Hello @keve,
Did you try the custom musl test release ?
I may have to make a new release sooner or later and would like to know if I will include the musl lib, as in my test, or if I have changes to make.
Hello @keve,
Did you try the custom musl test release ?
I may have to make a new release sooner or later and would like to know if I will include the musl lib, as in my test, or if I have changes to make.
Hi @dalgwen,
first: thank you for all the effort and energy you put into this binding. It is great to see people are still putting their capabilities into free and open projects. Thank you!
I probably donāt have the same problem than keve, but I encounter the challenge to register my landline via voice. First error codes refer to the wrong signal-cli version. That is why I would like to use your linked snapshot .kar file.
But I canāt figure out how I make my openHAB instance use it.
I may oversee some basic documentation, but putting it into the addons folder and modify the service.cfg does not seem to work.
Could one guide me to a documentation or even tell me how to deploy the kar file?
Thank you in advance!
Philipp
EDIT: I guess my problem has to do with my docker deployment of openHAB. I currently have no access to the environment and will check that later today.
Well, I donāt know why, but my land line is registered now. There was no problem with the add-on itself or the docker setup. I reset everything (deleted userdata/signal
folder) and followed the steps of the documentation.Thanks again for this binding!
Strange. Putting it into the addons folder should have been sufficient.
I didnāt answer your message because every passing day I thought I would be able to update the version soon. But as I didnāt manage to, Iām glad it now works for you. Less work for me
If someone has the same issue, please tell me.
There have been many fixes and updates since I last tried, which is great, but now I cannot seem to get beyond:
OFFLINE
COMMUNICATION_ERROR
RegistrationRetryException - StatusCode: 429
I cannot find much information on trying to unregister an old number to retry with, and hoped the API would automatically overwrite. Is something else going on?
The phone number I am using is a dedicated number that should not have another signal configuration associated with it. This is OpenHAB 4.1.3 with binding version marked with date May 28, 2024 7:57 AM
One other note, I found the github ticket that was opened for this error. One of the suggestions was to run the registration process from the server running the binding and OpenHAB. This did not seem to change anything.
Any help appreciated.
Hello,
There is no need to āunregisterā your old old registration, it should be overwritten automatically. At least itās what happened in all my tests.
The 429 is a http code meaning ātoo many requestā. It seems that you have tried too many times in a short period. You have to wait some time (donāt know how long, maybe hours ? but I guess a day is a safe duration). you should absolutely stop the signal thing during this wait period to prevent it from trying again.
Besides that, I also found, like you, the github ticket on signal-cli mentionning this issue as a recurring one for some users. It seems kind of random and no one can pinpoint precisely whatās happening ?
Are you trying a registration with a voice call or by SMS ? Voice call is trickier, as it involves the binding waiting for one minute after trying a SMS (procedure required by the signal server). (It should be automatic, you donāt have to do anything)
Donāt know if itās a real requirement or just a wild guess, but even if so, the same IP should be sufficient.
Is ipv6 used for this ? Do you have an ipv6 range at your home ? If so, then each one of your computer may have its own IP. Itās doesnāt seem likely but it could be worth investigating by disabling ipv6 on your network to force ipv4 ?
I also have the same advice as always : removing the binding, deleting everything in the openhab data /signal directory, reinstalling it, before trying again. Itās tedious but the registration process can be picky.
Hi, thank you for the quick reply. To answer your questions, I have tried with both SMS and Voice. It doesnāt get that far though, as the 429 comes before the verification code attempt.
I am using IPV4, so the uniqueness of IPV6 isnāt an issue.
It sounds like waiting some time might be the trick, so I will wait a few hours and try again. Then try again in a day and report back.
HTTP 429 Too Many Requests is triggered too, if captcha
signalcaptcha://signal-hcaptcha.5fad97ac-7d06-4e44-b18a-b950b20148ff.registration.[Very long string]
ist too old or was generated from another Network.
It has been a week, and I am just now trying it again. I did the captcha, and forgot to save, giving me the 429. After hitting save, to update the captcha, it sent the code, and I was able to complete the process. This makes me wonder if I failed to hit save last time after pasting the captcha in as well. At any rate, if it helps anyone else, donāt forget to hit save first.
Thanks for keeping me informed, Iām glad it works and that I donāt have to find a bug
I will emphasize in the documentation about the need to click on āsaveā at each step.
Hi all,
Iām currently trying to get the signal binding up and running. While creating the signalaccount thing Iām getting the following error:
[core.thing.internal.ThingManagerImpl] - Exception occurred while initializing handler of thing 'signal:signalaccountbridge:41239fe366': no signal_jni in java.library.path: /var/lib/openhab/tmp/lib
java.lang.UnsatisfiedLinkError: no signal_jni in java.library.path: /var/lib/openhab/tmp/lib
Iām running on the following environment:
runtimeInfo:
version: 4.1.3
system info:
javaVersion: 17.0.11
javaVendor: Debian
osName: Linux
osVersion: 5.10.0-30-686-pae
osArchitecture: i386
Could the i386/32bit be the reason for my problems?
Thanks for any hint.
Best
Andee
Yes it is.
Sorry, I didnāt think there was still some users with a i386/32 bits OS nowadays , and so didnāt add the relevant library (the goal was to prevent the binding from being to big, which is by the way a total failure). I never have such a request in nearly two years, so I suppose there are not many users in your case
I will add it in the next release.
If this would be possible without too many additional work from your side, I would be happy.
It runs on quite old but robust hardware on debian 11. I guess I should switch to 64bit, but itās up and running well.
I made a new release for i686 32 bits supports. I hope this works
Remove the old binding from the marketplace and reinstall.
Now it works perfectly fine! Thanks a lot.
Hi all,
Iāve been trying to make this binding work for months now, but I cannot figure out what goes wrong.
It used to work fine until around February when an upstream update of signal-cli broke it. Then, as advised here, when the update of the binding was released I removed the binding, restarted openHAB and reinstalled it, to no avail. It was some months ago so I cannot remember what was the error message. I tried deleting everything related to the binding from the system, then formatting the whole system and reinstalling openhab from a backup, then formatting everything again and adding every config file one by one to make sure there was not an old config file with some bad settings accidentally stored somewhere. Nothing worked, so at this point I gave it up.
After some months, I found that I really wanted to use that binding and gave it another try. Now, everytime I try the same thing happens:
InvalidTransportModeException - StatusCode: 400
.RegistrationRetryException - StatusCode: 429
.Incomplete registration: CAPTCHA_NEEDED
. Providing a new captcha link brings back to the second step. It was stated here that the captcha link must be generated on the same network and that is what I do (even though from a laptop and not directly from the Raspberry Pi that runs openhabian).At this point I wait 7 to 10 days and give it another try.
My understanding is that the error 400 is to be expected and related to the fact that I try to register through a landline. What I donāt get is the error 429: according to various sources, including this thread, it says that I tried to many attempts with the same line and that waiting about a week should fix the issue. I did that half a dozen time now (even waiting a few months on the first attempt) and the issue remains. If I try sending a message to this landline with Signal, the message is not delivered, so it looks like the line is not registered anymore.
Do you have any idea what the issue might be? Does it come from me, from signal/signal-cli or from the binding? Can you help in any way? Iām growing desperate.
Thanks for any help you might give.
Cheers,
Andstati
Wow, Iām sorry that you add to go through so many hinderances !
In this case, where you tried thoroughly nearly every solutions and workarounds, I suggest that you go āupstreamā.
Iām sorry, because it is totally the opposite of this bindingās purpose, but I think you could try to register and use signal-cli directly (use the last 0.13.4 stable version). Whatever the result, we would then know where the issue is coming from (from signal-cli or from the binding integration)
So,
Thanks for your quick answer! Registering through signal-cli worked. For future reference, I first tried to register with SMS verification (that triggered the error 400) then waited a minute and registered with a phone call (I remember reading somewhere some time ago that this is how it is supposed to be done). It then asked me for a second captcha with a challenge, but after that I was able to send and receive messages using the command line.
After that I disabled the signalaccountbridge thing in openhab, backed-up then deleted <OPENHAB_DATA>/signal/data, copied the data folder from signal-cli instead and made sure the files had the correct ownership (I left the permissions as is though). When re-enabling the openhab thing, it is stuck on āinitializingā. Do you have any idea why? I tried two things that didnāt work: adding the signal-captcha and verification code to the configuration of the thing in the openhab interface, and renaming the files in the data folder so that they match the name of the files I removed, making sure to also rename the corresponding line in accounts.json.
EDIT 1: I had a look at the logs and found the culprit: [ERROR] [gnal.manager.storage.AccountDatabase] - Database has been updated by a newer signal-cli version
. That being said, I have no idea how to proceed from there.
EDIT 2: Simply upgrading OpenHAB did the trick. Thanks for your help, Iām so happy to at last make your binding work!
then waited a minute and registered with a phone call (I remember reading somewhere some time ago that this is how it is supposed to be done). It then asked me for a second captcha with a challenge, but after that I was able to send and receive messages using the command line
Humā¦ Normally the binding does exactly that : it asks for a sms, then wait one minute and asks for the voice call. Does it now fail because it needs a new captcha during the second phase? In my previous tests, I could use the same captcha during the whole procedure.
I donāt want to mess with my setup so I will maybe wait for another user to manifest, to see if it is a widespread issue.
EDIT 1: I had a look at the logs and found the culprit:
[ERROR] [gnal.manager.storage.AccountDatabase] - Database has been updated by a newer signal-cli version
. That being said, I have no idea how to proceed from there.
It seems there is a version mismatch. Did you use the 0.13.4 stable version of signal-cli ? And did you install the binding version ā4.2.0 - BETA 10ā (released around one week ago).
EDIT 2: Simply upgrading OpenHAB did the trick. Thanks for your help, Iām so happy to at last make your binding work!
Oh, so it really was a version mismatch. Great news, happy to see that you make it works after all this time.
Hey,
Just to be clear, I used the same captcha for both the SMS and phone call verification. It was after that, when I tried to send a text through command line, that it gave me a challenge and ask for another captcha. I havenāt logged the exact message but it also said something about waiting 24 hours to be able to send a text if I didnāt want to provide another captcha. So whatever provoked the error 429 probably was not that.