.merge+ 7151 7152

Added

I'm trying to compile Monero, but the build script is totally broken. I mean ridiculously broken. How does something like this get released?

First, calm down, I’ve built it fine for years 🙂

What arch and OS are you trying to compile for?

And what issues are you having?

Actually let’s move this to #monero, likely doesn’t belong here.

I built an entire Linux system from scratch, with over a thousand packages. I've seen many, many broken build scripts. Monero is one of them.

Not the worst one, but close.

Are you able to give actionable data, or did you just want to complain?

It's hard to say where to even begin in telling you what's wrong. It would be much quicker to skip past the first 5-6 patches to various CMake files to get to the one error that seems absolutely unsolvable. There are two functions being called by the build script which do not exist in the release package. monero_crypto_generate_header being one of them

<sethsimmons "What arch and OS are you trying "> ^

Custom built, plain jane Linux from scratch distro. It's not the distro. It's your CMake files. Totally broken.

I’ve never had issues that weren’t easily solvable and my fault across many distros.

If they’re that broken I’m not sure why no one else calls building Monero “totally broken”.

I've never had issues with 80% of the thousand plus packages installed on this system.

Again, give some actionable data.

Any good sysadmin knows details are needed to troubleshoot.

Well, I'm not sure how your build system is expecting to call the function monero_crypto_generate_header, when it DOES NOT EXIST in the release distribution.

“It’s completely broken” is useless.

A simple git clone and make release works fine for me on Ubuntu/CentOS and has for years.

If you have specific issues and can give detail please open an issue on GitHub.

(clone --recursive)

Yeah I wasn’t being exact 😉

The instructions in the read me are plainly usable.

Look, I'm not sure how I can explain this to you more clearly. The official Monero release distribution DOES NOT CONTAIN SEVERAL FUNCTIONS CALLED BY THE BUILD SCRIPT. What does "git clone" have to do with it? Your release is BROKEN.

Not for me and many others, but maybe someone with a deeper understanding of cmake can help if you open an issue.

And no, I'm not going to create a Github account and give Microsoft my left arm, selfie in front of a mirror, phone number, kidney, or whatever else they want. I'm just telling you. Your release is apparently broken. So I guess I won't be using or mining Monero.

At least until I can figure out where to even begin unraveling this gordian knot...lol

Then give actionable detail some other way.

I already gave you plenty of actionable details. You don't seem interested in listening or understanding.

Where is that function located, and why isn't it in the release tarball?

You just keep screaming about a missing header (which I honestly can’t help with) and yelling about things being “completely broken”.

Like the 80% of thousand other packages on my system which somehow manage to get their build script right?

You didn't give anything.

Ah, tarball...

monero_crypto_generate_header

monero_crypto_generate_header

monero_crypto_generate_header

WHERE IS THIS LOCATED? NOT IN MY RELEASE TARBALL.

You either clone, or provide submodules manually.

..........

(but sounds like supercop to me)

Anyway, i don't have the patience.

Yeah, that's one of the offending problems.

Apparently this particularly CMake script isn't written to search for system dependencies and either find them, or download it, or quit with a well crafted error message explaining exactly what the problem in. Instead it sends me on a wild goose chase applying patches to CMake files in the hopes of accidentally figuring out what's wrong before going to frustration to an IRC channel where I'm told to "open a github issue." That is what I mean by

TOTALLY BROKEN. So I guess now that I've wasted an hour on this I'll now go 'git clone', and take a shot of whiskey to mourn the really sad state of computing in 2020. Thanks for your help iDunk, and THANKS FOR NOTHING to sethsimmons

Merry Christmas!

Glad I took the time to respond 🙃

hey, i'm building with the make target release-static-linux-x86_64, yet the resulting binary is looking for shared libraries....

specifically readline and zmq

do you have static copies of readline and zmq ?

no, they're the usual linked (system provided)

then that's what you get

hmm. but i've not run into this before the most recent release

if you don't have them as static libraries obviously it can only link against your shared libraries

indeed that does make sense

strange though

you sure your previous build was fully static? you still have a copy of the binaries?

debian testing x64 I am getting an error at 98%+ where do you want me to post the log file

I also have ubuntu lts 18 again x64 if you want me to try it there

monerod is build and seem to work

hyc: no, i doubt they were static, but i do find it interesting that the invalid objects error didn't show up until the most recent release. yes i still have copies. these are docker images based on alpine

i'm not going to dig into it though

pastebin.com/nc42sdUr (make) and pastebin.com/Qx89QUXN make release-test

Are you working on v0.17.1.8 yet?

.merge+ 7141 7142

Added

I got 2 auto banned IP(47.104.129.223 , 83.81.253.202) they are not in block list

auto ban can have various reasons

I only add IPs that are confirmed to do +2 attack

yes if you can can confirm they +2 attack

gah. ok, new version doesn't fix my rpc problems

it did not contain any rpc fixes

well i saw the bans happening on RPC connections so i thought perhaps

2020-12-16 13:55:28.559 [P2P4] ERROR net contrib/epee/include/net/abstract_tcp_server2.inl:1576 [sock 199] Failed to start connection, connections_count = 553

i guess too many socks. not enough space on the mantle

back to cronjob it is

just reduce out_peers on that node

the xmrchain.net has more peers but is fine. so i don't think its p2p

there is no reason for setting it that high on a normal node, I would try setting it to default 12 and see how long it takes to show up again

if you already have issues with too many connections reducing out_peers is the first thing I would try

ok we'll see

usually takes a while

Odd, I've been a pretty well used public node for a long time and have never run into that.

btw mooo asked a couple days ago: "Do you have loads of entries in /proc/`pidof monerod`/fd ?" ?

yeah i entered that command and got a lot of numbers. didn't really know what to do with it. just a grid of numbers perhaps 8 tall by 12 wide

I also don’t know what to do with it :)

Is someone actually looking at the GUI failing to send txes when target > current ?

.merges

I posted last night under a similar nick. If anyone replied could someone please post it here

any idea why this is giving me an error:

debian testing

works fine on ubnutu lts 18 btw

I assume you tried to install requests at least ?

Oh nvm, it goes on.

What version of boost ?

Also, uninbstalling gtest might do it, monero has a version in tree.

libboost-tools-dev 1.71.0.3

non clean build failed trying a clean build

that seems to have done it. Running make release-test noe

now

thank you moneromooo