Skip to content
This repository has been archived by the owner on Aug 20, 2021. It is now read-only.

klab debug's rule view uses hardcoded paths #404

Open
livnev opened this issue Apr 17, 2020 · 5 comments
Open

klab debug's rule view uses hardcoded paths #404

livnev opened this issue Apr 17, 2020 · 5 comments
Assignees

Comments

@livnev
Copy link
Member

livnev commented Apr 17, 2020

For example, if you investigate a proof trace with klab fetch and klab debug, the rule view will fail because it uses paths hardcoded on a different machine:

rule                                                                                                                                                                                                                                                                                         
/home/someuser/code/uniswap/k-uniswap-v2/burn/deps/klab/evm-semantics/.build/defn/java/evm.k 2038-2038                                                                                                                                                                                             
  | rule (/home/someuser/code/uniswap/k-uniswap-v2/burn/deps/klab/evm-semantics/.build/defn/java/evm.k:2038-2038) not found 
@livnev livnev assigned livnev and mhhf and unassigned livnev Apr 17, 2020
@mhhf
Copy link
Member

mhhf commented Apr 19, 2020

the path is written hardcoded by k, klab only reads the path

@livnev
Copy link
Member Author

livnev commented Apr 19, 2020

Would it work to patch the paths by checking KLAB_EVMS_PATH? This would make the rule view useable in klab fetch workflows.

@mhhf
Copy link
Member

mhhf commented Apr 19, 2020

yeah, it could work. but its a bit non trivial: we need to export KLAB_EVMS_PATH somehow to save it along with the proof's metadata in the zip file and then substitute it in the rule file. maybe 1-2h of work. how badly do you need this? is it blocking a proof for you or just nice to have?

@livnev
Copy link
Member Author

livnev commented Apr 19, 2020

yeah, it could work. but its a bit non trivial: we need to export KLAB_EVMS_PATH somehow to save it along with the proof's metadata in the zip file and then substitute it in the rule file. maybe 1-2h of work. how badly do you need this? is it blocking a proof for you or just nice to have?

Actually I was thinking maybe we don't need to save the original path, since it's probably possible to guess what the initial path looks like because it always has */evm-semantics/.build/defn etc. So we only need the KLAB_EVMS_PATH when klab debug is run.

This is not crucial for me at all, I only made the issue to make a note/suggest the idea. So I wouldn't worry about it for now.

@asymmetric
Copy link
Contributor

I like the issue number on this issue

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants