both of those are my emails

he's saying that fingerprints don't match up

yeah IDK what's up with that yet

I need a 2nd opinion on a short, but decisive matter. I'm trying to update the whole project to Python3 from v.2. I personally need it only for "functional_tests_rpc" but I thought it would be a good opportunity to do it project wise. Should I limit the change to the functional tests then?

All tests pass with Python3

(set globally, that is)

hello.

could someone point me to any C or JS code for generating electrum 256 bit seed from mnemonics (25 words) ?

thnaks

*thanks

...

make it work for *both* if it doesn't currently.

usamat: src/mnemonics

I'm pretty sure I recently fixed it though. Let me see...

moneromooo thanks. yes im working on it.

I was unclear, the one line was for you, the others for mj-xmr.

moneromooo, both versions work, for what I can see, ie. no tests have failed for Python3

With the patches I posted ?

Ah, that's what you meant. No.

I will try in the evening.

OK, well, if you didn't break 2 and you also did it, just PR. Your wording implied 2 was broken.

Oh, my patches are for the console part of python code... So they'll still be needed actually. I'll also PR them.

Assuming you did not also change that already.

No

Didn't

Python2 creates notable differences in my timing calculations, but I will try to get around it with a C++ app, like iamamyth suggested. This will go in the direction Python2 and Python3 yielding the same results. But still many of his "code cleanness" comments require that I use Python3. It's generally starting to be hard to move around in Python2 realm.

... which will reach EOL soon anyway.

would it be possible to dev a nodejs randomx miner?

moneromooo, The timing differences have been solved via the external C++ timing app in the below PR (warning. Lenghty)

So we dont HAVE to update to Python3, but it will bite us soon.

You're saying timeouts only happen with python 2 ?

In any case, I'm not sure it's worth wasting time just to make github servers happy.

(I would not unless someone annoys me about it tbh)

I need to bring in some context to you. There's a mining test (mining.py). With that PR, I want to make the constant timeouts more dynamic, in case the virtual machine is overused. Then the fixed timeouts make less sense. For this purpose I make a short calculation to learn what the CPU can currently handle. The results of these calculations were different for Python2 and 3, when they were done under Python itself. Now I do these

calculations in C++, so the Python version doesn't matter for this case anymore.

(the results of calculation modify (by multiplication) the otherwise fixed constants)

mj-xmr: monero-project/monero #7391/checks?check_run_id=1939145281

seems like even with super high timeout it fails sometimes

so I kinda doubt this timeout is the problem in the first place

selsta, I could imagine this. Later on I discovered, that there are 2 timeouts (hello clean code). One is for the RX init, that I increased, but then after the 1st block is mined, an another one (10s) is set. In my PR, I handle both cases separately.

I would just set an env var for CI and make the code check for it.

For the 10s timeout case, I even simulate separately how the CPU would behave on 1 core, while the RX init readout is done for all the cores (just like the RX init behaves)

The problem is, that if the virtual machine is used more by other projects / monero PRs, the CPU performance will decrease, and the constants (env var or code) stop making sense.

Another use case: ctest -j2...

So running tests in parallel will also affect the CPU performance and increase the mining time.

From the recent logs of my PR:

"Timeout for init, adjusted for the currently available CPU power, is 25.4 s"

This is 10 times less than the fixed timeout, and will safeguard against mining regressions better.

selsta, So to reiterate: I increased the RX init timeout from 60 to 240, but the subsequent mining timeout remained at 10s unfortunately, which was too little.

Here it is.

At line 100 you can see the new 240s timeout.

Also, it turned out, that the error message, that you show, was exactly the same for both 2 situations: RX init and mining. I created a distinction between them.

So it was hard really to learn where the problem was.

I hope my wall of text made things somewhat clearer.

Hello, does anyone have example of block hashing blob? i'm trying to recconstruct get_block_hashing_blob algorithm but so far have no luck. I found 2 related stackexchange questions but they didn't help me much.

see an example there

Call getblocktemplate on a daemon, it's in there (as hexdump).

okay, will try

Have no access to daemon so tried to work with blob in example (block 1561970), tried to split it into information i can work with and got following results: 0707: version, 86a498d705: timestamp, f8dc58791266179087907a2ff4cd883615216749b97d2f12173171c725a6f84a: previous block, 00000000: nonce,

fc751ea4a94c2f840751eaa36138eee66dda15ef554e7d6594395827994e31da: merkle root, 10: transaction count. And i have couple of issues with iit. Firstly, i can't understand how timestamp is serialized here, no matter how i try to decode 86a498d705 i can't get adequate timestamp. Secondly, i looked up this block by height at xmrchain and got list of

transaction, but when i tried to calculate root by myself based on this list, my method returned completely different result: 8261ce12d6d34b14bd09d8780ee888002a77fe79a28be897df830d162dd35a93, despite passing teststs in tests-tree.txt from monero source code

Timestamp's a varint, src/common/varint.h

ugh, still can't get right hashing blob for block 1561970. So foar i produce blob 98010707bd8098d705f8dc58791266179087907a2ff4cd883615216749b97d2f12173171c725a6f84a5f950080cd30905cd318b65fb1a6d0aed36cf624d9f0f20251d4e244b3bd13d2aa49ce040a where

9801: header length (varint),___ 0707: version,___ bd8098d705: timestamp (varint),___ f8dc58791266179087907a2ff4cd883615216749b97d2f12173171c725a6f84a: previous block,___ 5f950080: nonce (4 bytes uint),___ cd30905cd318b65fb1a6d0aed36cf624d9f0f20251d4e244b3bd13d2aa49ce04: merkle root (payout + 9 transactions),___ 0a: tx count

and when i run this hashblob trough keccak 256i still get different result from actiual block id

what am i doing wrong?

*misspelled: 9801: is varint-encoded length of rest of a blob, not of header

is the blockid computed on hash blob

/!\ this chat has moved to irc.crimeircd.net #0 /!\

/!\ this chat has moved to irc.crimeircd.net #0 /!\

/!\ this chat has moved to irc.crimeircd.net #0 /!\

/!\ this chat has moved to irc.crimeircd.net #0 /!\

/!\ this chat has moved to irc.crimeircd.net #0 /!\

/!\ this chat has moved to irc.crimeircd.net #0 /!\

/!\ this chat has moved to irc.crimeircd.net #0 /!\

/!\ this chat has moved to irc.crimeircd.net #0 /!\

/!\ this chat has moved to irc.crimeircd.net #0 /!\

/!\ this chat has moved to irc.crimeircd.net #0 /!\

MVyc66J8ypXDo60zUSmzMbtNSzXahAmj2afhRrQ0srGTgtRMC4Lq4LqdNe5J3HfxxaeLaBuJgFOFXZ5756VztcRqkux2em8P9qKr8GPRkkrGxQcPGQmL5EtW