[issue35377] urlsplit scheme argument broken

[devkral]

New submission from devkral :

the scheme argument of urlsplit/urlparse is completely broken.
here two examples:

urlunsplit(urlsplit("httpbin.org", scheme="https://";))
'https://:httpbin.org'

urlunsplit(urlsplit("httpbin.org", scheme="https"))
'https:///httpbin.org'

Fix: change urlsplit logic like this:
...
url, scheme, _coerce_result = _coerce_args(url, scheme)
scheme = scheme.rstrip("://") # this removes ://
...
i = url.find('://') # harden against arbitrary :
if i > 0:
...
elif scheme:
netloc, url = _splitnetloc(url, 0)  # if scheme is specified, netloc is 
implied

sry too lazy to create a patch from this. Most probably are all python versions 
affected but I checked only 2.7 and 3.7 .

--
components: Library (Lib)
messages: 330884
nosy: devkral
priority: normal
severity: normal
status: open
title: urlsplit scheme argument broken
versions: Python 2.7, Python 3.7

___
Python tracker 
<https://bugs.python.org/issue35377>
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue35377] urlsplit scheme argument broken

[devkral]

devkral  added the comment:

first a correction; there is one protocol where it make sense:
file
I would change:
...
elif scheme:
...
to
...
elif scheme and scheme != "file":
...


Second: no it is not a correct behaviour. Urlunsplit is supposed to create a 
valid url from a tuple created by urlsplit. :/// is not a valid url (except for 
the file protocol).

Third: rstrip could be simplified to rstrip(":/")

--

___
Python tracker 
<https://bugs.python.org/issue35377>
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue35377] urlsplit scheme argument broken

[devkral]

devkral  added the comment:

Ok, then the problem is in unsplit.
Because: :/// is not a valid url.

Next try: in urlunsplit:


if netloc or (scheme and scheme in uses_netloc and url[:2] != '//'):
if url and url[:1] != '/' and scheme in ("file", ""): # my change
url = '/' + url

--

___
Python tracker 
<https://bugs.python.org/issue35377>
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com

[issue35377] urlsplit scheme argument broken

[devkral]

devkral  added the comment:

ah, I get the idea behind urlunsplit. But still:
for e.g. http, http:/// is invalid. These urls are invalid for e.g. requests.
May I ask:
urllib.parse is for web protocol urls? If yes, then there should be an option 
which defaults to my version.

--

___
Python tracker 
<https://bugs.python.org/issue35377>
___
___
Python-bugs-list mailing list
Unsubscribe: 
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com