Accidental recursion error
New here? Learn about Bountify and follow @bountify to get notified of new bounties! x

There is a what appears like an infinite recursion error in the v0.6 branch of PyBitmessage. I cannot reliably reproduce it and I don't know where in the code it's triggered. The code isn't supposed to be recursive at all (apart from minor exceptions, which test negative for this bug).

More info:

A solution would, in the minimum, pinpoint the code location where it happens and provide a method for triggering it. A pull request against the v0.6 branch that fixes it will receive a tip. When creating a pull request, please follow the contribution guidelines, of most importance is that:

  • the pull request is against the v0.6 branch
  • all the commits in the pull request are PGP signed
  • the pull request can be merged using a fast-forward merge
Hi peter, hi have downloaded the code(V0.6) and all its dependencies, and tried to run the python , but I got no error.Any idea why that might be so?Thank you
Codeword 9 days ago
You may have to wait for hours for the error to appear. That's one of the problems, I don't know how to trigger it. Try to make sure you're accepting incoming connections, that may help it trigger faster.
PeterSurda 9 days ago
okay thanks.Let me run it for few hour, to see if i can get the error.
Codeword 9 days ago
It looks like g1itch's comment is correct, if you empty the inventory and join a chan, it will trigger faster. Tested with the "general" chan. To delete the inventory, execute while PyBM isn't running: sqlite3 ~/.config/PyBitmessage/messages.dat 'DELETE FROM inventory'
PeterSurda 8 days ago
Thanks, peter
Codeword 8 days ago
Hi Peter, I did exactly as you told, but the error doesn't seem to show up.I think it's a bit shy.Hahahaha
Codeword 7 days ago
Do you have any idea, as of where the recursion error might reside, I mean which script if you have an approximation.Thank you
Codeword 7 days ago
Try to repeat the process. After it's done syncing and still no trigger, repeat it again. Regarding where it is, that's the point, if I knew that I wouldn't ask.
PeterSurda 7 days ago
hey peter, atlast I got the error, let me see if i can find out the reason.Thank you
Codeword 4 days ago
I was able to replicate it without a chan as well by the way.
PeterSurda 4 days ago
You know, I actually got a crash at round the time when you made your comment. So maybe it was triggered by an object?
PeterSurda 4 days ago
yeah you may be right.Let me see if i can find any more info.Thank you
Codeword 4 days ago
awarded to Codeword

Crowdsource coding tasks.

1 Solution

Winning solution

Hey Peter, As you might already know, a good place to start debugging would be, also I found this file regarding the _ctype.DictRemover as or your application is also using ctype library which in turn is using _ctype
See if you can get any clue from as this file has the so-called _ctype.DictRemover in it.This is as far as I could do.Thank you

If you need more time, you can continue working and I'll send you additional bounty directly.
PeterSurda 3 days ago
If you say so then I continue searching for the error and if I found it I will report it to you here in the comment section.Thank you.Bye the way are u the author of this application?.If yes I must say this application is pretty useful.
Codeword 3 days ago
I'll award you the bounty now as I think you did improve on my understanding on the problem, but you can get more if you continue working and manage to fix it. I repeated tests over and over, and I think it's happening in the decryption thread (if I comment the "processxxxx" calls in, it doesn't seem to happen anymore). The decryption calls openssl functions through ctypes. You can join the "bitmessage" chan and post there. I'm not the original author of Bitmessage but I took over from him.
PeterSurda 3 days ago
Okay, Peter, I will go through the whole code thoroughly and see if I can find the error and will report you as soon as I get any clue.And thank you for awarding me the bounty.
Codeword 3 days ago