I want \"git log --format=\'(%h) %s\' --abbrev=7 HEAD\"
to be split into
[
\"git\",
\"log\",
\"--format=\'(%h) %s\'\",
\"--abbrev=7\",
\
As I understand, the idea is to split the string on contiguous spaces except where the spaces are part of a substring surrounded by single quotes. I believe this will work:
/(?:[^ ']*(?:'[^']+')?[^ ']*)*/
but invite readers to subject it to careful scrutiny.
demo
This regex can be made self-documenting by writing it in free-spacing mode:
/
(?: # begin a non-capture group
[^ ']* # match 0+ chars other than spaces and single quotes
(?: # begin non-capture group
'[^']+' # match 1+ chars other than single quotes, surrounded
# by single quotes
)? # end non-capture group and make it optional
[^ ']* # match 0+ chars other than spaces and single quotes
)* # end non-capture group and execute it 0+ times
/x # free-spacing regex definition mode
This obviously will not work if there are nested single quotes.
@n.'pronouns'm. suggested an alternative regex that also works:
/([^ "']|'[^'"]*')*/
demo
As often in life, you have choices.
Use an expression that matches and captures different parts. This can be combined with a replacement function as in
import re
string = "git log --format='(%h) %s' --abbrev=7 HEAD"
rx = re.compile(r"'[^']*'|(\s+)")
def replacer(match):
if match.group(1):
return "#@#"
else:
return match.group(0)
string = rx.sub(replacer, string)
parts = re.split('#@#', string)
# ^^^ same as in the function replacer
You could use the better regex module with (*SKIP)(*FAIL)
:
import regex as re
string = "git log --format='(%h) %s' --abbrev=7 HEAD"
rx = re.compile(r"'[^']*'(*SKIP)(*FAIL)|\s+")
parts = rx.split(string)
Write yourself a little parser:
def little_parser(string):
quote = False
stack = ''
for char in string:
if char == "'":
stack += char
quote = not quote
elif (char == ' ' and not quote):
yield stack
stack = ''
else:
stack += char
if stack:
yield stack
for part in little_parser(your_string):
print(part)
['git', 'log', "--format='(%h) %s'", '--abbrev=7', 'HEAD']
I found one possible (albeit ugly) solution in python (which also works with "
):
>>> import re
>>> foo = '''git log --format='(%h) %s' --foo="a b" --bar='c d' HEAD'''
>>> re.findall(r'''(\S*'[^']+'\S*|\S*"[^"]+"\S*|\S+)''', foo)
['git', 'log', "--format='(%h) %s'", '--foo="a b"', "--bar='c d'", 'HEAD']