Support Hashcat 7.0.0 with less-rigid status parsing #58
Conversation
|
Naively, wouldn't it be more robust to switch to using |
|
Sounds more robust, but it looks like |
In Hashcat 7.0.0, there is an extra field POWER which tells you how much power each device is drawing. This currently breaks the Hashtopia agent-python when using 7.0.0 because it expects the end of the line to be integer values for UTIL. In order to avoid this problem in the future, be less rigid in parsing how the status line is returned. Print out messages for any unexpected fields.
|
Fair. Normally I would argue that folks using hashcat pre-6.0.0 are probably missing the boat, but if someone has heavily modified some now-ancient version of hashcat and needs to keep coasting on that technical debt, I wouldn't be surprised ... That being said, do we have any numbers about how likely that is? Hypothetical backwards compatibility can only go so far, vs the huge number of improvements even since 6.0.0. |
|
Right now there aren't any tests that run by default during CI to even check that it still works with any of the agents. I don't see any tests that looks through different hashcat versions. So it's actually unknown if they're even supported. Perhaps a refactor using testcontainers could help establish whether the hashcat versions are even supported. |
|
Thanks for your pull request, looks good. You are right about tests, currently the agent is not optimal in terms of code structure. We may work more intensively on this, when the UI transition is over. |
3 participants
In Hashcat 7.0.0, there is an extra field POWER which tells you how much power each device is drawing. This currently breaks the Hashtopia agent-python when using 7.0.0 because it expects the end of the line to be integer values for UTIL. In order to avoid this problem in the future, be less rigid in parsing how the status line is returned. Print out messages for any unexpected fields.