homepage

The webbrowser module seems to treat URLs containing the "|" character
differently based on whether the browser is already running or not.
For instance, consider the following python script:
import webbrowser
url = "http://foo.com/bar.html?var=x|y|z"
webbrowser.open(url)
If you run this script while the browser is already running (so that
webbrowser.open creates a new tab) this behaves as you would expect,
with the given URL as an address.
However, if a browser is not already running, when webbrowser.open
creates it, it seems to interpret the "|" as a seperator character, so
that the browser will open with THREE tabs, one open to
"http://foo.com/bar.html?var=x", one to "http://y" and one to "http://z".
This is clearly a bug, webbrowser module should be smart enough to
escape the "|" character if the browser is interpreting that line
differently.
This happens in Linux with Python 2.5 and Firefox 3.0. Not sure if it
happens with anything else.

Just as an aside, the reason I consider this a fairly serious bug is
that the Google Charts API urls make heavy use of the '|' character,
which means if I want to have Python use it by opening the user's
browser, it won't work if they don't already have a browser open.

I was able to duplicate this, but it's an issue with Firefox, not
Python. webbrowser.open(url) just passes url as a command line argument
to the web browser; it doesn't do any manipulation.
Note that you get the exact same behaviour if you run Firefox from the
command line:
> firefox 'http://foo.com/bar.html?var=x|y|z'
Opens this URL in a new tab if it's already open, but splits on '|' and
opens in 3 separate tabs if Firefox isn't running.
Note also that while this string is a URL, it isn't properly normalized.
This works fine if you call
webbrowser.open("http://foo.com/bar.html?var=x%7Cy%7Cz")
(Which you can only obtain programmatically by generating the URL
properly in the first place, by using urllib.urlencode, or urllib.quote
on the value string "x|y|z").

Closing as invalid.