python nl80211 code

python nl80211 code

[Kalle Valo]

Hello,

I want to run some tests through nl80211 and, as usual, I prefer to
use python for that. Are there any scripts/examples using nl80211 with
python available? Preferably with an open enough license.

-- 
Kalle Valo

Re: python nl80211 code

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 492 bytes --]

On Tue, 2009-12-01 at 10:53 +0200, Kalle Valo wrote:
> Hello,
> 
> I want to run some tests through nl80211 and, as usual, I prefer to
> use python for that. Are there any scripts/examples using nl80211 with
> python available? Preferably with an open enough license.

http://git.sipsolutions.net/pynl80211.git/

Doesn't have much of a license but was intended to be GPLv2, but there
was only one commit from somebody other than me who'd probably relicense
if needed.

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 801 bytes --]

Re: python nl80211 code

[Kalle Valo]

Johannes Berg <johannes@sipsolutions.net> writes:

> On Tue, 2009-12-01 at 10:53 +0200, Kalle Valo wrote:
>> Hello,
>> 
>> I want to run some tests through nl80211 and, as usual, I prefer to
>> use python for that. Are there any scripts/examples using nl80211 with
>> python available? Preferably with an open enough license.
>
> http://git.sipsolutions.net/pynl80211.git/

Thanks. This looks interesting but doesn't compile right now:

$ make
Checking netlink numbers:
check.r5cOZHc/check.c: In function 'main':
check.r5cOZHc/check.c:63: error: lvalue required as left operand of
assignment
check.r5cOZHc/check.c:78: error: 'NL80211_CHAN_MAX' undeclared (first
use in this function)
check.r5cOZHc/check.c:78: error: (Each undeclared identifier is
reported only once
check.r5cOZHc/check.c:78: error: for each function it appears in.)
check.r5cOZHc/check.c:80: error: lvalue required as left operand of
assignment
check.r5cOZHc/check.c:86: error: 'NL80211_RATE_MAX' undeclared (first
use in this function)
[...]

The parsing of nl80211 seems to fail. If I find some time, I try to
look at it.

> Doesn't have much of a license but was intended to be GPLv2, but there
> was only one commit from somebody other than me who'd probably relicense
> if needed.

That's good enough :)

-- 
Kalle Valo

Re: python nl80211 code

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 1017 bytes --]

On Tue, 2009-12-01 at 12:31 +0200, Kalle Valo wrote:

> Thanks. This looks interesting but doesn't compile right now:
> 
> $ make
> Checking netlink numbers:
> check.r5cOZHc/check.c: In function 'main':
> check.r5cOZHc/check.c:63: error: lvalue required as left operand of
> assignment
> check.r5cOZHc/check.c:78: error: 'NL80211_CHAN_MAX' undeclared (first
> use in this function)
> check.r5cOZHc/check.c:78: error: (Each undeclared identifier is
> reported only once
> check.r5cOZHc/check.c:78: error: for each function it appears in.)
> check.r5cOZHc/check.c:80: error: lvalue required as left operand of
> assignment
> check.r5cOZHc/check.c:86: error: 'NL80211_RATE_MAX' undeclared (first
> use in this function)
> [...]
> 
> The parsing of nl80211 seems to fail. If I find some time, I try to
> look at it.

Oh, that's not surprising, it's over two years old and after I started
writing the C tool (iw) I never maintained the comments in nl80211.h
that this tried to parse.

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 801 bytes --]

Re: python nl80211 code

[Kalle Valo]

Johannes Berg <johannes@sipsolutions.net> writes:

>> The parsing of nl80211 seems to fail. If I find some time, I try to
>> look at it.
>
> Oh, that's not surprising, it's over two years old and after I started
> writing the C tool (iw) I never maintained the comments in nl80211.h
> that this tried to parse.

Ok, good to know. What do you think if I remove the parsing of
nl80211.h and instead add the nl80211 commands manually as python
code? Sure, it's monkey code but easier to maintain IMHO.

Or have some intermediate language, maybe. Any ideas?

-- 
Kalle Valo

Re: python nl80211 code

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 789 bytes --]

On Tue, 2009-12-01 at 13:47 +0200, Kalle Valo wrote:
> Johannes Berg <johannes@sipsolutions.net> writes:
> 
> >> The parsing of nl80211 seems to fail. If I find some time, I try to
> >> look at it.
> >
> > Oh, that's not surprising, it's over two years old and after I started
> > writing the C tool (iw) I never maintained the comments in nl80211.h
> > that this tried to parse.
> 
> Ok, good to know. What do you think if I remove the parsing of
> nl80211.h and instead add the nl80211 commands manually as python
> code? Sure, it's monkey code but easier to maintain IMHO.

Yeah, I agree.

> Or have some intermediate language, maybe. Any ideas?

I guess we could use sparse to output introspection information for
nl80211.h instead, but is it worth it?

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 801 bytes --]

Re: python nl80211 code

[Kalle Valo]

Johannes Berg <johannes@sipsolutions.net> writes:

>> Ok, good to know. What do you think if I remove the parsing of
>> nl80211.h and instead add the nl80211 commands manually as python
>> code? Sure, it's monkey code but easier to maintain IMHO.
>
> Yeah, I agree.

Good. I'll do that then.

>> Or have some intermediate language, maybe. Any ideas?
>
> I guess we could use sparse to output introspection information for
> nl80211.h instead, but is it worth it?

Increases coolness factory but that's about it =)

-- 
Kalle Valo