--- Log opened Mon Jul 04 00:00:12 2022
03:20 -!- gnusha [~gnusha@user/gnusha] has joined ##miniscript
03:20 [Users ##miniscript]
03:20 [@sipa         ] [ darosior ] [ jeremyrubin  ] [ MatrixBot1      ] [ sandipndev123] [ stevenroose] 
03:20 [ achow101     ] [ devrandom] [ jnewbery_    ] [ meshcollider    ] [ sanket1729   ] [ stickies-v ] 
03:20 [ ademan[m]    ] [ FelixWeis] [ kallewoof    ] [ michaelfolkson  ] [ sanket1729[m]] [ theStack   ] 
03:20 [ andytosh1    ] [ fjahr    ] [ kanishk779[m]] [ midnight        ] [ sanket_cell  ] [ warren     ] 
03:20 [ appservicebot] [ gnusha   ] [ katsu_       ] [ notmandatory    ] [ sebx2a       ] 
03:20 [ clewe        ] [ hugohn_  ] [ ksedgwic     ] [ RiccardoCasatta[] [ shesek       ] 
03:20 -!- Irssi: ##miniscript: Total of 34 nicks [1 ops, 0 halfops, 0 voices, 33 normal]
03:20 -!- Channel ##miniscript created Thu May 20 00:29:16 2021
03:23 -!- Irssi: Join to ##miniscript was synced in 173 secs
10:22 < darosior> So we removed the recursion depth limit when making the code non-recursive. I think we should have a limit nonetheless?  We would happily try to parse insanely large Miniscript descriptors at the moment and run out of memory
10:22 < darosior> (or insanely large miniscripts in the miniscript_string fuzz target, for instance)
10:32 <@sipa> We certainly could have a limit; IIRC you can't have more than 402 nesting levels.
10:33 <@sipa> Though that may change with tr support.
10:33 <@sipa> But also, does the recursion depth limit matter? Someone trying to run you out of memory can give a very big miniscript which doesn't have much recursion depth.
12:11 < darosior> Yes, not only the recursion depth, the number of nodes. Essentially i think of limiting the `stack` and `result` size in TreeEvalMaybe
12:11 <@sipa> oh, that'd be easy
12:27 < darosior> 🤦 we don't use TreeEvalMaybe to parse, obviously.. I'll just work on something and PR :)
12:32 <@sipa> Yeah, the conversion from Script or string to miniscript::Node doesn't use it.
12:34 < darosior> While at it we could limit the number of subs of thresh? I remember we discussed it already
12:36 < darosior> (i don't think it is necessary for not running out of memory)
12:40 <@sipa> The script parser could keep track of the number of non-push opcodes it has encountered.
12:41 <@sipa> The string parser could count how many fragments (incl wrappers) it has "started".
--- Log closed Tue Jul 05 00:00:12 2022