05:10:29 <Adreik> Does the RPC wallet have an equivalent to the freeze/thaw commands? It does not appear to be in the documentation, https://web.getmonero.org/resources/developer-guides/wallet-rpc.html
10:58:11 <moneromooo> hyc: I'm suspicious of Blockchain::block_longhash_worker. It's calling get_block_longhash, which ends up calling rx_slow_hash with no seed_hash, which means "not an alt block".
10:59:00 <moneromooo> But this is calling many hashes in parallel, so the hash being calculated may not be the one for the tip of the current chain.
10:59:57 <moneromooo> This "alt block" flag is very confusing, and I'm not sure what the fix should be (always set the is_alt flag when called from this parallel function ?)
11:13:22 <hyc> what do you mean it calls rx_slow_hash with no seed_hash?
11:13:53 <hyc> :q
11:15:07 <moneromooo> It goes through: bool get_block_longhash(const Blockchain *pbc, const block& b, crypto::hash& res, const uint64_t height, const crypto::hash *seed_hash, const int miners)
11:15:10 <moneromooo> seed_hash is NULL
11:15:23 <moneromooo> This gets used as !!seed_hash for is_alt.
11:16:04 <hyc> worker calls crypto::hash pow = get_block_longhash(this, block, height++, 0);
11:17:01 <hyc> which calls rx_slow_hash(main_height, seed_height, hash.data, bd.data(), bd.size(), res.data, miners, 0)
11:18:19 <hyc> seed is always provided
11:18:21 <moneromooo> Oh, the !!seed_hash is mine indeed, sorry. Still, 0 means "not alt" so the questiuon remaisns.
11:18:41 <hyc> the seedheight toggle will take care of it
11:18:56 <moneromooo> Yes, that's the thing that I do not understand.
11:19:15 <moneromooo> 0 here means not alt (so main). But those calls aren't necessarily ordered.
11:19:31 <moneromooo> And they could straddle an epoch (backwards, if unlucky).
11:20:49 <hyc> the seedhash slot for the current tip is shared across arbitrarily many threads
11:21:00 <hyc> access to the althash slot is mutex'd
11:21:04 <hyc> serialized
11:21:46 <hyc> crossing an epoch boundary is also already handled
11:23:34 <hyc> note /* RPC could request an earlier block on mainchain */
11:23:59 <hyc> even if is_alt is passed in as zero, we can toggle it on if the block isn't actually on the current epoch
11:24:04 <moneromooo> So does this imply that one should be able to call rx_slow_hash with 0 or !0 for is_alt, and the result will always be correct, it's just a performance thing ?
11:24:33 <hyc> mainly, yeah
11:24:50 * moneromooo looks suspciciously at "mainly"...
11:25:58 <hyc> is_alt means we expect it to be using a different seedhash than main, but it's possible that it will be the same
11:26:54 <moneromooo> Alright. Thanks.
11:30:18 <hyc> basically, is_alt=1 on entry definitely means it's an alt block, but if the divergence didn't cross an epoch boundary we can still use the same seedhash as mainchain
17:30:25 <rnhmjoj> hi all! i'm having trouble building monero-gui 0.16. it seems the application needs two headers (span.h and net/http_client.h) from monero, i think they are in contrib/. i'm already building monero with -DBUILD_GUI_DEPS=ON but these headers are not being installed. what am i missing?
17:32:19 <moneromooo> Try in #monero-gui
17:34:41 <rnhmjoj> moneromoo: i think the problem is in the build of monero, not monero-gui
17:35:32 <selsta> are you following the readme or are you trying to do something else?
17:38:43 <rnhmjoj> selsta: kind of, i'm updating the Nix package that (currently) builds monero 0.15 to 0.16. at the same time i have to update the monero-gui package, which depends on monero
17:39:57 <rnhmjoj> i can build monero 0.16 and it looks ok, but some headers that are apparently needed to build monero-gui from source are not being installed
17:40:34 <selsta> not sure what you mean with installed, monero is a submodule of monero-gui and it takes the headers from the the monero submodule
17:41:14 <iDunk> I think he's not using build.sh. Anyway, #monero-gui.
17:42:18 <rnhmjoj> all right, i'll ask there
17:42:31 <rnhmjoj> thank you