[Python-ideas] Making (?x) less scary

[MRAB]

On 2016-02-12 16:15, Ryan Gonzalez wrote:
> I was reading re's docs and came across this:
>
>
> Note that the (?x) flag changes how the expression is parsed. It should
> be used first in the expression string, or after one or more whitespace
> characters. If there are non-whitespace characters before the flag, the
> results are undefined.
>
>
> Is there a particular reason for this being undefined? It seems kind of
> un-Pythonic to me and more like C/C++. Would it be possible to instead
> throw an exception, e.g. ValueError?
>
In Python 3.5 I get this:

 >>> re.compile(" (?x)a b", re.DEBUG)
LITERAL 32
LITERAL 97
LITERAL 98
LITERAL 97
LITERAL 98
re.compile(' (?x)a b', re.VERBOSE|re.DEBUG)

So, apparently, if the 'flags' argument doesn't specify VERBOSE, it 
parses the pattern, and if it then finds that it's now marked as VERBOSE 
because of a "(?m)" in the pattern, it parses it again. Unfortunately, 
it generates the debug output while it parses, so you get output from 
both passes.