New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
These should probably not be possible #10165
Comments
For FileSystem provider. Trailing whitespaces is removed. Leading whitespaces is meaningful. |
/cc @mklement0 For information. |
/cc @SteveL-MSFT How we should resolve names (variables, aliases, functions/cmdlets) with whitespace characters (space, tab)?
Suggestions is
|
Why are we considering that we own the mandate for how people consider naming their variables. I really don't think there's a problem here. If you don't like this, then don't use it, but please don't try to tell me what I can name my variables. |
@JamesWTruher The problem is that bugs and unexpected behavior shows up from these types of edge cases. It isn't about telling you what you can/cannot use for variable/function/alias names. It is about ensuring there is some expectation of order though, so that you don't run into obscure bugs or unexpected behavior that show up when you use names that are not expected. Such "restrictions" would help the scripter, not just in their use of PowerShell, but in being a scripter/developer in general. On the flip side, if we don't do this, do we have tests in place for such things (variable/function/alias names that are whitespace only or that start and/or end with whitespace), so that the scripter can reasonably expect them to work just fine? I'd wager that we don't, and I think it would be better to guide PowerShell scripter here in the right direction than to spend time testing/fixing the engine so that it properly supports such things throughout the language. A few well-intentioned limits here would be a good thing in my book. |
Here's another example that came up in an email list today, where appropriate boundaries are not in place and what may appear as strange bugs show up as a result: # You can define a function with a parameter name that is numeric.
# This should not be allowed.
function Test-ParameterName {
[CmdletBinding()]
param([Switch]$7)
# The body doesn't matter
}
# You can define a function with a parameter alias that is numeric.
function Test-ParameterAlias {
[CmdletBinding()]
param([Alias('7')][Switch]$NumericAliasTest)
# The body doesn't matter
}
# You can't invoke either of these function with a parameter name or
# alias that is numeric.
# Strange bugs show up when you don't put appropriate boundaries
# in place around command/variable/etc. names
Test-ParameterName -7 # Returns an error
Test-ParameterAlias -7 # Returns an error In this case, we should have rules for parameter names and parameter aliases. For example, if they start with a number, they must contain at least one alphabetic character. But even then, there are other considerations that need to be taken into account, like this one: function Test-ParameterName {
[CmdletBinding()]
param([Switch]$10mb)
# The body doesn't matter
}
# You can't invoke this with the parameter either because the name is
# still numeric, even though it contains alphabetic characters.
Test-ParameterName -10mb # Returns an error I think we should support parameter names that are numeric, so that a command could have a parameter name such as |
@PowerShell/powershell-committee reviewed this, we don't see the value to spend time addressing this as this has been existing behavior forever and hasn't been an issue. Filenames allow whitespace and this follows that convention. Users can choose to use this if they choose and we shouldn't break them. |
@KirkMunro, you can use parameters that start like |
Steps to reproduce
Expected behavior
Command and variable names should either be trimmed so that they contain no extraneous whitespace, or they should raise an error if they contain extraneous whitespace so that there is no ambiguity on the part of the person creating the command or variable.
Set-PSBreakpoint should trim whitespace from string parameters.
Actual behavior
As indicated in the comments above.
How we might fix this
Define an AutoTrim attribute
This attribute would automatically trim string values before they are validated against validation attributes. That corrects the contents for commands, which is helpful; however, it means users can pass in values and get results that are different than those values (i.e. create an alias with leading/trailing whitespace and end up with alias that doesn't have the whitespace). This could hide bugs that exist in calling code.
Define a ValidateNoExtraneousWhitespace attribute
This attribute would result in an error in the parameter binder if a parameter of type
string
or a parameter that is a collection of typestring
contained strings that included leading or trailing whitespace, forcing the caller to deal with the problem. This is my preferred partial solution to this issue.Questions
Is there anything we can/should do about provider paths with leading/trailing whitespace? For PowerShell entities such as functions/aliases/variables, we can obviously generate errors if leading/trailing whitespace is used, but since files exist outside of PowerShell and can be created with spaces in Windows (not sure about macOS/Linux, haven't tested there), we probably can't provide a generic solution that prevents use of paths with extraneous whitespace.
Environment data
The text was updated successfully, but these errors were encountered: