<josie>
I'm reading through some old discussions regarding obfuscating bitcoin data and while it was done for chainstate, it was never implemented for blocks. I haven't seen any objections to doing it on blocks in the discussions (aside from general objections to obfuscating the any data to begin with), but perhaps I'm missing something. Was there a reason it was never done for blocks?
Norrin has quit [Remote host closed the connection]
<josie>
darosior: that seems to be the original reason for not including blk files when this was mainly about getting around AV scans. for me, I was thinking more in the context of people putting illegal content into the witness data, but given the fact that's been possible for awhile and hasn't caused any issues that I'm aware of, perhaps it doesn't matter?
<josie>
on the other hand, if obfuscating blk data would help avoid some annoying headaches and there aren't technical objections, seems like it would be a nice to have
<darosior>
Yeah i don't know about previous discussions, nor have i thought if it's worth doing something about it, but i can see how it can be an issue if our nodes start storing -say- child porn ordinals..
Guyver2 has joined #bitcoin-core-dev
<josie>
given that storing arbitrary has been possible for awhile and no one has reported having issues, I'm inclined to think I'm missing something and this is a non-issue. But if not, seems worth having a discussion about
Norrin has joined #bitcoin-core-dev
Guest80 has joined #bitcoin-core-dev
Guest80 has quit [Client Quit]
Guyver2 has left #bitcoin-core-dev [Closing Window]
sahil has joined #bitcoin-core-dev
Norrin has quit [Ping timeout: 255 seconds]
<sipa>
I believe the original reason was that big files weren't scanned by antivirus software, but maybe we should just do it anyway.
Norrin has joined #bitcoin-core-dev
bitdex has quit [Quit: = ""]
<josie>
sipa: only downside that I can think of is breaking software that parses the blk.dat files directly, but afaict that's not a use case core has ever guaranteed support for
<sipa>
yeah, and it will prevent downgrading
<josie>
sipa: but isn't downgrading already an issue due to the chainstate db being obfuscated?
<sipa>
i mean: it will prevent downgrading from a version post-blockfile-obfuscation to a version pre-blockfile-obfuscation
<sipa>
presumably a similar issue existed when chainstate obfuscation was introduced, but that's not relevant anymore now
<josie>
yeah, that's what I was trying to say: this isn't a new issue because we are doing blocks specifically. if you wanted to downgrade, you would download an older client and redo IBD, right?
<sipa>
another option is only doing block obfuscation for new installs
<sipa>
or reindexes
<josie>
seems reasonable, that way it would be piggybacking off an already expensive operation
yanmaani has quit [Ping timeout: 255 seconds]
<bitcoin-git>
[bitcoin] SomberNight opened pull request #27097: descriptors: fix docstring (param [in] vs [out]) (master...20230214_desc_docstring) https://github.com/bitcoin/bitcoin/pull/27097
sahil has quit [Ping timeout: 252 seconds]
yanmaani has joined #bitcoin-core-dev
hg has joined #bitcoin-core-dev
sahil has joined #bitcoin-core-dev
bugs_ has joined #bitcoin-core-dev
<bitcoin-git>
[bitcoin] fanquake opened pull request #27098: refactor: don't avoid sys/types.h on when building for Windows (master...windows_can_use_sys_types) https://github.com/bitcoin/bitcoin/pull/27098
<sipa>
I think there are much better checksums than CRCs available if we need one.
<sipa>
CRCs are great at detecting single-bit errors, but relatively slow to compute absent specific hardware acceleration
<andrewtoth_>
Interesting thanks. Any specific checksums you would recommend I should research?
<sipa>
even siphash is faster than a CRC, and that's a cryptographic-quality hash (apart from its length)
<sipa>
there are some non-cryptographic super fast ones
<andrewtoth_>
leveldb uses crc32c, so I thought it would be easy to reuse that since it's already in our codebase
<sipa>
yeah, and on systems with hardware acceleration for crcs, that library will exploit it, and be very fast
<andrewtoth_>
I see but if we were to use one for blk.dat, we can pick one that is faster without hardware acceleration.
<sipa>
Possibly. It's worth e.g. trying a few and benchmarking on varied hardware (e.g. including raspberry pis and similar hardware, which seem popular for bitcoin nodes)
<andrewtoth_>
ok thanks
<fanquake>
I wonder how crazy would it be to deprecate / remove our libconsensus library
<fanquake>
Does anyone know how much real world use it actually sees?
<sipa>
probably not much, as the PR for adding taproot support was never merged iirc
<fanquake>
I have been waiting for someone to report the fact that the Windows DLL has been broken for ~1 year, but it still hasn't happned
<fanquake>
Yea, and the macOS lib didn't even export any symbols for some period. That went unnoticed for a while.
<bitcoin-git>
[bitcoin] fanquake opened pull request #27099: build: produce a .zip for macOS distribution (master...macdeploy_use_zip) https://github.com/bitcoin/bitcoin/pull/27099
sahil has quit [Quit: Leaving]
PaperSword has quit [Ping timeout: 246 seconds]
andrewtoth_ has quit [Remote host closed the connection]
<brunoerg>
is there an easy way to read data from msgs_recv{sent}.dat?
<theStack>
brunoerg: ./contrib/message-capture/message-capture-parser.py could help you
<brunoerg>
theStack: thank you!
jarthur has joined #bitcoin-core-dev
Talkless has joined #bitcoin-core-dev
<bitcoin-git>
[gui] hebasto opened pull request #711: refactor: Disable unused special members functions in `UnlockContext` (master...230214-unlock) https://github.com/bitcoin-core/gui/pull/711
brunoerg has quit [Remote host closed the connection]
brunoerg has joined #bitcoin-core-dev
roze_paul has joined #bitcoin-core-dev
Talkless has quit [Quit: Konversation terminated!]
vasild_ has quit [Ping timeout: 255 seconds]
vasild has joined #bitcoin-core-dev
roze_paul has quit [Remote host closed the connection]
Norrin has quit [Remote host closed the connection]
aielima has quit [Quit: Ciao]
brunoerg has quit [Remote host closed the connection]
brunoerg has joined #bitcoin-core-dev
dzxzg has quit [Quit: Client closed]
___nick___ has quit [Ping timeout: 255 seconds]
Norrin has joined #bitcoin-core-dev
<jamesob>
anyone ever try and generate negative CScriptNums in the functional test framework? I'm finding that bitcoind doesn't recognize them and I'm trying to figure out if there's something wrong about the way the func tests encode CScript([-50]), say.
<jamesob>
nevermind, operator error
<_aj_>
jamesob: you need to use CScriptNum to encode numbers, not just pass them into CScript directly?
salvatoshi has joined #bitcoin-core-dev
hsmiths has quit [Quit: hsmiths]
hsmiths has joined #bitcoin-core-dev
Guest2943 has joined #bitcoin-core-dev
martin_1 has quit [Ping timeout: 252 seconds]
Guest2943 has quit [Quit: Ping timeout (120 seconds)]