外壳文本到python字符串

Question

I'm writing a little python utility to help move our shell -help documentation to searchable webpages, but I hit a weird block :

output = subprocess.Popen([sys.argv[1], '--help'],stdout=subprocess.PIPE).communicate()[0]
output = output.split('\n')
print output[4]
#NAME
for l in output[4]:
    print l
#N
#A
#
#A
#M
#
#M
#E
#
#E

#or when written, n?na?am?me?e

It does this for any heading/subheading in the documentation, which makes it near unusable.

Any tips on getting correct formatting? Where did I screw up?

Thanks

Answer 1

The documentation contains overstruck characters done in the ancient line-printer way: print each character, followed by a backspace ( \\b aka \\x08 ), followed by the same character again. So "NAME" becomes "N\\bNA\\bAM\\bME\\bE" . If you can convince the program not to output that way, it would be the best; otherwise, you can clean it up with something like output = re.sub(r'\\x08.', '', output)

Answer 2

A common way to mark a character as bold in a terminal is to print the character, followed by a backspace characters, followed by the character itself again (just like you would do it on a mechanical typewriter). Terminal emulators like xterm detect such sequences and turn them into bold characters. Programs shouldn't be printing such sequences if stdout is not a terminal, but if your tool does, you will have to clean up the mess yourself.