--- Log opened Wed Aug 18 00:00:41 2021
06:22 < meshcollider> sipa: so what are the main things that need resolving in the miniscript repo before merging into core?
06:24 <@sipa> i think we should first resolve all the ongoing work in the miniscript repo
06:24 <@sipa> there are some PRs open, which i owe darosior and sanket1729 reviews on
06:27 < meshcollider> yeah, I did notice while rebasing 16800 that test_miniscript.cpp was not fixed with the new changes (thresh_m -> multi, etc.) 
06:27 < meshcollider> It appears that's covered in #58 though
06:34 < meshcollider> sanket1729: I was just saying to sipa that I'd like to help get this into core as soon as we can, so I am glad to see you also have the same goal :)
06:35 <@sipa> excellent
07:06 < darosior> meshcollider: fwiw i have the same goal, reached out to sipa a month ago and now working with sanket on that :) Great to see there seems to be a momentum for Miniscript!
08:15 -!- muhblockchain [~muhblockc@user/muhblockchain] has joined ##miniscript
09:03 -!- muhblockchain [~muhblockc@user/muhblockchain] has quit [Ping timeout: 240 seconds]
12:36 -!- muhblockchain [~muhblockc@user/muhblockchain] has joined ##miniscript
15:51 < meshcollider> darosior: \o/
19:10 < meshcollider> Do you think issue https://github.com/sipa/miniscript/issues/14 can be resolved simply with this: https://github.com/meshcollider/miniscript/commit/470a43316aede6c04a813e65a7315fa3712146fb 
19:11 < meshcollider> (I haven't got compilation working yet)
19:50 < meshcollider> (I fixed compilation)
19:52 <@sipa> meshcollider: probably!
19:52 <@sipa> i was thinking we may consider switvhing to a different way of parsing that doesn't involve recursive stack.usage
19:52 <@sipa> but this is likely sufficient
19:54 < meshcollider> I see that the actual Parse() function already has a MAX_PARSE_RECURSION implemented for protection
19:55 < meshcollider> avoiding recursion would also make it a bit harder to review wouldn't it?
19:59 <@sipa> maybe; i think it could be clearer too if done well
19:59 <@sipa> but that's probably an orthogonal thing that could be worked on
20:05 <@sipa> i'll have a deeper look tomorrow
22:10 < meshcollider> Hmm yeah after thinking more I think it is feasible. I guess you'd have to make a root node with empty children, push (child, sub-expression) tuples to a stack, and for each tuple on the stack, parse the sub-expression and store it in the child node pointed to, etc.
22:11 < meshcollider> I think you could do it directly with the Node struct in miniscript.h with some modifications, e.g. I don't think c++'s shared_ptr would work if all the fields in Node are const
22:13 < meshcollider> If they weren't const you could make a blank Node and fill its data and children in later
22:14 < meshcollider> Keen to discuss tomorrow with you then sipa!
--- Log closed Thu Aug 19 00:00:42 2021