Add PowerShell 7 support

[ronan-lg]

Is your feature request related to a problem? Please describe

As Get-ItemProperty output is different → parsing fails

• Number of lines
• Number of digits

Windows PowerShell 5.1

> $PSVersionTable

Name                           Value
----                           -----
PSVersion                      5.1.22621.2506
PSEdition                      Desktop

> Get-ItemProperty -Path C:\Windows
0|
1|
2|    Répertoire : C:\
3|
4|
5|Mode                 LastWriteTime         Length Name
6|----                 -------------         ------ ----
7|d-----        29/12/2023   6:01 PM                Windows
  < 6  >

PowerShell 7

> $PSVersionTable

Name                           Value
----                           -----
PSVersion                      7.3.10
PSEdition                      Core

> Get-ItemProperty -Path C:\Windows
0|
1|    Directory: C:\
2|
3|Mode                 LastWriteTime         Length Name
4|----                 -------------         ------ ----
5|d----          29/12/2023  6:01 PM                Windows
  < 5 >

In addition the output is colored:

b'd----          29/12/2023  6:01 PM                \x1b[44;1mWindows\x1b[0m'

Describe the solution you'd like

facts/windows_files.py → class File(FactBase): → process method

• from

  def process(self, output):
      if len(output) < 7:
          return None
  
      # Note: The first 7 lines are header lines
      return parse_win_ls_output(output[7], self.type)

• to

  def process(self, output):
      if len(output) < 5:
          return None
  
      # Note: The first lines are header lines
      return parse_win_ls_output(output[-2], self.type)

facts/util/win_files.py → WIN_LS_REGEX definition, relax the filetype and mode pattern

• from

        # filetype and mode
        r"^([darhsl\-]{6})\s+"

• to

        # filetype and mode
        r"^([darhsl\-]{5,6})\s+"

facts/util/win_files.py → parse_win_ls_output function, remove ansi sequences

• from

  if output:
      matches = re.match(WIN_LS_REGEX, output)

• to

  if output:
  
      # Patch to remove ansi sequences from PowerShell 7
      # @see https://stackoverflow.com/questions/14693701/how-can-i-remove-the-ansi-escape-sequences-from-a-string-in-python
      ansi_escape = re.compile(r"(\x9B|\x1B\[)[0-?]*[ -/]*[@-~]", flags=re.IGNORECASE)
      output = ansi_escape.sub('', output)
  
      matches = re.match(WIN_LS_REGEX, output)

[vojta001]

I think a much more future-proof, clean and safe (with respect to various kinds of injections when running commands on untrusted hosts) way is to use a proper serialization (JSON/XML) for transport. It is even much easier to handle the results on pyinfra's side while simply switching from Get-ItemProperty -Path C:\Windows to Get-ItemProperty -Path C:\Windows | ConvertTo-Json and possibly requesting only a subset of fields for efficiency.