell as Administrator.
Type cmd and confirm it opens cmd on the current Powershell terminal. Type exit to get out of the cmd console.
Run the following code:
Import-Module WebAdministration
cd iis:\apppools
new-item c:\temp\1 -itemType directory
Push-Location
cd c:\temp\1
Pop-Location
remove-item c:\temp\1
Type cmd and now you see it opens cmd on a new window.
Run cd c:\
Type cmd and now you see it opens cmd on a current window as expected.
Run cd iis:\apppools, and then cmd, and you see it opens on a current window as expected.
Why did #4 open in new window, and then why does cd c:\ fix the problem?
The real problem that I am dealing with is that when the console gets into the state of step #4, it launches any console command in a new window, so while I would expect the script to wait until the console command is finished, it launches the console command in a new window and move on to the next line without waiting for the console to finish.
Reproduceable on Powershell 5.1.14393.4350 (Windows 2016), 5.1.18362.1171 (Windows 10) and 5.1.19041.906 (Windows 10).
#Daniel's comment had the answer, and I am writing it here to comply with StackOverflow's rule:
When calling cmd from powershell, it tries to set path to the current path that you are in. However, iis: (env:, alias:, cert:, etc) are not available paths in cmd. So it will try to fallback to the latest folder you were in under cmd-accessible path (C:, D:\ etc.)
And when the folder does not exist, it appears that the failsafe for this in the code is to open a real cmd window instead.
This can be replicated without IIS, push-location and pop-location.
mkdir c:\temp\deleteme;
cd c:\temp\deleteme;
cd env:;
rm c:\temp\deleteme -Recurse;
cmd
#Daniel's comment had the answer:
When calling cmd from powershell, it tries to set path to the current path that you are in. However, iis: (env:, alias:, cert:, etc) are not available paths in cmd. So it will try to fallback to the latest folder you were in under cmd-accessible path (C:, D:\ etc.)
And when the folder does not exist, it appears that the failsafe for this in the code is to open a real cmd window instead.
This can be replicated without IIS, push-location and pop-location.
mkdir c:\temp\deleteme;
cd c:\temp\deleteme;
cd env:;
rm c:\temp\deleteme -Recurse;
cmd
I am trying to do the following steps from Powershell. Following is just a pseudo-code which explains my requirements.
# This is a powershell function
function load(){
cd C:\my_path\scripts
invoke-expression -Command C:\Program Files\Git\git-bash.exe
# I want to go to this path into the git-bash.exe window
cd C:\my_path\scripts
# I have bash script here. I want to excuete this script.
./loadData.sh
}
how can I achieve this by Powershell? Thanks!
try git-bash.exe --help to see what parameters it has. after a short google search i assume you can run Start-Process -FilePath "C:\Program Files\Git\git-bash.exe" -ArgumentList '--cd="C:\my_path\scripts" --exec="loadData.sh"' maybe even Start-Process -FilePath "C:\Program Files\Git\git-bash.exe" -ArgumentList '--exec="C:\my_path\scripts\loadData.sh"' might work.
If you simply want to execute the bash script loadData.sh, run :
cd C:\my_path\scripts
path/to/bash.exe loadData.sh
This will create a bash shell (in the current console, not in a separate one), run your script, and exit, returning to your powershell.
If the current working directory is not set as you expect within your bash shell, you can perhaps pass it as an argument to your script :
# not 100% tested, I don't have a Powershell at hand
path/to/bash.exe loadData.sh /c/my_path/scripts
# and in your loadData.sh, add an instruction :
cd "$1"
I can open PowerShell window in any directory using Windows File Explorer.
I want to run a script every time a new PowerShell window is open and use current directory where it was open in the script.
Using $profile let me for automatic script execution but $pwd variable does not have directory used to open PowerShell window but has C:\WINDOWS\system32. I understand PowerShell starts in C:\WINDOWS\system32, run $profile and next change location used with File Explorer. How can I get file explorer current directory it when my script is executes from $profile or maybe there is another way to automatic execute my script after PowerShell window is open?
Note: The answer below provides a solution based on the preinstalled File Explorer shortcut-menu commands for Window PowerShell.
If modifying these commands - which requires taking ownership of the registry keys with administrative privileges - or creating custom commands is an option, you can remove the NoWorkingDirectory value from the following registry keys (or custom copies thereof):
HKEY_CLASSES_ROOT\Directory\shell\Powershell
HKEY_CLASSES_ROOT\Directory\Background\shell\Powershell
Doing so will make the originating folder the working directory before PowerShell is invoked, so that $PROFILE already sees that working directory, as also happens when you submit powershell.exe via File Explorer's address bar.[1]
Shadowfax provides an important pointer in a comment on the question:
When you hold down Shift and then invoke the Open PowerShell window here shortcut-menu command on a folder or in the window background in File Explorer, powershell.exe is initially started with C:\Windows\System32 as the working directory[1], but is then instructed to change to the originating folder with a Set-Location command passed as a parameter; e.g., a specific command may look like this:
"PowerShell.exe" -noexit -command Set-Location -literalPath 'C:\Users\jdoe'
As an aside: The way this shortcut-menu command is defined is flawed, because it won't work with folder paths that happen to contain ' chars.
At the time of loading $PROFILE, C:\Windows\System32 is still effect, because any command passed to -command isn't processed until after the profiles have been loaded.
If you do need to know in $PROFILE what the working directory will be once the session is open, use the following workaround:
$workingDir = [Environment]::GetCommandLineArgs()[-1] -replace "'"
[Environment]::GetCommandLineArgs() returns the invoking command line as an array of arguments (tokens), so [-1] returns the last argument, assumed to be the working-directory path; -replace "'" removes the enclosing '...' from the result.
However, so as to make your $PROFILE file detect the (ultimately) effective working directory (location) irrespective of how PowerShell was invoked, more work is needed.
The following is a reasonably robust approach, but note that a fully robust solution would be much more complex:
# See if Set-Location was passed and extract the
# -LiteralPath or (possibly implied) -Path argument.
$workingDir = if ([Environment]::CommandLine -match '\b(set-location|cd|chdir\sl)\s+(-(literalpath|lp|path|PSPath)\s+)?(?<path>(?:\\").+?(?:\\")|"""[^"]+|''[^'']+|[^ ]+)') {
$Matches.path -replace '^(\\"|"""|'')' -replace '\\"$'
} else { # No Set-Location command passed, use the current dir.
$PWD.ProviderPath
}
The complexity of the solution comes from a number of factors:
Set-Location has multiple aliases.
The path may be passed positionally, with -Path or with -LiteralPath or its alias -PSPath.
Different quoting styles may be used (\"...\", """...""", '...'), or the path may be unquoted.
The command may still fail:
If the startup command uses prefix abbreviations of parameter names, such as -lit for -LiteralPath.
If a named parameter other than the path follows set-location (e.g., -PassThru).
If the string set-location is embedded in what PowerShell ultimately parses as a string literal rather than a command.
If the startup command is passed as a Base64-encoded string via -EncodedCommand.
[1] When you type powershell.exe into File Explorer's address bar instead, the currently open folder is made the working directory before PowerShell is started, and no startup command to change the working directory is passed; in that case, $PROFILE already sees the (ultimately) effective working directory.
1.open the registry (command:regedit)
2.find out the path \HKEY_CLASSES_ROOT\Directory\Background\shell\Powershell\command (not \HKEY_CLASSES_ROOT\Directory\Background\shell\cmd\command)
3.the default value should be powershell.exe -noexit -command Set-Location -literalPath "%V"
4.you can change some param,
ps: you change the command to cmd.exe /s /k pushd "%V". if you do so, shift & right button in the explorer will open the cmd, not powershell
I have an extensive set of CMD scripts for an automation suite.
In a console using CMD.exe, everything works fine. After installing the Windows Creator's update, where PowerShell becomes the new default Windows shell via Explorer's menu options, my scripts break at-random. I can't provide any meaningful code for repro for two main reasons:
No obvious error is occurring; my automated scripts just hang, eventually
The halt isn't even occurring in the same place each time
What I can tell you is that the suite heavily relies on exit codes, use of findstr.exe, and type.
I know things like Windows macros, e.g., %Var% are not compatible, but I was assuming that since the only call I did was to a .bat file, .bat behavior would be the only thing I would need to worry about.
If that's not the case, should my initial .bat be triggering the direct execution of a CMD.exe instance with my parameters? If so, what's the best way to do that, PowerShell-friendly?
eryksun's comments on the question are all worth heeding.
This section of the answer provides a generic answer to the generic question in the question's title. See the next section for considerations specific to the OP's scenario.
Generally speaking, there are only a few things to watch out for when invoking a batch file from PowerShell:
Always include the specific filename extension (.bat or .cmd) in the filename, e.g., script_name.bat
This ensures that no other forms of the same command (named script_name, in the example) with higher precedence are accidentally executed, which could be:
In case of a command name without a path component:
An alias, function, cmdlet, or an external executable / PowerShell script (*.ps1) that happens to be located in a directory listed earlier in the $env:PATH (%PATH%) variable; if multiple executables by the same name are in the same (earliest) directory, the next point applies too.
In case of a command name with a path component:
A PowerShell script (*.ps1) or executable with the same filename root whose extension comes before .bat or .cmd in the %PATHEXT% environment variable.
If the batch file is located in the current directory, you must prefix its filename with ./
By design, as a security measure, PowerShell - unlike cmd.exe - does NOT invoke executables located in the current directory by filename only, so invoking script_name.bat to invoke a batch file of that name in the current directory does not work.[1]
Instead, you must use a path to target such an executable so as to explicitly signal the intent to execute something located in the current directory, and the simplest approach is to use prefix ./ (.\, if running on Windows only); e.g., ./script_name.bat.
When passing parameters to the batch file:
Either: be aware of PowerShell's parsing rules, which are applied before the arguments are passed to the batch file - see this answer of mine.
Or: use --% (the PSv3+ stop-parsing symbol) to pass the remaining arguments as if they'd been passed from a batch file (no interpretation by PowerShell other than expansion of %<var>%-style environment-variable references).
[1] eryksun points out that on Vista+ you can make cmd behave like PowerShell by defining environment variable NoDefaultCurrentDirectoryInExePath (its specific value doesn't matter).
While ill-advised, you can still force both environments to always find executables in the current directory by explicitly adding . to the %PATH% / $env:PATH variable; if you prepend ., you get the default cmd behavior.
As for your specific scenario:
After installing the Windows Creator's update, where PowerShell becomes the new default Windows shell via Explorer's menu options
This applies to the following scenarios:
Pressing Win-X (system-wide keyboard shortcut) now offers PowerShell rather than cmd in the shortcut menu that pops up.
Using File Explore's File menu now shows Open Windows PowerShell in place of Open command prompt (cmd).
However, nothing has changed with respect to how batch files are invoked when they are opened / double-clicked from File Explorer: The subkeys of HKEY_CLASSES_ROOT\batchfile and HKEY_CLASSES_ROOT\cmdfile in the registry still define the shell\open verb as "%1" %*, which should invoke a batch file implicitly with cmd /c, as before.
However, per your comments, your batch file is never run directly from File Explorer, because it require parameter values, and it is invoked in two fundamental ways:
Explicitly via a cmd console, after entering cmd in the Run dialog that is presented after pressing Win-R (system-wide keyboard shortcut).
In this case, everything should work as before: you're invoking your batch file from cmd.
Explicitly via PowerShell, using File Explorer's File menu.
Per your comments, the PowerShell console may either be opened:
directly in the directory in which the target batch file resides.
in an ancestral directory, such as the root of a thumb drive on which the batch file resides.
In both cases, PowerShell's potential interpretation of arguments does come into play.
Additionally, in the 2nd case (ancestral directory), the invocation will only work the same if the batch file either does not depend on the current directory or explicitly sets the current directory (such as setting it to its own location with cd /d "%~dp0").
This is a non-answer solution if encountering the question's specific behavior. I've verified all my halting scripts stopped halting after implementing a shim-like workaround.
As erykson said, there doesn't appear to be a reason why using a shim would be required. The goal is then to explicitly launch the script in CMD when using PowerShell, which aligns with Jeff Zeitlin's original suggestion in the question's comments.
So, let's say you're in my shoes with your own script_name.bat.
script_name.bat was your old script that initializes and kicks off everything. We can make sure that whatever was in script_name.bat is correctly run via CMD instead of PowerShell by doing the following:
Rename script_name.bat to script_name_shim.bat
Create a new script_name.bat in the same directory
Set its contents to:
#echo off
CMD.exe /C "%~dp0script_name_shim.bat" %*
exit /b %errorlevel%
That will launch your script with CMD.exe regardless of the fact that you started in PowerShell, and it will also use all your command-line arguments too.
This looks like a chicken egg problem, wihtout knowing the code it's difficult to tell where the problem is.
There are a ton of ways to start batches with cmd.exe even in win10cu.
Aliases are only a problem when working interactively with the PowerShell console and expecting behavior as it used to be in cmd.exe.
The aliases depend also on the loaded/imported modules and profiles.
This small PowerShell script will get all items from Help.exe and
perform a Get-Command with the item.
internal commands without counterparts in PoSh are filtered out by the ErrorAction SilentlyContinue.
Applications (*.exe files) are assumed identical and removed by the where clause.
help.exe |
Select-String '^[A-Z][^ ]+'|
ForEach-Object {
Get-Command $_.Matches.Value -ErrorAction SilentlyContinue
}| Where-Object CommandType -ne 'Application'|Select *|
Format-Table -auto CommandType,Name,DisplayName,ResolvedCommand,Module
Sample output on my system, all these items will likely work differently in PowerShell:
CommandType Name DisplayName ResolvedCommand Module
----------- ---- ----------- --------------- ------
Alias call call -> Invoke-Method Invoke-Method pscx
Alias cd cd -> Set-LocationEx Set-LocationEx Pscx.CD
Alias chdir chdir -> Set-Location Set-Location
Alias cls cls -> Clear-Host Clear-Host
Alias copy copy -> Copy-Item Copy-Item
Alias del del -> Remove-Item Remove-Item
Alias dir dir -> Get-ChildItem Get-ChildItem
Alias echo echo -> Write-Output Write-Output
Alias erase erase -> Remove-Item Remove-Item
Alias fc fc -> Format-Custom Format-Custom
Function help pscx
Alias md md -> mkdir mkdir
Function mkdir
Function more
Alias move move -> Move-Item Move-Item
Function Pause
Alias popd popd -> Pop-Location Pop-Location
Function prompt
Alias pushd pushd -> Push-Location Push-Location
Alias rd rd -> Remove-Item Remove-Item
Alias ren ren -> Rename-Item Rename-Item
Alias rmdir rmdir -> Remove-Item Remove-Item
Alias set set -> Set-Variable Set-Variable
Alias sc sc -> Set-Content Set-Content
Alias sort sort -> Sort-Object Sort-Object
Alias start start -> Start-Process Start-Process
Alias type type -> Get-Content Get-Content
I'm pretty new to scripting (especially powershell) and new to Stack Overflow, so please, excuse my ignorance and please bear with me! I will do my best to specifically explain what I'm looking to do and hopefully someone can give a detailed response of what I could do to make it work..
Intended Process/Work Flow: A co-worker downloads "Install.zip" file that has all the necessary files. This "Install.zip" file contains "Setup.bat" file (for computer config), "Fubar.zip" file, 2 powershell scripts, and a custom powerplan (.pow) file. Once downloaded they will run the "Setup.bat" file and it will pretty much do all the work. Inside that batch file it calls 2 powershell scripts. 1)"Download.ps1" - Downloads some other files from the web. 2.) "Unzip.ps1" - Unzips "Fubar.zip" and places contents in another folder - C:\TEST\
Issue: I've recently gotten familiar with using %~dp0 in batch files. I want to make sure that the location where my co-worker initially downloads the Install.zip doesn't throw off my batch file. So for example.. some people will download .zip files to the "Downloads" folder, then extract contents to proper destination. Others will download the .zip to a specific folder, then extract it within that folder. [Ex: C:\Alex\Install.zip --Extraction-- C:\Alex\Install\((Content))] So I tried to not pre-define file locations due to the variables. I've gotten the %~dp0 to work everywhere I need it to in my batch file. The only issue I have is getting my powershell scripts to use same working directory that my batch file is in. *My batch file and my powershell scripts will always be in the same directory. (Wherever that may be)
Goal: I want my powershell script ("Unzip.ps1") to look for my "Fubar.zip" file in the same directory that its currently running in. (Again - Wherever that may be) I basically want to remove any variables that may throw off the powershell script. I want it to always use it's current working directory to locate Fubar.zip. I basically need powershell to either use its current working directory OR figure out a way to have it pull its current working directory and use that to look for "Fubar.zip".
my current "Unzip.ps1" powershell script is extremely basic.
Unzip.ps1:Expand-Archive -Force c:\ALEX\Install.zip\Fubar.zip -dest c:\TEST\
Batch File Command that calls the Unzip.ps1 script: Powershell.exe -executionpolicy remotesigned -File %~dp0UNZIP.ps1
Please keep in mind, I'm just learning scripting and I'm teaching myself. My knowledge is limited, but I've made it this far and this is the only part I'm stuck on. Please give clear responses. Any help or advice would be extremely appreciated! Using PowerShell 5.0
Thanks in advance!
The equivalent of cmd.exe's %dp0 in PowerShell v3+ is $PSScriptRoot:
Both constructs expand to the absolute path of the folder containing the batch file / PowerShell script at hand.
(The only difference is that %dp0 contains a trailing \, unlike $PSScriptRoot).
Thus, to make your Unzip.ps1 script reference Fubar.zip in the same folder that the script itself is located in, use:
Expand-Archive -Force $PSScriptRoot\Fubar.zip -dest c:\TEST\
Constructing the path as $PSScriptRoot\Fubar.zip - i.e., blindly joining the variable value and the filename with \ - is a pragmatic shortcut for what is more properly expressed as (Join-Path $PSScriptRoot Fubar.zip). It is safe to use in this case, but see the comments for a discussion about when Join-Path should be used.
The automatic $PSScriptRoot variable requires PS v3+; in v2-, use
Expand-Archive -Force ((Split-Path $MyInvocation.Mycommand.Path) + '\Fubar.zip') -dest c:\TEST\
From your description, I gather that Fubar.zip and Unzip.ps1 are in the same directory. We'll pretend this directory is C:\Users\Me\Temp; although I understand that may vary.
Powershell's working directory will be the directory you're in (if called from CMD) when you launch; otherwise, it'll be from $env:UserProfile. Since the .bat file always call Unzip.ps1 from the directory that it's in (C:\Users\Me\Temp), powershell.exe will find it with this command (you can still use %~dp0 here; it's not hurting anything):
powershell.exe -ExecutionPolicy RemoteSigned -File Unzip.ps1
Inside of Unzip.ps1, you'll use Get-Location:
Expand-Archive -Force "$(Get-Location)\Fubar.zip" -dest c:\TEST\
However, if the .bat file does a cd into another directory, this won't work. From your %~dp0UNZIP.ps1 example, I assume this isn't the case, but let's address it anyway. If this is the case, you need to process from where the location of the script is. So for this call the full/relational path to the .ps1:
powershell.exe -ExecutionPolicy RemoteSigned -File C:\Users\Me\Temp\Unzip.ps1
Then, your Unzip.ps1 will need to look like this:
Expand-Archive -Force "${PSScriptRoot}\Fubar.zip" -dest 'C:\TEST\'
Alternatively, you can also do some fancy path splitting, as #JosefZ suggested. The $PSCommandPath and $MyInvocation variables contain the full path to your script; which you should familiarize yourself with:
$location = Split-Path $PSCommandPath -Parent
$location = Split-Path $MyInvocation.MyCommand.Path -Parent
Expand-Archive -Force "${location}\Fubar.zip" -dest 'C:\TEST\'
Note: Of course, you wouldn't set $location twice. I'm just showing you two ways to set it.
I hope this helps!
Prior to Powershell 3
$currentScriptPath = Split-Path ((Get-Variable MyInvocation -Scope 0).Value).MyCommand.Path
Otherwise, $PsScriptRoot will work. If you're going to depend on it, though, I'd make sure you mark the script with #Requires -version 3
I'd advise against changing the PWD unless you must. Which is to say, reference the variable directly.