Catch CSV read error gracefully for large field limit

[elyezer]

Specify type:

• Enhancement

---

Description:

rho fact encrypt will re-encrypt the facts every time its run, would be better to avoid that and print a message stating that those facts are already encrypted?

Eventually it will raise an exception if re-encrypted multiple time:

$ rho fact encrypt --reportfile report.csv  # run this 5 times
Please enter your encryption password: 
Fact connection.host encrypted
Fact connection.port encrypted
Fact uname.hostname encrypted
Fact uname.all encrypted

$ rho fact encrypt --reportfile report.csv
Please enter your encryption password: 
Traceback (most recent call last):
  File "/home/elyezer/.virtualenvs/camayoc/bin/rho", line 6, in <module>
    exec(compile(open(__file__).read(), __file__, 'exec'))
  File "/home/elyezer/code/rho/bin/rho", line 28, in <module>
    CLI().main()
  File "/home/elyezer/code/rho/rho/cli.py", line 139, in main
    cmd.main()
  File "/home/elyezer/code/rho/rho/clicommand.py", line 93, in main
    self._do_command()
  File "/home/elyezer/code/rho/rho/factencryptcommand.py", line 114, in _do_command
    keys, data = self.read_and_encrypt(normalized_path, vault)
  File "/home/elyezer/code/rho/rho/factencryptcommand.py", line 92, in read_and_encrypt
    for row in reader:
  File "/usr/lib64/python3.6/csv.py", line 112, in __next__
    row = next(self.reader)
_csv.Error: field larger than field limit (131072)

---

Version of rho:

6409363

[chambridge]

@elyezer I don't think we need to prevent the user from doing this over and over. Its strange use case. Also less likely now that we have an outputfile.

I do think that this could be converted to a low severity bug. We should catch the _csv.Error and fail gracefully instead with something like CSV file input could not be read.