Unable to acquire request token from Twitter
I wonder if this the problem?
Although the dates don't seem to match exactly.
If it was working and nothing has changed then it must be twitter (AGAIN).
God damn it. More busy work just to keep up with vendors who constantly screw with their API's.
Why do they do this? *sigh* Let me have a look (again). I did make changes the last time they mess around with all this, now they are doing it again.
UPDATE: I ran some tests and I found that http:// connections to twitter always return a 403. I've update xap-twitter to use https:// and now its working again. I guess the grace period for doing this is over and twitter are enforcing it.
Anyway all fixed in 311.6 beta - if you've not migrated do read http://www.homeautomationhub.com/content/and-coming-changes which has some background about some of the changes.
Brett
Kema,
I haven't tried the update yet (will try have a go later) but looking through the changes Brett has made today, I believe this should fix functionality for those already authorised with Twitter, not sure the authorise button via Kloned is fixed yet though. This might explain what your seeing.
ill see if I can test the update but I'm running on a Pi at the moment and cannot just run the update-dev command like I used too!.
Garry
Edit: can confirm twitter now working on an already authorised system. Cannot test authorisation as I don't have Kloned, but like I say I suspect this still broken.
Thanks Gary... I never checked that the Authorization service is still working
UPDATE: 311.7 - the twitter auth from the web page is good now.
It getting time to push this as 312 given stuff like this has to be fixed.
Brett
Kema thanks for the confirmation. googlecal is a bit of a sad story the library that I've been using has not had any updates since 2011 http://code.google.com/p/libgcal/source/list so I'm thinking either a) I have to debug this library myself or b) switch to a supported gcal framework which means another language and as space is VERY tight on the livebox that might be problematic c) give it up as dead :( I'm still pondering the effort reluctantly.
Cron should however still be fully functionaly.
1. Make sure cron is enabled http://livebox/services.kl1
# cat /etc/xap.d/system.ini
<other bits removed>
[cron]
enable=1
2. Get a command line prompt and check for a crond process with 'ps'
# ps | grep crond
203 root 1808 S crond
219 root 1800 R grep crond
3. Check there is a linkage in /var/spool/cron.
# ls -l /var/spool/cron
lrwxrwxrwx 1 root root 13 Jan 1 1970 crontabs -> /etc/crontabs
Make sure this points to a /etc/crontabs directory if you have someting setup you'll see a file called 'root'
# ls -l /etc/crontabs/
-rw------- 1 root root 25 Jan 30 09:25 root
4. Paste in the output /etc/crontabs/root so we can see what you have.
Brett
Well, I've talked to Brett about this a fair bit over the last month or two. The Livebox as a platform is kinda getting a bit hard to come by and its lack of space is beginning to limit the extra things that might be done.
However, I'm keen not to lose some of the core 'addon offerings' that come with the HAH PCB. Platform wise, I reckon that the Rasp Pi should be a focus. Simply because there are a lot of them out there. BeagleBone too as it's just another xcompile.
I'm playing with an idea to make an ethernet connected version of the HAH PCB. Same relays/inputs/RF Tx and 1-Wire ... but in a standalone, xAP compliant unit.
Hard to really know if there would be demand for such a thing ...
Derek.
Derek,
I agree that moving to other platforms will help with the longevity of the project and open up the project to a wider audience.
I can't talk for the beaglebone but certainly moving HAH onto the pi would not render the HAH pcb useless.
With absolutely no changes to the code and just the price of a gpio breakout board (<£3) I have the pcb working fine. Work in progress pic attached .
An Ethernet version would open up other possibilities but like you say, difficult to guage what the demand would be.
Garry
Attachment | Size |
---|---|
image.jpg | 43.89 KB |
Kema,
Is it definitely Cron that is the issue? Do your alias commands work straight from the command line?
Garry
The stringx is my bad - this happend when I moved to using the 'hostname' instead of defaulting to livebox as I added a new function into xap/init.lua which used stringx and I didn't explicity require the module.
Although this fixed aliasmsg I also found that I could not run bscmsg due to a bug in the penlight library so I uplifted that as well.
released: 311.8
Brett
If you are not using the IRnode then just comment out this line in your jeenodeApplet.lua
--IRNode = require("xap.irnode")
I will fix this in the next beta drop. If you are using the IR node let me know and I'll push up the beta sooner.
Its a bug in the decoder it doesnt return the class like the other now do.
If you are using it then you can workaround the bug (until the next beta drop) like this:
IRNode = require("xap.irnode").IRNode
Brett
Kema,
The error you are getting out of twitter indicates that an internal array ran out of room after parsing all the tokens it could hold. This was set to 60 tokens, I've increased this to 100. The reason it worked even thou you get this error is because the token we are after was in this batch of 60. Still we don't want the error as it will fill up the logs and eventually crash the livebox out. So even thou its benign it can have a nasty side effect if not dealt with.
Brett
I know twitter was working for me a couple of months ago becaue I used it to adjust my heating while I was away. But I have tried to use it a couple of times recently and it is broken for me as well. Also using 311/3.4