I forgot to mention, this uses the python-libpcap and libpcap libraries as 
dependencies. Both were available via apt-get on ubuntu.

On Saturday, October 22, 2016 at 11:21:22 PM UTC-5, Jerome Helbert wrote:
>
> Finally got some time it port my libpcap based stuff into the driver. I 
> figure I'd post it here in case you had any interest in incorporating it 
> into the main driver. I could see it having its advantages over the socket 
> based approach.
>
> I'm able to run this on a server already running a webserver or port 80, 
> the myacurite functionality still works, and (my favorite part) all 
> functionality is encompassed within the driver and nothing is required 
> outside of it (ie no ngrep, or tcpdump redirects.
>
> My formal OOP skills are a bit rusty, and I am sure the pcap stuff could 
> be incorporated into the architecture of the base driver better than I 
> have. My intent was to allow for both methods of operation, defaulting to 
> the socket approach unless a flag was specified. FYI, I didn't resolve the 
> standalone portion with its current form (initially I passed in the whole 
> options dict for the various parameters, now it reads those out of 
> stn_dict), just an FYI if you try to run it as is - it will probably fail 
> without some code massaging,
>
> On Sunday, October 16, 2016 at 12:40:14 AM UTC-5, Jerome Helbert wrote:
>>
>> ok, I see what radar is doing... Most of the "magic" is going to happen 
>> in a perl script that he is piping the stream into. His tcpdump and stdbuf 
>> all pretty much give the same stuff as the ngrep/sed stuff I have been 
>> trying to use (spread across multiple lines still) and then he has a perl 
>> script that glues them all together.
>>
>> I already wasn't super thrilled about having to run the helper script 
>> outside the driver to sniff the data, and I really want to avoid a third 
>> step...
>>
>> I am brushing up on my iptables... hoping that some combination of a TEE 
>> and a DNAT or REDIRECT can sniff off the relevant stream and change the 
>> dest ip/port. That then feeds into the interceptor driver as normal. If I 
>> cant get that to work, I am thinking about bringing in the libpcap stuff I 
>> had added to my hackulink driver and eliminating the helper script all 
>> together.
>>
>> If I got the libpcap working, is that something we'd be interesting in 
>> bringing into the driver itself?
>>
>> On Saturday, October 15, 2016 at 3:04:22 PM UTC-5, mwall wrote:
>>>
>>> On Saturday, October 15, 2016 at 12:21:13 PM UTC-4, Jerome Helbert wrote:
>>>  
>>>
>>>> Two things here...
>>>>
>>>>    1. I need some way to get ngrep or tcpdump to piece together the 
>>>>    http stream properly... I tried to use sed, but since sed operates on a 
>>>>    line by line basis - it is not very easy (if possible) for it to remove 
>>>>    newlines...
>>>>    
>>>> have you tried radar's approach with stdbuf and strings? once the 
>>> fragments are reassembled, you should be able to send the data along using 
>>> curl.
>>>  
>>>
>>>>
>>>>    1. I seem to have found a bug with the version reporting... since 
>>>>    it reported a different version than what was in it the bridge went 
>>>> into 
>>>>    reprogramming mode and would *not* leave until it updated firmware 
>>>>    (undid the DNS Hijack, etc and let it re-update to 224 even though it 
>>>> was 
>>>>    already at that version)
>>>>
>>>> there is an (untested) fix for this at commit b27acc1 (interceptor 
>>> driver 0.13a)
>>>
>>> [Interceptor]
>>>     ...
>>>     firmware_version = 224
>>>
>>> m
>>>
>>>
>>>  
>>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to